The Beamway – Technical Details

by Olav Næss


The Beam

The Track Curvatures

Are Strong Beams Needed?

The Poles

In City Streets

Along Roads

Beam Suspension without Poles

Movable Beams

Crossing Water

The Submerged Floating Tunnel

The Wheels

The Trains

The Elevator

The Electric Motor

Hovering Trains

Compressed Air Hovering and Propulsion

No Wheels at all?

A Universal Bogie

Double Action Air Cushion

The Linear Motor

The Suspension

Proactive Suspension

Transversal Force Control

Spine Trains

The Cabins

The Load-carrying Capability

The High-speed Train

Beam Irregularities

A Compromise Bogie

A Lighter Beam


Wagon Exchange

The Control System

Collecting Energy

Building a Beamway


A 2C Metropolitan System






High-speed trains and other passenger trains should carry 6-8 tons in a wagon, so why should they use a train concept for carrying 100 tons? The discrepancy is partly due to the need to go on the ground.

The minimum weight for trains running on the ground was demonstrated in February 2007 when a train was derailed by snow masses on a Norwegian mountain line. The train — called a Signature train — consisted of motorized wagons, each with 72 seats and weighting 54 tons. A local train expert told the media that he had warned against using such a light train, as it could slide like a snowboard and derail. A heavier locomotive, which could push an efficient snow plow, should be used in front.

In warmer climates, where snow masses may be disregarded, trains must still be built to survive collisions with cattle and other large animals, as well as crossing road vehicles. If the track were lifted 4-5 meters above the ground, resting on pillars, both snow mass disturbance and collision danger would be virtually eliminated, so now lightweight technology could be employed. A positive feedback is encountered here: Rail elevation enables light trains, and light trains are easy to elevate above the ground. Such positive feedback situations can create quantum leaps in technology, and this effect seems to be effective here.

An empty lightweight wagon for 70 passengers could weigh only 5 tons, against 40 tons for the railway wagon (according to the now closed Hytran website). If the total weight is used, with motorized bogies, it becomes 6.4 tons, but this weight can be compared with a 40 ton wagon among 6-7 such wagons pulled by an 80 ton locomotive. In both cases the beamway train has only 1/8 of the weight. The 70 passengers in the railway train have a 25 meter long wagon, while the 70 in the beamway train have 16-18 meters, so it may be more correct to use the ratio 1/6. If the weight of the passengers equals the weight of a light wagon, the ratio becomes 2/7.

The use of such elevated rail creates a new paradigm for ground traffic: multi-level ground traffic – without the clumsy creation of new ground levels by using huge amounts of concrete. This elevation could reduce the railway's ground occupation, ground razing and blocking effect (preventing access/crossing for people and animals) by something like 99%.

A conventional railway track is so wide that the train can easily balance on top of it, and this works fine on a flat surface. But when the track is to be elevated on top of pillars, it should be designed like a stiff beam. This means: It should be thick vertically, but narrow. A wide beam will be heavy, expensive, and often annoyingly overshadowing. The train should not try to balance on top of a wide beam, but achieve stability by wrapping partly around the upper part of a narrower beam. We now have a straddling monorail, also called the Alweg type train.

The train itself can be lightweight, but the track — normally a concrete beam resting on concrete pillars — has some problems:

So we let the train hang under a steel beam, and now we have a suspended monorail, also called a SAFEGE-type monorail, or a beamway.

Both the straddling monorail and the suspended beamway have found limited use, having been found suitable only in cities which have been diluted by having been shaped by The Car. They both have been given elevated stations having stairs and elevators. And with this complexity, it may suddenly be found that stations may have to be manned. Old, compact cities haven't room for such station buildings in their streets, and for small rural places, stations should not have to be manned. If trains can have built-in elevators, they can stop almost anywhere – a huge boon for both old cities and for rural areas. The straddling train is now disqualified in this discussion, as its beam prevents the use of an elevator.

The other improvements to be proposed here mostly aim at making the beamway better for rural use – with high speeds. The most important points of this proposal (making this design novel and unique) are:

This technical document doesn't contain all the improvement suggestions. Several other points (easy to explain and understand) are described in the introductory article.

The Beam

The beam is the central element in this transport technology. The other elements, like trains, wagons, bogies and poles, can easily be exchanged at any time, but the beam is the immutable factor which is most responsible for the success or failure of a beamway line.

Our point of origin is SwedeTrack's FLYWAY design, which had chosen the monolithic German SIPEM steel beam. Monolithic means: The beam is not split up lengthwise, but has contiguous steel in a cross-section. This gives the beam high strength and stiffness in all directions - horizontally, vertically and combinations of these, as well as against twisting. But, as beamways should be built with graded turns, only the ability to resist downwards-directed forces needs to be strong.

It should now be noted that beamways are presently only used for local, low-speed lines, having failed to be chosen for long distance, high-speed lines. (The top speed is 65 km/h on SIPEM beams.) For high-speed applications, the builder needs to bend the beam for controlling the curvature and twisting of the line.

Each side of the beam will normally need a gutter - a track (for wheels and/or air cushion pads) with side walls, and this should be firmly attached to an I-shaped beam, which is optimal for resisting vertical forces. The combination will give a J-shaped beam component whose gutter part shape may be more or less rounded. (An exchangeable gutter is likely to be dead weight rather than a beam-strengthener.)

The significant characteristics of the beam is:

When these properties are given, the beam may be produced in a monolithic version, in which a steel roof connects the tops of the J-beams. A holder for the power line (probably going down in the middle) may also be a part of this complex steel profile.

The monolithic beam may be replaced by two J-shaped or C-shaped half-beams, having the walls out to the sides, one half-beam for each track. The C-shaped version is symmetrical, having similar gutters in top and bottom.

This symmetry has some advantages:

Using a pair of half-beams (supplemented by a plastic "raincoat" and power line holder) has several advantages compared to a monolithic beam:

The proposed design, as shown in the illustration, is based on two symmetrical C-shaped half-beams, so we may call it the 2C-beam.
The top of the two half-beams are clamped together so that they act as a single beam – mainly to prevent derailing. Such steel clamps should be placed on at least the middle part of the beam – to limit the separation of the half-beams. Strong versions of these clamps can be used as supplementary suspension points when the beam is hanging under e.g. a cable. They may be long enough to hold a (single half-beam) splice in a staggered beam.
The wheels shown have reinforced hard rubber without air, and they contain a hub motor. This picture is not a true cross-section, as the wheels may not run side-by-side (which will improve the aerodynamics), and the little pantograph going up to the (copper and aluminium) power line will not be adjacent to a motor.

The half-beams should be more economical to produce, and simplify the logistics, transport, and track building. The lengths supplied will depend on the splice method:

Conventional trains and ships can transport 70 meter long goods, so the 60 meter lengths should not be too problematic. This transport is only needed to the most accessible part of the beamway line, like a harbor or railway line. (A temporary line extension down to such a place should be quite feasible.) This is where the work with building the line starts, and then half-beams are transported by the beamway which extends itself.

The half-beams should have the same profile along the entire length, and not have flanges for splicing. This means they can be cut as desired. Flanges on the inside would be a waste of space, hard to connect/disconnect, and they would transform minor run irregularities into major crashes. Flanges on the outside would be clumsy and ugly.

During the construction work, or during later maintenance, half-beams can be bent as follows by a benderan extra strong, mobile beam running like a wagon. It has a truss design, being perhaps two meters high and one meter wide, and is suspended under two extra strong bogies at its ends. It can grab the middle part of a half-beam either mounted above it or being transported near it, and then bend by pushing or pulling. Alternating between bending two half-beams is much easier than bending a monolithic beam.

One kind of construction time bending is valuable: The SwedeTrack page says: "the beams could be prefabricated with an upward bent".  This means: When the beam is bent down by the weight of a passing train, it becomes straight. But this assumes a certain speed.  If the speed becomes higher, the train has to lower itself at each mid-beam.

If there are concerns about metal fatigue caused by repeated bending, a sagging half-beam can be turned upside-down instead. (A half-beam in a curve must be rotated so that its ends exchange positions.)

The pole separation can be considerably increased if there is a tall mast on top of each pole, and a strong cable connects the top of each mast with the middle of each adjoining beam. (Carbon fibers are becoming strong contenders to steel for this purpose, and nanotube fibers will probably become cost-efficient in the future.) The middle of the beam can then be pulled up as much as needed. If sagging at the 50% point is then replaced by sagging at 25% and 75%, a properly positioned bender can fix this, too.

Bending the beam vertically – for a track passing a hilltop or a valley bottom – is a more awkward matter, but it may not be needed, thanks to two design features which we will come back to later:

(More about this in The Track Curvatures below)

The track should have banking in curves. Hanging cabins get their correct side tilt quite automatically, but the wheels need to be relieved of brutal side forces. A simple thing to do is to mount a curved beam in a tilted position, so that its middle part is elevated. To maintain a level track, implies bending half-beams also in the awkward vertical direction, and this may not be practical. When banking starts and stops, the bogies need some time to tilt to the side, so they can't suddenly be introduced to a tilted beam. A long transition zone between beams, having a twisted track, could do the job for low speeds, but it shouldn't be too difficult to twist the half-beams which are at the ends of a turn. It would be a logistical nightmare to provide monolithic beams with different combinations of curvature and twisting; It is far simpler to twist and bend half-beams.

So far, with only the half-beams held up by the poles, we have only a coarse railway, useful only at moderate speeds, and for self-powered (e.g. diesel) trains only.  But this will be OK while the railway is being built.  We will now consider some additions before the beam is ready for ordinary use.

The asymmetric profile of a half-beam will cause it to be twisted when the middle part is weighted down: The lower part will be bent outwards, and the upper part will be bent inwards. This will alter the track separation and cause problems for bogies having fixed axle lengths. The FLYWAY design calls for ribs on the outside of the beam to prevent this twisting. But these increase the cost and constitute a serious visual pollution which will give people the impression of being in a factory, so we should try to dispense with them by avoiding wheels being paired by axles – see A Universal Bogie and A Compromise Bogie, both able to adapt to varying half-beam separation. And the distribution of steel in the thick (upper and lower) part of the beam should be fine-tuned for minimizing this twisting.

Finally, the beam is provided with a plastic cover which:

Some additional details will become evident when we describe how the railway can be built, and some beam profile characteristics will also be discussed below – under The Wheels.

The Track Curvatures

The horizontal curvature is limited by the sideways movement capability of the suspensions, as described in In City Streets below. But this applies to low speeds only. For normal running speeds, the curvature will be limited by the maximum allowable centrifugal acceleration at the speed to be used. A suitable limit might be .5 G, or 4.9 m/s². This will cause a 30° tilt, and a 12 % ("weight") force increase in the suspension. The minimum track radius for a speed v will be: r = v²/4.9. For v = 200 km/h = 55.6 m/s, r = 630 meters. When the curvatures are hundreds of meters, the beams should be bent at the construction site. But this calls for beams split lengthwise: the 2C beam.

The vertical curvatures are determined by the peaks and valley to be traversed. When height differences are small, the track should be kept even by adjusting the pole heights. But if the landscape has major peaks and valleys, there must be slopes between them. The beamway trains will not have any problems with slopes like e.g. 10 %. The slope transitions, however, are potentially more problematic in practice. This could be called the vertical curvature, so should we try to use beams having vertical curvature? Beams must have large vertical stiffness, so it would be difficult to bend them vertically.

There are two considerations here:

Having the beams curved vertically (or having transition track pieces with intermediate slope between succeeding beams) will smoothen the ride, but not remove the real problem: A wagon in the middle of a valley will have its end bogies high and its central bogies low – and conversely when on a peak. The wagon's ability to handle this situation depends on the piston travel length in the lifter – see the cross-section diagram in A Compromise Bogie. A suitable travel length for this piston could be 20 cm. A 20 cm height difference between a middle and end suspension (11.4 meters apart?) of a 24 meter long wagon means: .2/11.4 = 1/57 radian = 1°. That is: The slopes of two succeding beams should differ by less than one degree.

If maximum vertical curvature is needed, the beam length should equal the wagon length. Starting or ending a 10% slope will then take at least 9 intermediate beams, with a total length of 216 meters.

On a city track, where wagon lengths up to 12 meters are used, the slope increment will be doubled to 2 degrees, and if anybody should care to use 12 meter long beams (feasible in tunnels), the 10% slope could be started or ended (with four intermediate beams) in just 48 meters. If long wagons are needed in such difficult places, they can have a simple articulation in the middle. These need only bend 2 degrees in the middle, and perhaps only vertically, so those awkward bellows shouldn't be needed.

It might be useful to have beams with a vertical curvature delivered from the factory – at least as an alternative to using 12 meter beams. Or old, sagging beams will be appreciated. Anyway, the symmetry of the 2C beam will be valuable now, as we needn't distinguish between valley and peak beams.

In a track with staggered beams – where the half-beams are not paired, but are splints for adjacent splices – there is no opportunity for proper slope increments. If a span with staggered beams has to curve over or under a central obstacle, vertically curved beams are needed.

Are Strong Beams Needed?

It may seem obvious that heavy trains will need a strong and rigid beam, but this isn't really the case if we are able to control the load distribution. We can check this out with a multi-step scenario. We prove our point already in step 1, and then we will relax our assumptions and become more practical.

  1. We have no beam at all – just short rings on top of the poles, and they are fully lined up. The train is like a spear: stiff, but unable to rotate, and may be quite heavy. It is more than twice as long as the distance between the poles. There could be motor driven wheels in the bottom of the rings.

  2. We reduce the stiffness of the train by giving it some ability to bend sideways, but not vertically. We also connect a flexible tube between the rings. Its purpose is to guide the train to the next ring, but it needn't support any weight. The first part of the train (10 %?) should be able to bend a little in any direction, so that its sides will receive the guiding forces. Our train can now do turns in the horizontal plane.

  3. We transfer the passengers to wagons hanging under the spear (because of numerous complaints about the bad conditions in that narrow tube). The spear is now just a bogie, acting as a spine, but its beam (tube) loading is still negligible.

  4. We want to put the wheels on our spine bogie instead.  (It was impractical to spin up and power all those pole wheels, and we wanted to distribute the weight over many wheels.)  We also want conventional independent bogies, so the spine stiffness is moved down to the train instead.  And still it works: The rigidity of the train causes the weight of the train to be carried only by the (somewhat lengthened) rings on the poles, not on the weak and yielding tube connecting them.

  5. But now we must leave paradise: We can't stay in one plane, as the train has to go uphills and downhills, and it has to bend also vertically when passing the top or bottom of a hill.  (We had a train for the comfortably flat Netherlands and Denmark, but not the beamway-needing Switzerland and Norway.  There is a limit to how much height variations can be eliminated by means of tunnels, bridges and varying pole lengths.)

We have now two tricks to remedy this situation:

  1. The simple: Spread the beam loading to the first and last part of the train. We do this by pulling the wagons together at the bottom and pushing them apart at the top.  Our special “Exchangeable cabin” philosophy is well suited for this: The rigid frame on top of the modules is eminently suited for pushing apart, while a lightweight cabin easily can exert a pull through its floor.
    This should, however, be done to a limited extent with long trains, as the load at the ends would become too great.

  2. The complex: Use the active suspension of the train in such a way that the weight of the train is shifted to the bogies which are nearest to a pole.  (See “Proactive Suspension” below.)  This method is important for long trains, which could transfer their load to at least two poles, and may not really need to put load on beams. The spine, originally able to bend only sideways, must be replaced with a version with dynamic rigidity: It will vertically always actively assume a dynamically rigid shape which is a copy of the track profile covered at that point of time – except that it will not follow a yielding beam downwards.

It should be noted that if these load distribution schemes fail, the beam will not collapse, but will (due to the excellent overload characteristics of steel) cause added downwards bending (sagging) of the beam, thus necessitating half-beam reversals.

A useful design for the spine train in point 4 above: The train has a vertical chassis plate (a light framework) which hangs below the center of the beam. It becomes a flexible spine by being divided (by vertical cross-sections near each bogie-held suspension) into sub-plates, which are connected to each other by double hinges. That is: There is a hinge at both the upper and lower end of a sub-plate connection line. This point is important when we try to adapt to point 5 above (vertical flexing): The upper or lower hinge must be designed so that the sub-plates can here be pulled together and/or pushed apart. (See the cargo wagon picture below.)

Such a central chassis would be OK to hang cargo containers on (on both sides), and it can distribute the weight over a much greater distance than the total cabin length, but the passengers would not like to be separated by such an iron curtain. We might now try to find remedies for this problem, and should bear in mind that we could make quite large holes in the chassis plate without losing much strength. In fact, it is common practice to make large holes in such structures in order to save weight, but this works best with circular holes.

We could make holes for doors: would work, but the wagon would be strange, and we would have no useful cabin for use elsewhere.

The entire cabin could be placed in a huge hole of the chassis: Satisfactory chassis strength might be obtained if the chassis gets ample space around each cabin, but the cabin could not be lowered down.

The chassis could be shaped to fit around a decent cabin: along half the roof, then down along a side wall, with large openings for the windows.  The main problem with this chassis is that it is absent at the middle of the floor level, where wagons should be firmly connected.  Besides, it would probably be ridiculed for its strange asymmetric shape, and the aerodynamics of such a perforated plate would be bad. (This was the design I played with in the beginning.)

The exoskeleton

The exoskeleton is symmetric, and it fits around a cabin on three sides. The lower crossbars improve the solidity and enables solid connection between wagons also at floor level. The exoskeleton would probably be so strong that vertical bars between the windows can be dispensed with. This will improve the aerodynamics as well as the appearance. Quite weak cabins can now be used, but it should be asked: Wouldn't the extra weight of the exoskeleton offset the strength gain? Is there any point in its ability to push at the floor level if just a pull is needed there (which a cabin can easily do)?

If we don't need the exchangeable cabins, it will of course be simple to make the exoskeleton invisible by integrating it into the walls and roof of a quite ordinary passenger wagon.

Regarding the vertical chassis, it should be remembered that the train must be able to bend sideways at every second suspension point, and passenger cabins will normally need at least three suspension points. Bendable cabins would be rather awkward, and so would a flexing chassis plate near a rigid cabin. The solution is clearly: Let the suspension points be able to shift sideways above the roof of the cabin, preferably integrated in a suspension unit which also deals with vertical movements. This was anticipated in the picture above.

Conclusion: Try to transfer the functionality of the vertical chassis into the design of the cabins and their suspension.

The Poles

The purpose of the poles is: to lift the beam so high over the ground that there is room below for the train + ground activities which should not come in conflict with the train.

A pole consists of:

The base should preferably be embedded in the ground, but if this is not practical, it should extend along the ground surface in perhaps both dimensions, depending on whether or not the upper part of the pole has any support from the side. If the railway track is to follow a city street, the edge of the sidewalk may be the only suitable place for the poles. It may then be practical to use a base which extends only along the edge of the sidewalk. But that would be OK if the poles stand in pairs, one on each side of the street, forming a rack , a rectangular arc, by being connected by a horizontal support beam under which the rail beam is mounted. 

Two beams, for two-way beam traffic, can thus go along the middle of the street, and that may be the only practical placement if the street is lined with trees.  Considerable street zig-zagging can be smoothed out for the beamway if beams are suitably placed on such wide arcs, and by alternating between arcs over the street and poles e.g. near street corners.

The double pole shown above is very stable if it holds the beam at a fixed angle. All the poles may then simply stand upon the ground – although wide feet as shown below will be convenient before the beam is fastened. Even the forces from emergency braking can now be handled, as the beam will distribute the forces between several poles/racks. The stability may be so good that every second support can be a single pole (as shown in the background in the picture above) which simply stands upon the ground, thus dealing only with gravity.

Such poles – able to simply stand upon the ground – can really simplify beamway construction across shallow water.

The pole stem may be telescopic, at least in uneven terrain, to contend with varying height requirements, and to fine-tune the track for high speeds. Along the stem it may be nice to have one or two channels, covered by a lid which can be opened when cables are to be mounted in a channel. A channel may be dimensioned so that an ordinary collapsible fire escape ladder fits into it when the lid has been removed.

The beam holder on the top of the pole is bent to the side so that the train passes at a safe distance from the stem – even if it swings to the side. A pole should be designed to carry two beam holders, pointing in opposite directions, so that it can carry two beams, for two-way beam traffic.  (This is seen in the background in the above picture.)

Here we see some free-standing multipod “poles”, suitable for a provisional beamway line. The type to the right may become the standard type, normally used with the feet buried.

Poles should also be prepared to hold street lights and signs.

The upper end of the pole should not be attached directly to the beam. The end should hold a strong transition bracket, shaped like an upside-down U.

Here we see how the transition bracket holds 4 steel tongues, each able to hold the end of a half-beam.  Each tongue is attached to an internal slider on the inside of the transition bracket, so that varying half-beam lengths can be accommodated.

The transition bracket is held up by two horizontal steel rods which are held by a pole or between two poles.

Here is a section of the same. (The tongue must be thinner if large wheels are used.)

If the beam is fastened only by the tongues which are inserted under the top of its ends, only careful construction traffic is possible – mainly transport of the splints, which will give the beamway improved solidity.  

This picture clearly shows a gap between the half-beam and the bracket wall.  This is the space for the splint.The splint is a steel plate as high as the beam, a few centimeters thick, and extending perhaps 10 meters to each side, firmly holding the half-beams, and strengthening the whole structure like the splint at a broken bone.  It should have tapered ends which visually blend with the beam – no point in getting visual noise here.

When the beam has been properly fastened, any gaps in the railway can be bridged by means of short rail pieces which are attached directly to the splints.

The splint may be mounted inside the bracket before the half-beam is mounted, but it should be possible to press it up afterwards, as the half-beams should be able to move their lower parts sufficiently sideways – even if the splints are mounted from a vehicle which rolls on this unfinished beam.

Here is a bracket which can be used for clamping together a half-beam and a splint.  
Similar brackets may be used for cable suspension, and the splice for a single half-beam (in staggered configuration) may be covered in this way.
The curled end of this bracket should preferably go into the split edge of the beam.

When I propose that beams should be joined at the poles, I disagree with FLYWAY, who hold that beam joining should be done at the inflexion points, because this is where forces trying to bent the beam are minimized. Yes, it is tempting to choose these points, but the bending forces are easy to deal with when we can use strong (and heavy) splints which are mounted with their centers of gravity on the poles. If the beams are firmly fixed for 20 meters at a pole, it is like having a 20 meter wide pole. The 30 meter pole separation required by the Sipem/FLYWAY design will now become a 50 meter distance between pole centers, but it may be more realistic to assume this job of “pole widening” is only done halfway, so we will operate with a pole distance of 40 meters.

Another advantage of this firm beam fixation: Between the beam ends there may be a short stretch with rails which are held by the splints. (There could be a slit or ridge along the middle part of the splint for holding rail pieces.) This has several valuable possibilities:

If the transition area (and its transition bracket) is several meters long, new possibilities open up:

The use of firmly fixed beams gives the whole beamway added strength, so that it can withstand strong forces, like from an emergency braking train.

In City Streets

In streets without trees, the beam train will normally be able to take wide, lazy turns by choosing whatever position that suits it:

As we see here, even a double-tracked train can have a turning radius which is several times greater than what a bus or tram can get.
We can see from this picture that it is natural to have beam section lengths corresponding to 1/8 of a circle, and the section length will be (Pi/4=) 78,54 % of the turning radius.

If we assume a 3 meter track separation, it will be natural to produce and stock half-beam sections in 3 meter turning radius increments, and with these values for turning radius (and length): 50 (39.3), 47 (36.9), 44 (34.6), 41 (32.2), 38 (29.8), 35 (27.5), 32 (25.1).  The two last ones seem to be appropriate in this example.

If we use the same beam type in the two tracks, the length error for a quarter circle turn will be a quarter of the circumference increase caused by a 3 meter radius increase: 4.71 m, or 1.57 m in each of the three beam joins at the turn.

A long transition bracket is ready to fill in this gap, but may not be practical to use regularly on street corners, particularly if rail curvature is to be maintained.

The two half-beams in a 2C-beam have just 20 % of this separation, so they can certainly have the same length.

The ride will be more comfortable if the centrifugal acceleration could increase and decrease gradually (perhaps following a Gauss curve?).  That means the curvature should be strongest near the middle of the turn, and gradually weaken further away.  Perhaps a hyperbolic shape would be ideal.  It is, however, practical to use straight beams as close to the turn as possible, and that should not be so difficult, as the suspension of the beamway train can absorb quite much sideways irregularity.

It may not be so difficult to get a near perfect path if the previously described beam-bending schemes really work.  Then the half-beam parts near the center of the turn can be made more curved, while the parts further away are made less curved.  If this is done in a balanced manner, the distance between the ends can be kept reasonably constant.

So: How small can the turning radius be?  Such turn-on-a-dime stunts are easy for small vehicles with just one or two suspension points, but we are here concerned with real buses/trams/trains having at least three suspension points, so now the sideways mobility of the suspension points becomes the limiting factor.  If this is ± 40 cm, the turning radius for our little yellow friend with 8 meter long wagons will be 8 m, but values less than 20 m are unlikely to be required in city streets.  For the brown train with 24 m long wagons, the turning radius could be 75 meters, but empty wagons could be transported through much tighter turns if they for the occasion were using only two or three suspension points. (The design goal should be ±80 cm sideways movement.)

The sight of such a beast above might scare the city dwellers, but vehicles passing a step beside them should really scare them much more.

These two beam-holding tripods (attached to house walls) are actually the same design, but mounted in two different configurations.  
Here we also see the use of (triptych-style) hinged splints in a turn.  Curved splints would be preferable.

Along Roads

As beamways can follow roads, the railway and the road needn't give conflicting influences to the population distribution. The beamway can in many ways improve its path relative to the corresponding road. New roads for fast and high-volume traffic are often routed outside cities, but the beamway can jump over to old roads in order to come near the city center. Roads will often have to take detours around e.g. river bends – at least old roads will be found to have many detours which newer roads would have avoided. The beamway, however, will take shortcuts over rivers, gardens, parking lots, garages etc..

In order to achieve high speed, the beamway will try to straighten out its path, and frequently cross a road at acute angles.

Such a beam-holding rack gives high track configuration flexibility.
The poles should be telescopic. The horizontal part should be displaceable and perhaps telescopic.

The use of these racks along a road (compressed lengthwise).
The first and last two could have been replaced with simple poles, but with the shown arrangement, it is simple to add a second beam for dual-track operation.
If the road is altered, the poles of these flexible racks can easily be moved.
A widened road will probably mean narrower racks, as one pole in each rack may be placed in a central double fence.

The map above shows how a beamway (blue) can position itself in relation to a new road (red), old road (yellow) and old railway (black). The rack described above is used in various configurations and position above and at a road or railway. The terrain may be flat in other places, too, but agriculture and other land uses will be less disturbed through this positioning strategy.

Beam Suspension without Poles

Between two strong buildings/rocks, some beam splices may be suspended in a V-shaped wire.

In difficult terrain, it may not be possible to have a pole every 40 meter along the track. If the beamway follows a road across a bridge, it should generally not be difficult to let the beam follow the road at a higher level. If it is a suspension bridge, the construction may be simpler then normal if the poles can be tied to the vertical cables. If the road has to make 90 degree turns before and after a short bridge, the beamway can take shortcuts at the ends, and may be able to cross at a more convenient angle like 45 degrees. And if this bridge is high, the beam could hang under it.

If there is no bridge, the beam could be suspended in the same manner under cables. In this case the joining of the half-beams should be staggered: Instead of a double join with splints every 40 meter, there should be one half-beam join in a lightweight, wire-suspended bracket every 20-30 meter.

It should be noted that the beam shown above is hanging under straight cables, and not under the catenary cable commonly used for bridges. This is because a train will significantly increase the bridge weight and deform the catenary cable, bending the beam where it passes. A road bridge, in contrast, has a quite constant weight, so its catenary cable will retain the curvature.

If the beam follows a suspension bridge or hangs under its own cable, it should be practical with a short distance between the suspension points. The beam could then be thinner and lighter here, enabling a valuable weight reduction.

This should be the preferred design for long, straight tracks. The beam there should be ready for being moved along its length now and then. When the beam sags down between its suspensions, it should be moved a distance equal to half the suspension separation (and beam pieces equal to this distance should be moved from the front to the rear of this movement). The beam will now be held up at the sagging curve, and further sagging will straighten out the beam. This applies to beams held by poles, too, but the suspension separations should be uniform.

Such an A-rack can adapt to uneven terrain by having its crossbar in various positions, and by varying its leg lengths.
The lowest position shown allows normal elevator use. The higher positions allow emergency evacuation – with an unsupported elevator hanging under long cables.
The distance between these A-racks can be at least twice as long as between normal poles.
With beams in the lower position, support of the intermediate beam part can be from cable suspension as shown in the previous picture.
With beams in the higher positions, the support can be from simple poles (standing freely upon the ground, giving only vertical force). Or the A-rack can be extended by a simple cable-holding pole on the top.

The beamway can get a stiff and long beam by using an I-beam or truss like this.
The 2C-beam hanging below it can be much thinner and lighter (and asymmetrical) now when it needn't function as a beam.

If the beamway follows a road tunnel, the beam should be embedded in the tunnel ceiling, or it will be in the way for high trucks. It may be feasible to affix it in a groove in the rock, or it may be necessary to set up large holding brackets which follow the profile of the tunnel.  A vehicle with the same shape (minimally obstructing the traffic) could be used for cutting the groove in the tunnel ceiling, and for mounting the beam.

The train itself would not be in the way, as it would act like just another bus. In such a case it may be wise to let the poles just outside the tunnel stand in mobile racks, so that the whole beamway can be moved to the side if it gets its own tunnel.  A tunnel for a beamway will be just a simple little hole (< 4 m Ø).  If it is placed near the main tunnel, it can double as an escape tunnel.

Movable Beams

A simple trick with movable beams is: Make a track switch by pulling one end of the beam so much sideways that this end is aligned with another, adjacent track.  That should be about 80 cm to the side, so such a switch can be passed at a quite high speed.  If the two half-beams can slide in relation to each other, the problem with track length variation at the gap will become quite small.  Anyway, some transition track telescoping, involving obliquely cut tracks, may be useful.  It will be difficult to fix the movable beam end properly with splints.  It could be fixed with half-splints to a pole whose top can be shifted sideways, or the movable beam could simply be made a little shorter.

A more drastic beam movement is: Get the beam out of the way for crossing ships.  A beam is a far simpler object to swing away than a road or railway track. 

Here are two designs for bridges that can be opened.

And here are two more – for small rotating bridges.  

The rotating bridges could, if made large and strong enough to turn while carrying a train, be used for turning a train at the end of the track, or for sending a train/wagon into one of several parking tracks.

And now a more impressive trick:

The train takes the elevator up to the mountain plateau

In this picture we can also see: 

The beam hangs under brackets which are attached to the vertical cliff, and it hangs in a V-shaped wire when jumping from one mountain to the other.

The elevator could easily be designed to rotate to align with any one of several receiving tracks.

The ultimate use of a movable beam would be: Mount a pole on each of two ground train wagons, going side by side on two parallel tracks, so that the beam between them can sweep over a farming field. If the two poles are tall towers, each with wires from its top and down (on both sides) to adjacent parts of the beam, they could hold a beam which is more than 100 meters long. An unmanned train on this beam can cover the entire area and automatically harvest food or energy crops in also difficult environment – from deep water to steep hills, and also over trees. When the beam is in the middle of the field, it should line up with a beamway line through which trains can go out and deliver the crops at remote destinations. The low tracks may be somewhat elevated (or fenced in), and then automated machinery may not interfere with most animal life. People may find paths in this terrain ok for walks. The low tracks could be very steep, with pulling wires like a funicular. If the earth is not to be worked with, the machinery needn't go on the ground – just up on the beams – so the whole process can easily be automatized. The lower parts of the plants may be left untouched. This means that Coppicing for energy harvesting may be replaced with pollarding over a relatively undisturbed animal life. Chopping the wood to pellets may be done during harvesting. Very steep or uneven terrain, useless for conventional agriculture, may be used. More details here.

Crossing Water

The light beamway is well suited for crossing water, with the beam being carried by pontoons. If the half-beams are staggered at the crossing, the beamway line will constitute a quite solid long beam. There will be a range of designs to choose from, depending upon the size (and number) of boats that have to cross underneath:

  1. No boats: The poles on the pontoons can be very short.

  2. Only small boats: The poles are like those used on land – for lifting the beamway over normal land traffic/activities.

  3. Larger boats: Special wide pontoon configurations may be used (for towers or tall poles) in the parts of the crossing where large boats may pass. If there is deep water near one shore, it may be convenient to locate this passage here, as it is easier to build a tall pole/tower on land – especially if the shore is elevated.

  4. Ships and tall sailboats: For light ship traffic, a beam that can be lifted out of the way (as depicted above) may be used, but a pontoon-carried beamway will probably not have sufficient position accuracy for such a mechanism. For crossing heavy ship traffic, a submerged floating tunnel should be used.

The height available for the boats can be either: below the beam, or below the bottom of the train. Letting masts – and perhaps other, more dangerous parts of boats – pass in collision course with trains, might be feasible if boats tall enough to collide were only those piloted by professionals who could be expected to observe train schedules. The only safe precaution may be to block the crossing path with wires at train bottom height, although parts of this blockage could be booms that are lifted when no train is near.

The water surface area of the pontoons should be so large that they don't sink much while they are carrying a train. The stiffness of a staggered-spliced beam should ensure a good weight distribution over many pontoons. Large and slow waves may cause a danger of metal fatigue in the beam, and therefore necessitate rigid anchoring in exposed waters. The tidal movement should not be problematic if beams crossing the shores are mounted with articulation – prepared for some bending in joins. The last pole on the shore may have to stand on a little bridge having one end on land and one end on a pontoon – for halving the tidal amplitude there, ensuring a small enough vertical slope change. Or a long span will be needed: A one meter tide change at the end of a 57 meter long span will give a one degree slope change – the limit for long trains.

For crossing long distances (and/or for moderate train traffic), the beamway train could use a ferry having on its deck a beam which (at either end) is connected to a beamway line when the ferry is moored:

A catamaran ferry for short trains. The receiving beam has a movable end – for obtaining connection with the ferry's beam.
Also electrical connection is obtained, so that the ferry can run on rechargeable batteries.

For still higher flexibility, a cabin could be lowered upon a ferry/barge, but this would necessitate another train ready to receive the cabin after the crossing.

The Submerged Floating Tunnel

A round steel tube, 3-4 meters in diameter, should be a suitable tunnel for the beamway. Ventilation and illumination should be unnecessary. The light weight and quite derail-proof traction of beamway trains will make such tunnels far simpler and safer than what conventional trains require.

A simple submerged floating tunnel. (Without reinforcing structures, and without supplementary pontoons at the sloping tube)
can be towed out from a shipbuilding yard.
(Vertical dimensions are exaggerated in this picture; the tube diameter will be less than 4 meters. The anchor wires/rods should actually be spread in the image depth dimension.)

The tunnel consists of three parts (and soft transitions between them):

  1. Sloping down from one shore

  2. Horizontal in the middle

  3. Sloping up to the other shore

The sloping parts will hang under pontoons having a quite small separation. Small boats should be able to pass over these parts. The slope might be as for undersea tunnels (10% in Norway).

The middle (horizontal) part should be 20-25 meters below the surface to allow passage for ships. It would be advantageous if pontoons could be used in only three places: In the middle (to separate the two sailing directions) and at the transitions to the sloping parts (to indicate the sailing channel limits for ships). Under the middle (and slightly lower) part of the tube there might be a reservoir for collecting any leakage water. It should have a water level detector and a bilge pump. This reservoir could be shaped like a V-shaped keel, which could reinforce the tube. (But, as the bottom of the tube isn't used by the train, it is ok if any water forms a long pond there. The bottom part is also needed for passing creatures – on two or four legs – straying in there: Topple them over and run over them.)

An important design consideration is: The tunnel tube should be designed to not collapse if it accidentally became filled with water, even though a water-filled interior will dramatically increase the weight. It will be easiest to design a tunnel strong enough for both situations if the upwards forces from an empty tunnel were as strong as the downwards forces from a filled tunnel. The tendency for the empty tunnel to float up should not be attempted neutralized by means of weights attached to it, as this would make a water filled tunnel too heavy, but by means of anchoring wires to the bottom. (Anchoring stiff rods, if possible – for holding up a filled tunnel.) And the pontoons should be heavy (and attached with stiff rods) to counteract the buoyancy.

This strong tube enables heavy trains to pass there. (It is easy to attach the beam very firmly, and of course with short intervals, inside the tube.) Trains in the tube are consequently well suited for transporting cars safely under the water.

For increasing the stiffness of the tunnel, a T-shaped steel profile could be welded to the top of the tube (where it would also protect against ships and perhaps bombs dropped from above, but necessitate a larger tunnel depth). And/or a similar inverted profile could be welded to the bottom. The vertical part of this profile should have large holes – for reducing the drag from cross currents.

The trains to be used in the submerged floating tunnel should preferably have backup batteries (for leaving the tunnel in case of a power loss), and they should have rubber wheels pinching the beam from above and below (see picture in The Trains below). This will enable trains to pull themselves out of a tilted/flooded tunnel. (If wagons float up, ordinary top wheels will loose traction.) Another alternative is that a locomotive with such double traction is stationed at each end of the tunnel – for automatized rescue operations.

Both cars and conventional trains would be helpless in such a situation.

The Wheels

Conventional trains use wheel flanges for ensuring precise tracking. Even if rubber wheels are used, it may be necessary to use such metal-against-metal contact if extra strong forces become required for tracking. The picture below show three ways to do this:

An important function of rubber is: To prevent the huge forces caused by a metal object lying in the track. It will be advantageous to have springing action in the zone between the hub motor and the rubber rim, mainly as an extra insurance against such dangerous jolts.

Another method to ensure good tracking for rubber wheels: Let each bogie have horizontal wheels running against the side walls.  But the simplest tracking method for rubber wheels will probably be to let them have steel walls.

Outsiders: alternative beam-wheel combinations:
Left: A flanged wheel in a grooved beam. (Unlikely to be chosen, as it just follows an old tradition.)
Middle: A beam for flangeless wheels, with a steel wheel normally following the weak tracking force from the wheel path curvature. Could have a zone of vibration-dampening rubber or plastic inside the outer steel rim.
Right: A rubber wheel runs better in this grooveless beam.

The right part of the picture shows a cross section of a special rubber wheel designed for this beamway. This wheel has two circular steel plates which can take the strong guiding forces from the side walls of the beam. These plates also function as a steel wheel if the tire is destroyed or just looses air pressure. When the outer wheel plate is removed, it is easy to change tire. There are only tiny holes between the tire cavity and the remaining space between the wheel plates. This means that the short-term springing action of the tire has the progressive characteristic determined by the small tire volume, and still has its average tire pressure evened out by the larger adjacent volume.

These rubber wheels should automatically be kept fully inflated by an air pump which is supplemented by a safety valve limiting the pressure in each tire. This air pump could have a very low capacity, and could be:

The optimal tire pressure may depend on the speed. This could be controlled by a microprocessor on the bogie, or by a simple valve in the entrance to the tire cavity, if this valve is controlled by the centrifugal force.

It may be useful to have a microphone on the bogie – for detecting unpressurized tires. This might work best with a microprocessor on the bogie.

The rubber wheels on such a smooth surface can do with very little air – just so much that there is a thin air layer in the lower part when the train is standing still.  This small air layer will prevent the rubber from being deformed.  When so little air is used, getting a flat will not be so serious.  A good, elastic material which does not deform permanently, could do without air altogether.

A rubber wheel must be quite wide, but only in the outer part. The inner part – without rubber – can be quite thin. Car wheels are shaped this way, having the brake mechanism in the space made available at the thin part of the wheel. This place is well suited for mechanisms involving the interaction between rotating and non-rotating parts. This could mean motors instead of brakes, and as motors can do regenerative braking (supplemented by beam-pinching emergency brakes, as depicted below), the beamway should use this opportunity to get a motor in each wheel. This will ensure good hill-climbing capability, but perhaps not give enough total power for high speed, so additional locomotive power might be needed.

The alternative beam profile presented here has some noteworthy properties:

Air-filled tires like this are not very reliable, and could be avoided if air cushion hovering is used for carrying most of the train weight, thus decreasing the forces on the wheels. See A Compromise Bogie below.

The Trains

A conventional train rides upon its bogies (trucks); a beam train hangs under it.  The beam train bogie is essentially a mini-version of the old heavy-train bogie. If the bogies are motorized, they may be regarded as small locomotives – microlocs.

The conventional train has traditionally consisted of a locomotive pulling some unpowered wagons, but modern passenger trains tend to have the locomotive functions (electric motors) integrated in the passenger wagons. This is like a laptop computer: elegant and convenient when everything works well, but awkward when a problematic part needs to be replaced. A wagon or train full of passengers is ill suited for motor replacement.

The beamway train can and should be designed to combine the advantages of both approaches.

The elements making up a small beamway train/tram/bus.
(Shapes are not shown in this diagrammatical representation.)

A beamway train should have one or two wagons. An elevator should be in the end of the first wagon, or constitute a separate "wagon" employed between the two real wagons. The wagons hang under a series of "microlocs" – small locomotives handling both the suspension and propulsion. These have traditionally consisted of motorized bogies running inside a hollow beam, but we will mainly use air cushions for holding up the train, and then it will be more appropriate to speak about sleds instead of bogies.

From each sled goes a suspension connection (controlling the height and providing a soft suspension) down to the wagon. The wagon only needs to have the small motors which displace the suspension sideways.

Microlocs might be programmed or remotely controlled for running alone, but will normally work in teams, and usually for carrying a train as depicted above. Alternatively, a team of something like 3-10 microlocs may be connected to a very thin "wagon" which is hardly more than a beam. When this is connected to an end of a train – perhaps the nearest suspension – we have something like a conventional locomotive. Such a miniloc, used before and/or after a train or wagon, can be useful for giving a train some extra pulling power.

Beamways may run on different kinds of beams. We are concentrating on the 2C-beam, but the train shouldn't be concerned with what kind of beam it is running under. Different kinds of beams will require different kinds of microlocs, but the lower part of the microlocs – the suspension rods and their connection to the wagon top – should be standardized, so that the same wagons may be used with different beam types.

Braking is primarily done by letting motors work as generators (regenerative braking), but for strong emergency braking, friction brakes are essential. These conventionally act through the wheels, but beam trains can do it much better:

Here we see the brake pads of the beam train.
They brake much more efficiently by pinching the beam.
And the beam carries the heat away rapidly.

If motors on the bogies turn out to be too weak, too heavy, or causing too much air resistance in that narrow beam, a miniloc like the one depicted below may be useful

A large motor should be placed in a nacelle, resembling a bomb or torpedo when it hangs under the beam.  It could then function as a locomotive which can run by itself to another train which needs to be pushed or pulled.

Meet the free-lance locomotive: It has a lance, perhaps 10 meters long.  This is a beam containing a motor shaft, and the mechanical power is transferred to the bogies in both ends, as well as to the connected bogie of a preceding or following train, through either end seen in this picture. If the motor is displaced (along its beam) to the remotest position, the weight distribution of the train is improved.  

When the middle wagon of the train pulls the two adjacent wagons at floor level, more of the train's weight comes on the first and last bogie.  This not only improves the weight distribution, but also improves the traction.

A separate loc may be used at the same time as small motors in the wheels. The loc may be needed for high speeds. A heavy loc is valuable for providing high pulling force. But if the problem is to achieve high speeds, high power can be delivered while using low force, and then a light loc can be efficient.

The free-lance locomotive will run by itself to help another train, if needed.  A bag with a few hundred kilograms of luggage could also be hung under it.

The Elevator

Having an elevator in the train dispenses with the need for elevated station buildings at the stations – which would need to have elevators. Station buildings with elevator could easily become so complex that they would have to be staffed. Being able to do without station buildings, saves costs and prevents cluttering of the urban and rural environments, and it gives valuable flexibility, enabling the stops – and even the beamway lines – to be relocated. Station buildings cannot replace elevators, as elevators are valuable for emergency evacuations.

The elevator should be in the middle of the train, easily accessible for passengers from fore and aft. If full reconfiguration flexibility is needed, the elevator can have its own wagon, so that both the fore and aft wagon can be replaced (and dispatched to side lines) separately. A separate elevator wagon is also valuable for a minimum configuration, when a few passengers control the train movement by pressing buttons for selecting the destination. That is: The train is controlled like an ordinary house elevator.

But most trains will not need to send away the front wagon, so the elevator should generally be built into the rear end of the front (main) wagon.

Whether or not the elevator has its own wagon, it could be produced as a module, with the roof from which it all hangs. This roof could then (in a service facility) be inserted from back into a long, short or minimal (elevator) wagon.

The elevator to the left hangs in a single cable and is stabilized by four telescopic rails. If there were a cable in each rail, as shown to the right, very long cables could be used for evacuations at difficult places – e.g. when the train is 20 meters above the ground. The separated telescopic rails would then become reassembled when the elevator goes up again. (The lowest rail set – the outer tubes – would be four vertical cable channels in the elevator walls. The other tubes should all be hanging together from the roof, with decreasing diameter downwards.)

The elevator should be able to eject a little ramp from under its front door, as shown to the right, so that wheelchairs can be admitted from flat ground. (The external ramp depicted to the left can be regarded as a station area marker/placeholder.) The bottom of the elevator could also have a foot contacting the ground for stopping penduling. If side doors are not used in the elevators (as in the picture to the right), any remaining penduling will not be a threat for the passengers.

The stationary wall behind the elevator compartment should contain the elevator motor(s), as well as backup batteries ensuring evacuation capability.

The picture shows how the elevator ends have doors leading to fore and aft wagon. These doors should be as wide as possible, so that both bikes and wheelchairs could be rolled to parking positions in the wagon, beside the aisle. If there are windows in these doors, the train's conductor (seated at the elevator's back wall) has a good view around, being able to see passengers approaching – both in high and low elevator position. (S)he may then choose to start the train when the elevator starts going up.

The Electric Motor

The electric motor consists of two sets of magnets, one fixed set (the stator) and one moving set (the rotor).  At least one set must have electromagnets – able to adjust the direction and magnitude of its magnetism in such a way that the desired force is obtained.  The other set may be permanent magnets, electromagnets, unpowered coils magnetized inductively, or a ferromagnetic material.  A practical and efficient motor design is: Use a brushless design – with permanent magnets in the rotor and intelligently controlled electromagnets in the stator.  The strategy for powering a stator electromagnet is:

This analytical approach may seem artificial, but we really should be using computers (microprocessors) for controlling those quite powerful motors, and the above algorithm will determine the programming.  The commutator used in traditional motors is simply a "computer" for computing the magnetizing current (positive, negative or off) as a function of the rotor position.  Being a mechanical switch, it is bad for efficiency.  A computer, however, can use smarter strategies like providing magnetizing current pulses whose duration is determined by the motor power needed, and correct for power line voltage fluctuations.

So – what kind of voltage should we have on the power line?  The natural point of origin will be the 50 or 60 Hz AC available practically everywhere, and we can have it transformed to whatever voltage suits our trains.  A single power line (using the beam as the return path) will have an inconveniently low voltage for a significant time – around the time when the sinusoidal voltage curve crosses the zero line.  (The polarity of the voltage is unimportant – the electronics can easily reverse the polarity.)  Electronic equipment can generally remove these voltage lows by means of filter capacitors, but these will be impractical in high-power applications like trains.  If we use a three phase supply, we will have a decent voltage all the time, but will need two power lines in the beam.  If we, however, full wave rectify the three phase power, we have a quite well stabilized DC – with only 4.2% ripple (AC component).  The computer can now get a good voltage anytime.  It just sends the current through the electromagnet in the opposite direction if it needs to reverse the magnetism.

The problem with using DC is: When you rectify AC to DC, you need to know which voltage you will eventually be using.  With AC, you can easily use transformers later.  With a well-defined application field like train motors, there can't be much uncertainty, but it may be wise to specify a reasonable wide voltage range the trains must be ready to deal with. The small amount of AC needed by various equipment in the trains, can easily be made by chopping up the DC.

Another advantage with DC is: It is easier for the motors to send the power back to the power line during regenerative braking when they needn't check which voltage the line is ready to deal with at the moment.  (But if there is only one train supplied by a line rectifier, that rectifier station must deal with this problem when the train brakes regeneratively.) DC also doesn't pollute the environment with biohazardous electromagnetic fields the way AC does.

Power will be supplied to the line from many sources – including regeneratively braking trains. All this power should be pooled together, so that if there is a break along the line, power will still come from the opposite direction.

Hovering Trains

For high-speed trains, it may be advantageous to avoid wheel-rail contact altogether.  A new technology called maglev (from Magnetic Levitation) is being developed, and even used on some short stretches.  It uses electromagnetic forces to lift the train off the ground, so wheels, which are quite impractical at high speeds, are dispensed with.  Maglev needs a quite special track which is incompatible with ordinary railway tracks, and this is a serious impediment for its acceptance.  For the beamway, however, maglev operation may be compatible with our half-beam design, because:

But the problem with maglev is:  Its track must everywhere be able to exert a lifting magnetic force.  That means: It will be expensive and bulky.  The strength to weight ratio is critical for the beamway, so it can't afford to be weighted down by additional paraphernalia all the way along.  An extra light (and cheap) maglev track design has been introduced quite recently: Inductrack, whose magnets are simply unpowered coils in the track.  They become powered and magnetic by induction due to the magnetic field from the moving train.  The faster the train goes, the more efficient is the levitating mechanism.  This mechanism is really efficient: It can keep the train lifted, down to walking speeds.  This has an interesting consequence for us:  If we are content with using levitation only at higher speeds, the array of induction coils can be much lighter (and cheaper).  

This starts looking like another kind of levitation: windsurfing. 

Bogie with airfoils for windsurfing

If the train can catch the air in the beam, and press it down towards the rails by means of suitably curved airfoils, it can slide on a cushion of air – provided it can close well to the sides.  One advantage of this windsurfing is the progressive spring action: The pressure from the trapped air increases strongly if the surfboard – shaped like the bottom of the beam interior – comes too low.  This holds also for displacement to the sides, so the stability will be good, but this depends on how easily the air can move sideways under the airfoil.  If the air is compressed to half its volume over a 40 cm wide bottom, the 1 kg/cm² lift becomes 4000 kg for each meter of the beam, so only a fraction of the train will need this lifting surface.  If the geometry of the lifting airfoils is fixed, the lift-drag ratio can probably not be good over a large speed range, so wheels may have to be used up to the speed where they start getting problems.  But those wheels can't be small, so they will disturb the aerodynamics.  The bridging solution here may be to use small wheels in conjunction with either a variable airfoil geometry or Hovercraft-like compressed air cushion flight. But variable airfoil geometry doesn't simply mean: Put a hinge on the place where the airfoil starts rising. This is because the airfoils are also curved sideways to fit along the lower part of the C-shaped half-beam, and to provide stabilizing sideways pressures.

 If the transition between rolling and flying occurs at a high speed, there will be a problem during slowdown, when wheels which don't rotate, suddenly and brutally come in contact with the track.  It will be difficult to make all the wheels rotate with a reasonable speed before touchdown.  Rubber wheels should be better than steel wheels because:

The simplest solution anyway will be to let touchdown occur at a low speed.

Even if levitation can work at all speeds, some kind of bogie set will be needed in the beam for holding up the train when it is parked or powerless, and it might as well have some small wheels for low speeds.  The question is now: At what speed should the wheels be lifted from the track?  And should some sort of intermediate levitation mechanism be used before high-speed levitation is employed?  A maglev solution for moderate speeds could be electromagnets which pull the steel beam from below, but this may not be very energy efficient, due to power loss caused by induced eddy currents in the steel.

Those two levitation methods may very well be combined – as long as the extra equipment doesn't become too heavy.

It may be a good idea to do the levitation only partially, e.g. by letting windsurfing airfoils relieve the wheels of only half the weight. But the sideways acting airfoils can still have full stabilizing effect, so that the wheels may not be subjected to any forces from the sides. The reduced strain on the wheels should enable significantly higher speeds. The reduced high-speed traction should not be a problem, as hill-climbing is unlikely to be required at full speed. The amount of regenerative braking needed for normal operation will still be available, and for emergency braking, brake pads can still efficiently pinch the beam as described under The Train above. The beamway is unique in its ability to use partial levitation, because of its immunity against derailing and side slipping. The partially levitating bogie (possibly supplemented with a little aerodynamic lift from the train itself) is likely to be the most important bogie design for all applications except the very highest speeds.

Compressed Air Hovering and Propulsion

Another alternative is the compressor-driven train. The compressor could be like the (free-lance) locomotive described above, but it need not exert a pulling force upon the train. Instead, it delivers compressed air to all the air cushion bogies. In these bogies, the wheels are replaced by air cushion units - lids covering the channel (with the track) in the bottom of each half-beam. The front and back ends of this lid should have valves which could be swung down to almost close the channel, or swung up to let the air escape – in the direction from which the train will be pushed by rocket action. If the closed valve could receive 20% of the total pressure force, the train could climb grades approaching 20%, or accelerate at .2 g at low speeds, or reach a top speed for which the drag is 20% of the train weight. Several air cushion units may be required in each bogie for obtaining sufficient pulling force.

The two next pictures show (with a design not made aerodynamic) how the lower part (“wheel groove”) of a half-beam can be sealed off quite well by a metal lid on top (shown in blue) and valves fore and aft (shown in yellow). The valve in the foreground is shown opened, and the valve in the background is closed. Opening/closing is done by electric motors or pneumatic actuators (in the violet cylinder on top of the lid). The thick tube supplies air to the two or three compartments under the lid.

An air cushion unit of type 1: Thick air cushion (with two compartments) under the (blue) lid

An air cushion unit of type 2: Thin air cushion (with three compartments) under the lid (very thick, blue – partly concealed by an opened valve)

Type 1 provides simple hovering and position control. The two compartments provide forces with the directions / and \. The lifting force is the sum of the two vertical components. This is quite simple to control: Just keep the airflow strong enough. If it is very strong, there will be wide openings around the edges of the raised lid. (And unexpected vertical forces are rare.)

Controlling the sideways position is not so easy. It can be monitored with position sensors, and then adjusted by altering the air distribution to the two compartments. Or windsurfing could be employed against vertical surfaces, using supplementary vanes.

But a safer and more elegant method is to utilize the progressive spring action of an air cushion: If air without an escape route is forced to compress, it gives a progressively stronger counterforce.

The type 2 cushion is thin, so that the progressive counterforce comes before the position error becomes too large (with direct metal contact). Also this design could use two compartments, but is here depicted with three, as the added control might be valuable.

The different compartments give different kinds of force, end these can be paired in antagonistic pairs: towards left vs. towards right; rotating clockwise vs. rotating counterclockwise. Airstreams to antagonistic compartments can be connected with a symmetrical suction valve so that if an airstream A is reduced to zero while its antagonist B is strong, the (Venturi) force from B gives suction – perhaps approaching vacuum – in A, thus augmenting e.g. the sideways positioning. If a strong wind gust from the right hits the train, the train should try to resist it in order to prevent an annoying train swinging. A train with wheels could only stiffen its suspension until the left wheels almost were lifted. The air cushion train, however, might increase the lifting pressure on the right side and turn off the airflow to the left side. With an antagonistic interconnection, suction would be generated on the left side, to perhaps double the stabilizing resistance (torque). The train should have wind pressure sensors on the sides, so that wind swinging can be anticipated.

Power loss during flight should rather not lead to a crash, so there should be some small emergency landing wheels or rollers. The vertical compartment walls should contain thin steel wheels protruding a few millimeters, and so should the outer edge of horizontal lid plate. But such small wheels are only suitable for low speeds, so if power is lost at high speed, a pressure tank or backup battery should at least enable some emergency actions like opening the fore valve and closing the aft valve – for giving a little soaring (airsurfing) capability. This emergency lift capability could be increased if each air cushion compartment had a little plasma generating spark plug for generating extra lifting force from the local backup batteries. Such a spark-generated force pulse may also be needed to compensate for unexpected forces, like from a wind gust. (Or a little fuel may be injected and ignited, so that jet or rocket propulsion is obtained. But this is possible only to a small extent, as the beam interior – especially with plastic components – is likely to be damaged by the heat.)

But the pressure tank should preferably have so large capacity that it after a power loss enables hovering during a not too brutal train slowdown. If the compressor delivers air at e.g. 3 atm in the long delivery tube, and the average air pressure in the cushions is (increased by) .5 atm, the air volume expands about six times. A fast train is likely to have the tube built into the roof (instead of the cabin lowering mechanism), so by having two parallel tubes, the capacity is doubled. And/or the compressor loc could have a 300 atm air tank with a matched compressor stage. Its volume would be about 100 times more efficient compared to the tube.

The compressor drive is unique with respect to its capacity for storing energy (in pressure tanks). If a compressor can run at reduced power, its energy efficiency will be better – because the air will then be heated less during compression.

A compressor “locomotive” (left), and one of the hovering bogies powered by the high pressure air tube. The compressor hangs under two such bogies.
(This loc could move by itself, and even provide some maneuvering train moving even if the air tube is disconnected. It could have some pneumatically driven, pinching rubber wheels, engaged in an emergency – for submerged tunnel use.)

The proposed bogie shown above consists of four air cushion units mounted at the ends of the strong X-shaped air distribution tube/beam. Each arm of the X should have two or three ducts for air to the different compartments, distributed under control by the bogie's central processor/regulator.

The X should have some scissor-like adjustment capability, so these bogies tolerate some half-beam separating movement. The thin rods going along the X arms enable opposite air cushion units to be parallel displaced together – by a parallelogram movement.

This bogie may be designed so that it could be converted into a wheeled bogie – for use when there is much hill-climbing and slow movement. Each air cushion unit will then be replaced by a cylinder with a sideways-directed axis. This is an air turbine motor. The wheel fits on the outside of it – like a car wheel outside its brake. To the free end surface of the cylinder (facing out of the wheel) could be attached an airfoil unit giving partial lift as well as sideways stabilization – away from the side walls of the beam. This breaks the symmetry of the bogie, as the airfoil is adapted to running in one direction.

Another kind of wheeled operation: The cylinder contains an electric motor instead of the air turbine motor. This should give maximum energy efficiency, but less power and speed. The compressor will be running on low power, but should then be more efficient due to lower air heating during compression.

Only the tops of the four units are common to wheel use and air cushion use. These tops will ensure correct alignment along the beam for whatever is attached below them – thanks to the X arm parallelogram movement. When also wheeled bogies are able to adapt to varying half-beam separation, the beam need not be designed for constant half-beam separation, as it will not matter if half-beams are bent apart when loaded.

The compressor should be efficient, silent and with low wear, so it is likely to be of a rotary type – perhaps the scroll type. The air inlet should not be from an end of the nacelle, but around the middle part. In this way the compressor, like the depicted hovering bogies, will be symmetrical, able to run equally well in both directions, so that a compressor locomotive can be used before and/or after a train. If the air intake is through a grille in the middle part of the nacelle, not much debris and birds will be sucked in. The compression (in atmospheres) should be high for getting low air resistance loss in the tubes (and for getting much pneumatic emergency energy stored in the tube), and low for avoiding the energy loss caused by (nearly adiabatic) compression heating. The compression should be variable. The compressor might also have a tank for compressed air – so that additional pneumatic emergency energy can be stored (for a safe slowdown in case of loss of electrical power).

The compressed air from the “locomotive” goes out through the long horizontal tube (which can flex a little in track turns) and up to each bogie at or in its (hollow) suspension rod. The compressor might be something like 20 meters from the train itself. (There could be an additional compressor loc at the other end of the train, connected to the other end of the long air tube.) This separation from the train has important advantages:

But passenger evacuation through end doors of the train will now be impeded.

The long horizontal tube will actually be two tubes – one on each side of the center line – and these will be integrated in the roof of streamlined high-speed train. (These high-speed trains will not have exchangeable cabins, and the elevator will be integrated in the rear end of the first wagon.)

The train may occasionally need some extra pulling power, so it may be wise to have air outlets some places at the train, on any surface except the front. These outlets should be covered by flaps going flush with the train surface, and hinged in the front. When compressed air is sent out the outlet, it will press the flap open, which will then provide a surface for the expanding air to push (with rocket reaction).

Such flaps opening in the other direction, to catch the wind, could be used as air brakes, and if the air pressure here became strong enough to overcome the internal tube pressure, there would be regenerative braking. Its efficiency would probably not be great, but regenerative braking would not be so important on long-distance beamway lines.

Each of a bogie's four air cushion units has a footprint of 20x60 cm, giving a lift area of 4800 cm² for the whole bogie. If a bogie is to carry three tons, the pressure in its cushions need only be 3000/4800 = .63 atm. An advantage with using so low pressures is that (wind gust) forces threatening with lifting one side of the bogie can be efficiently counteracted by a (local vacuum tank and/or) suction device providing down towards -1 atm. Normal wheels (running on a track below them) are unable to pull downwards, but these air cushion units can pull both downwards and upwards.

This example, with three tons on one bogie, assumes even weight distribution. But use of proactive suspension for avoiding loading the middle part of the beam, implies putting more force at the poles, more than tripling the air pressure there, to 2 atm or more, but this is still a moderate value. Low pressures, for suction, should still be applied also at the middle of the beam, and for this purpose it is valuable for air cushion units to go down there. (If a sagging beam has just been turned upside down, the lowest part of the beam will be at poles, so it is really important for air cushion units to go down and apply pressure there.)

Intelligent air flow control enables two beamways to cross in the same plane. It is easier to reroute air flow when the bogie has to cross a track gap than it is for wheels to negotiate such a gap. It is also easier to provide extra lifting force to the bogies which are near a pole – assuming the train now has a strong spine.

Beam joints are difficult to make good enough for wheel traffic if the noise level and wheel wear are to be acceptable. For air cushion hovering, the gap can be many centimeters, and a simple gasket in the gap will be fine. It may be wise to skip wheel trains completely for rural beamway lines – except for steep grades and submerged floating tunnels, where sqeezing rubber wheels may occasionally be activated.

No Wheels at all?

If it were decided that the trains could dispense with wheels, and use only air cushion gliding, a completely different, and more efficient beam design could be used. The problem with wheels is that they are very large: A train needs a wheel diameter of at least half a meter. This means they have to be kept vertically like a car's wheels, and hence be paired, with a stabilizing axle between the two. If the wheels were running in-line like bike wheels, and tilt to the sides up to perhaps 45° (like bike wheels), the beam would have to be very wide and bulky, and hence very inefficient as a load carrier. But air cushion units can be very small, and able to tilt sideways in a quite small groove.

The S-beam for air cushion gliding – with the air cushion units running in-line in the bottom of a single-profile beam.
They swing freely 35-40° to each side when the train swings outwards (banks) while going through a curve.
Compressed air emerges from slits in the bottom of the track – and perhaps from slits on the sides, particularly if control of side forces turns out to be needed. There may be small steel rollers, lifted by air pressure when not needed, before and after the air cushion units. These will be useful for (emergency) maneuvering at speeds up to e.g. 30 km/h.
(The air-conducting “suspension rod” seems flimsy, but should be long, e.g. half a meter along the track.)
The train should also have rising arms which are held in fixed position relative to the beam. They should contain brushes (small pantographs) for contacting the power line (depicted orange). And they could exert anti-swing force (to prevent wind swing) upon either the bottom fin of the beam or above the power line holding isolator.

The S-beam will have its steel extending vertically rather than having the 2C-beam's side-by-side duplication, which gives less load-carrying strength. A single beam needn't be twice as high, but rather 40-50% higher. It may be possible to strengthen such a beam with embedded, diagonally running, carbon fibers, preferably nanotubes.

It should be quite easy to shape this beam so that it doesn't bend (twist) to a side when loaded. And if it bends a little to a side, it is rather unimportant, as proper space for an axle length is not to be reserved.

In tunnels, where the distance between beam suspension points can be small, only the lower 1/3 of the beam may be needed. (High stabilizing arms should then be able to disengage and lower themselves.) The small crossectional area of such a beam makes it easy to put it into a slit in the ceiling of an old road tunnel, in which the train can go like a bus. And this narrow plate can have expansion mechanisms which affixes it in a slit.

If the train has its own tunnel, a double-track tunnel need be little wider than a single-track tunnel, as the total width of the two adjacent beams is very small. Such narrow track parts are called gauntlets. Trains must be scheduled to meet only outside the gauntlets.

Special beams, with track grooves on both sides, may be produced for this purpose, and for bidirectional traffic on a single line. Beams hanging in bridge spans will be another application for such gauntlet track parts.

It will be simpler to make this beam as a plate with a gutter-like track channel attached. Such a plate may be made from reinforced concrete. But it may now be difficult to make the channel contribute to beam strength rather than being dead weight. Weak splices in the beam are avoided if the plate and the gutter have their splices at different places.

If the beam is a vertical concrete plate, it starts looking like the “People Cargo Mover”, but if it is unable to use metal beams held by metal poles/racks, it may be too heavy and awkward for crossing awkward terrain. And without an elevator, the safety in emergency situations is unlikely to be acceptable. This transport concept is well adapted to high traffic volumes, having a smart mechanism for lowering the stopping trains/wagons out of the way at stations, but the concept is less down-scalable and flexible for lower population densities.

The S-beam may be vertical all the way, even in turns, as a suspended train will automatically get the correct banking angle in turns. This will make such a beamway simpler to build. If a centrifugal acceleration of .5 g is permitted, the train will swing 30° to the side. When wind swing and other irregularities are added, there could be danger of derailing for outward swinging – away from the beam wall, so an anti-swing mechanism or some other bracket going to the back of the wall will be needed for derail prevention. The beam may be mounted non-vertically – for giving a bank angle – in turns. It will be elegant, but difficult, to twist the beam at both ends of the turn – for providing beam continuity, but it should be ok to aim for continuity only down at the track, and accept sudden angle changes for the wall parts of adjacent beams. The mechanism for exerting anti-swing force should just be prepared for discontinuities in the surface it operates on.

A Universal Bogie

This chapter shows variation possibilities which may be bewildering. The preferred design uses a different suspension, as described in A Compromise Bogie below.

The high-speed convertible bogie can take four air cushion units or four wheels containing electric or compressed air motors (with a central stator).
This picture shows both the bogie (with some modules attached), its suspensions and parts inside the roof of the wagon.
The modules attached here – an unrealistic combination – are:
The right side of the bogie: A wheel with a motor, and an air cushion unit
The left side: A motor lacking a wheel, and (violet) a unit for partial levitation (for use with wheels: full sideways stabilization, and about 50% wheel load reduction)

The white arc is one of the two rails along which the suspension can move sideways. The nearest of these rails is not shown, so that the small wheels rolling on it are visible. The minibogies with these small wheels – one on each end of the (green) suspension box – can be displaced passively or actively along the curved rails, and they can be rotated – for adjusting the sideways tilt of the train. Passive displacement will occur when the train is going through a curve, and letting the train be displaced to the side by the wind may be suitable for keeping the train vertical in a side wind. If not, some active displacement will be needed. These rails follow the curvature of the roof, so that the passive displacement is self-centering. It is important to compute the roof curvature accurately, so that these self-centering forces are optimal.

Each side of the universal bogie has a separate L-shaped suspension plate. These suspension plates can bend apart a little, and thanks to this movement the two independent sides of the bogie can adapt to varying half-beam separation.

Compressed air comes from the tube going along the train, and it continues through a flexible and/or telescoping tube (not shown) to the suspension box (green), which controls the air supply to the bogie. It may be necessary to measure and actively control the air cushion levitation for the bogie's four corners separately, so the suspension box will then send out these four air streams, plus air streams for right and left side path stabilizing air cushions, plus a high-pressure air stream for extra lifting (compressed by a small compressor in the suspension box). These air streams (and electric cables) go through ducts inside the suspension plates.

The (electrical) power collector should be centered between the two suspension plates. Perhaps springs to both sides may give adequate centering.

The air cushion unit used is the type with varying height. Its outer box is plugged into the end of the suspension plate. The rest of the unit can be pushed down by extra-compressed air let into the top of the outer box, then pressing down the telescoping inner piston-like box and telescoping air supply tubes. These tubes send cushion air for lift, left displacement and right displacement.

Such telescopic movements are prone to jam when the width is large compare to the length (in the movement direction), so it may be safer to replace linear telescoping movement with a corresponding circle arc (hinged) movement.

Air cushion drive may do without springs for softening the ride, but if wheels are used, springs and shock absorbers should probably be used. This could be where the suspension plate is attached down in the suspension box, but, for minimizing the unsprung mass, it should be near the wheels. A wheel module should then be a rectangular frame with spring and shock absorber before and after the wheel. A wheel can have a soft part (rubber, plastic) around the rim, inside the rim, or no soft part at all. If the soft part is air-filled, its air pressure can be maintained by the supplied compressed air, through a hollow axle.

A rubber wheel module might have provision for attaching a supplementary rubber wheel module below it. This lower wheel can be engaged to run (air pressure driven) on the underside of the beam when hills are to be climbed.

Air cushion drives may be made so small that they can be retracted out through the bottom of the beam. A small vehicle for inspection (and delivering urgent parcels) could have arms which can grab the upper part of the beam. If it then retracts its air cushion drives, it can swing itself to a position above the beam and be out of the way for passing trains.

Having separately operating wheels and/or air cushion units on the two sides, enables designing a bogie which can hang, and perhaps move, on one half-beam if the other half-beam collapses.

Double Action Air Cushion

Air cushion operation can be more efficient if the beam is designed for double action air cushion hovering. Double action implies having air cushion units both over and under the horizontal steel surface of the beam. The upper air cushion lifts in the familiar manner, while the lower air cushion, having a low pressure approaching a vacuum, "sucks up" (letting the atmospheric pressure lift from below). One air pump can serve both cushions, pumping from the low pressure cushion to the high pressure cushion. Or the low pressure may be obtained from a high-pressure air stream by means of the Venturi effect. Valves should be able to redirect the high-pressure stream to the lower air cushions (and the low-pressure stream to the upper) momentarily when needed for counteracting a sudden (wind gust) swing.

This gives several advantages:

The upper air cushion unit also acts on the side walls – because a high-pressure cushion gives progressively higher force when compressed. This is not the case for the low-pressure air cushion unit below the beam, so it only acts on the beam surface above it. (Also air surfing vanes may give sideways stabilization).
The suction below should be carefully controlled to prevent mechanical contact.
If the lower air cushion unit has a ferromagnetic (steel) body, a coil (electromagnet) could be wound around it – for giving some maglev lift.

The jet action obtained by opening ends of air cushion units may not be enough for giving high speed, so these minilocs should be combined with the linear motor described below. (Somewhat larger motors and compressors can be placed on top of the wagons.) It is still valuable to have this jet action, as trains can then go (with reduced speed) where linear motors are not mounted. When linear motors are used, it may not be wise to try to use jet action, as the wrong air pressures will now be imposed upon the air cushions.

Air cushion units below the track could be replaced by electromagnets – for a partial maglev. It will then be very easy to control the separation and lifting force, but the force between an electromagnet and iron can not be reversed.

There could be Inductrack coils in the bottom of the beam if those brutal wheels are not used there. These coils may not be needed for maglev hovering when compressed air is used for this, but they may still be valuable for propulsion, as a linear motor. In this case, they may not (as explained in the next chapter) be needed all the way along the track, as it is probably ok to pull only a part of a train. They may also be omitted at the middle half of each beam, and at stretches where acceleration or high speed isn't needed.

Inductrack coils for merely propulsion could probably go partly up the side walls of the track, and not only lie on the bottom. This should increase the pulling power. If a train needs to be pulled on some places (after a station, up a hill), Inductrack could be used reversed – with those light coils on the train, and electromagnets in the track. Those coils are light and cheap enough to be used along the entire train, and then it is enough to have electromagnets near poles. But now we are dealing with linear motor – to be discussed in the next chapter.

If wheels are not used, beams with perhaps only 1/3 of the normal height might be used in tunnels. It may be room for these in the ceiling of an unmodified road tunnel, in which a train may move like a bus.

The Linear Motor

Levitated movement implies: The wheels lose contact with the track, so conventional wheel-driving motors will not work anymore.  If we use a linear electric motor instead, the lost mechanical connection between the train and the track is replaced by magnetic forces, and these forces work just like in a conventional motor: between stationary and moving magnets.  The difference is: In the conventional motor, the movement is rotational; in the linear motor, it is straight like the track, so the moving magnets (corresponding to the rotor) might as well be fixed on the train, while the stationary ones (constituting the stator) can be fixed along the track.

Having magnets or coils all the way along the track, will be expensive – and add very inconvenient weight on the beams, so we will try to reduce the size of the apparatus.  We may assume that the linear motor will only be used for real trains, and not for local trams.  We may also assume the train will not be shorter than the distance between the poles (40 meters).  We may then place stators (propulsion modules) at each pole, and a part of the train will always be adjacent to a stator which can pull (or brake) it.  If each wagon of the train has a tube containing permanent magnets over its roof, the electromagnets in an e.g. 5 meter long stator set on the nearest pole will always have at least four meters of the train magnets under their influence.  The pulling force should be about 10000 Newton ("1 ton") per meter for a module pair.  At places where hill climbing or strong acceleration is required, longer stators, giving more pulling force, can be mounted.

The sausages over this train are real hot dogs, containing (4-5?) powerful permanent magnets.

The magnet sausages can be replaced by flat (vertical) plates containing passive induction coils (Inductrack), or simply aluminum plates in which currents may be induced. They will be lighter and cheaper than the real hot dogs, but may not give enough pulling force, and they will require different stator modules (with a slot-formed gap). If the induction coils are embedded in the wall of a tube for compressed air, we can use our sausage configuration – perhaps with a rectangular cross section. But, unlike sausages, the air tube must be continuous, so it must either let the suspension rods go up in the middle of it, or be mounted 20-30 cm to the side of the suspension rods. (But not be mounted on the roof, which can move to the sides relative to the suspension rods)

Inductrack coils may also go around the edge of each air cushion unit. This would require a special beam or transition area having electromagnets in the track.

The stator modules can be mass produced in a factory, and easily deployed from a passing service train.  Each module is then pushed into a bracket at the beam join at the pole (the lower middle part of the splint), and one module is mounted on each side, so that the chain of train magnets pass closely between the two.  

Here we see how the magnets are passing between two stator modules.

Each stator module contains:

The electromagnets nearest the ends (or Hall effect sensors) will sense the approach of train magnets, and the module will by a radio signal receive a request for a certain pulling (or braking) impulse from the train.  The speed of the train is evaluated and compared with the permitted speed (as given by radio from the central computer).  The electromagnets may then be ordered to push and pull the train magnets forward, or to slow them down if this had been requested by the train – or if the train speed is too high.  The braking is regenerative: Regained energy is returned to the power line.

It may be advantageous to avoid all mechanical contact between the train and the track, so also the brush getting power from the power line should be disengaged.  Power needed in the train may then be fetched inductively by coils in the magnet tube.

The pulling (or braking) impulses affect the train quite unevenly, so some force smoothing will be required.  The train magnets should be positioned between springs in their tube, and there is not much mechanical rigidity on the way from the magnets and down the suspension to the passengers.  The few wagons in a beam train will be more tightly connected than in an ordinary train, so the uneven force from the linear motor should not be a problem.

When axle-connected wheels are not used, the wheels or air cushion units needn't be paired, but may follow after each other. The suspension plates (of e.g. the universal bogie) – one for the left side and one for the right side – needn't be paired, but these two and the current conductor can follow each other in-line. This will give the whole suspension the narrowness which is advantageous for use with the linear motor described here. (The train magnets may be attached to either of the suspension plates, as both the leftside and rightside plate will be in proper position at poles, where stator modules are – even it the half-beams are somewhat twisted out in the middle.)

The various magnetic solutions mentioned here may be replaced with electric equivalents. But these require high voltages which are dangerous for people, and the voltages may cause sparks. It may be assumed people will not come near the locations in question – in or near the beam. And magnets can be replaced by much lighter conductors, so linear electrostatic motors are certainly worth considering.

The Suspension

Hanging below the bogies – which may or may not be rolling on wheels – is the suspension, which ensures that the passengers move more smoothly than the bogies do.

But first some words about the use of the underside of the beam, and this underside must, as we now will see, be smooth.

Braking should mainly be done regeneratively by the motors, that is: The kinetic energy of the train should be converted to electric energy which is fed back into the power line instead of being wasted on wearing down brakes.  When the connection with a power absorbing grid/batteries/resistors is lost, and during emergency braking, old fashioned friction brakes are needed.  The beamway has the ideal solution for this: Let the brake pads pinch the lower part of the beam! (Or two brake pads could be pushed apart against the interior beam walls.)  This is far more efficient than the conventional solution: slowing down the wheels. And overheating should not be a problem, as the beam has a great heat dissipation capability.

Here we see the brake pads of the beam train.
They brake much more efficiently by pinching the beam.
And the beam carries the heat away rapidly.

This strong braking puts extra strain on the suspension. Acceleration and regenerative braking may give about ±0.2 G horizontal acceleration, while friction braking should be dimensioned for 1 G. While braking with 1 G, the force from the brakes goes down and forward at a 45° angle. It could go through strong straps attached at the lower end of the suspension rod, where it acts directly on the roof arc, and may block the movement of the wheels rolling there. But if straps tend to vibrate like a wind harp, rods might have to be used. It may, however, be enough to let the straps be in contact with the nacelles they pass.

This supplementary suspension should also act as a regular vertical suspension, able to hold the train if a regular suspension rod fails or needs help with lifting. During a 1 G braking, the resultant force in the supplementary suspension will increase by 41 %, and its length (if it originally were vertical) should increase by the same percentage. The brakes should slide backwards from the bogie, through a distance corresponding to the height of the suspension. In order to do this, they should be connected to the bogie with telescopic rods, heavily springed. The supplementary suspension needn't be fully vertical initially, but start from perhaps half the backward movement. The regular suspension rods should be articulated in both ends (for swinging backwards), so that they don't receive orthogonal forces.

An additional use for these straps is: If the bogies crash completely in a destroyed beam, the suspension rods will be broken loose, but the straps could give a more even (but brutal) braking if they had several meters spare length to slip along in e.g. the lower end.

Rubber wheels can run on the underside of the beam.  They may be just millimeters away from contact, to ensure against derailing.  Or they may be engaged to squeeze during hill-climbing.  If powered rubber wheels are used both above and below the track, the hill-climbing ability could really be great. But those supplementary wheels will not be needed if there are motors in the main wheels.

Now, down to the real suspension system at the top of the wagons.  A wagon should be able to swing to both sides like a pendulum, but in a controlled manner.  The axis for this movement should ideally be near the bogie, but then it would be difficult to control the swing.  Besides, magnets for the linear motor drive could not be mounted on anything swinging under the beam, as they must pass precisely between the stator modules.  Placing the pendulum axis in the roof of the wagon, where the suspension control mechanism is, enables the suspension control to control the swing of the wagon optimally.  It may be enough simply to control the dampening of the swing, and let gravity do the rest.  In that case, we could let the swing move a piston in an air-filled cylinder, and just choose the size of the opening the air in the cylinder can leak through.

Each of the suspension rods, from which the wagon hangs under the bogies, should be able to move sideways across the roof, both passively and actively.  They must be displaced passively when a wagon with more than two suspension rods goes through a curve, and they should be shifted actively to the side e.g. to compensate for a side wind or uneven sideways weight distribution in the train.

Proactive Suspension

Vehicles normally have reactive suspension: By yielding for sudden jerks, they can smoothen the ride.  But the beam train moves in a simple and thoroughly mapped environment.  It identifies every pole it passes, and from its database it retrieves some parameters about the next beam:

This enables the train – like a downhill skier – to assume the best body position and degree of dampening. The train can compute how much the beam will bend in the middle – determined by factors like wagon weight, train speed, and bending caused by the wagons ahead. It can then compute how much the wagon should be lifted up at the middle of the beam.

If the wagons are rigidly connected with respect to vertical displacement, the proactive suspension mechanism can be used for determining which part of the beam shall get the task of carrying the train weight, and that generally means the train should pull itself up primarily by the bogies which are closest to a pole. In other words: try to achieve the above mentioned lifting at mid beam not by really exerting force there, but by exerting the force well before and after mid beam. This should not be difficult to do if there are strain gauges measuring the load at each suspension.

In connection with the linear motor, we found that we could move the whole train by exerting force only from certain parts of the track, and we save both equipment money and beam loading this way.

Now we can use the same principle: dynamically applying forces to parts, but vertically instead of horizontally, and with the object of reducing beam loading. Distributing the force (or momentum, as the train mass inertia permits some rapid force variation) through the train was simple in the horizontal case: The train simply hangs together. (And the magnets can slide, with spring damping, in their sausages.) But in the vertical case, we find that selecting the right train stiffness/bending is more tricky. The most important point is: Don't apply so much stiffness (or upward bend in the middle) that wheels in the middle are lifted from the track when the train passes the bottom of a vertical curve.

Transversal Force Control

The transversal forces are those acting perpendicularly to the direction of movement, i.e. vertically and/or sideways. The most important of these, the gravity, need not be discussed further here. The two next to be considered are: the wind and centrifugal forces. For both of these, there will be a linear lateral force (directed sideways) and a rotational force (threatening with lifting the wheels on one side), and there are different solutions for the steady and momentaneous components.

The sentrifugal forces are simply determined by the speed (and the load of passengers/cargo). The steady component is here the sentrifugal force at the normal speed. The track is of course banked, so that the sentrifugal force at normal speed is experienced as increased gravity, for which the track and the train must be dimensioned. The momentaneous component will now be experienced when the speed is different from the usual speed, for which the track is banked, and lateral forces will now act. These will just cause the train to hang at a different sideways angle, so that the passengers experience no lateral forces – only what feels best for them: seemingly a varying gravity.

The crosswind forces will originally be linear lateral forces, but these will, due to the pendular nature of the train, be partly transformed into a rotational force (roll torque). The steady component is determined by the mean wind speed (averaged over several seconds), and the momentaneous component is due to wind gusts.

The principles for transversal force control will then be:
The steady rotational forces will be counteracted by means of a sideways displacement in the train suspension: When the train lets itself be pushed/blown to the side, the center of gravity is no longer below the suspension points. This creates a (gravity) torque which can counteract the imposed (wind) torque, consequently greatly reducing wind swings. But the momentaneous wind forces are too rapid and unpredictable to be controlled in this way. They should be resisted by stiffening the suspension, but if the force becomes so strong that wheels on one side are about to be lifted, additional resisting roll torque will be needed. With compressed-air hovering, the ability for suction (reducing the air pressure to below one atm) can give a higher roll resisting torque.
This torque will be still higher if the Double action air cushion principle is used.

It can be difficult for the train to distinguish between the swing caused by the wind (which should be counteracted) and the centrifugal swing (which should be permitted). There should therefore be an XY-accelerometer below the pivot, and it should check that the acceleration vector goes orthogonally towards the floor.

Another stabilizing mechanism is available, at least at high speeds: Let the train have ailerons (rudders) on the bottom. It is easy for a computer to control them to counteract swings, especially if the computer is connected to wind sensors which detect gusts from the sides.

If the imposed roll torque is still too high, the force control system must give in and allow the train to swing to the side. This will be an annoyance comparable to the turbulence annoyances during flights. If this happens often, it may (for low-speeds, at which stabilizing ailerons are not so efficient) be necessary to use a side rail at the bottom of the train.  Support wheels there could run horizontally on the side of the side rail (which can be shared by two adjacent beamway tracks), and the suspension rods should be shifted towards this side, so that the bottom of the train swings towards this side.  Or the support wheels could be vertical, running in a groove on top of the side rail.  The suspension rods should then be shifted away from this side, so that some of the train weight rests on the side rail.  (This variant will be more disturbed by snow in the support track).

Air cushion trains can replace support wheels with air cushion units, and may not need to displace the train sideways, as these units can both blow and suck. But these high-speed trains are less likely to need such side support, due to the the strong stabilizing effect of ailerons and suction-enabled compressed air cushion drive. Besides, a high-speed line is more likely to be moved down into a tunnel on windy stretches.

The strength of the beamway will be significantly improved if there is a double track with a T-shaped support bracket midways between the poles.  The pole separation can then be significantly increased.

No matter which support method is used: Where side support is not needed, the train can simply center its suspension and optionally retract its side wheels. Exchangeable cabins should not have to be equipped with such wheels, but it may be enough to have these wheels at the front and rear end of the center wagon, and at the locomotives (shown at each end of the brown train).

If maglev propulsion is used, it should be possible to make maglev side support rails, but the side support should be less necessary in conjunction with maglev, because the magnetism can simply be reversed for pulling down on the side where the bogie is about to get lifted.

The (non-rotational) lateral forces at the wheels will, if they are small, be counteracted by means of a curved track profile: Lateral forces cause the wheels to go higher up on the side of the track, and the gravity will then give a progressively increasing corrective lateral force. Flanged steel wheels will be ensured strong corrective forces from their flanges. Rubber wheels fit well in a curved track, but may not get sufficient track width, so the proposed rubber wheel depicted here has steel sides which can act as firmly delimiting flanges. These forces will become rather brutal at high speeds, but for high speeds with wheels, partial levitation should be used. This method uses airsurfing against vertical beam surfaces for obtaining the corrective forces in a soft manner.

A special kind of transversal force control is: Prevent the train from penduling when the elevator is going down – to position the elevator at the edge of the (very low) passenger platform. The active suspension might do some extra stabilization now, but the main trick is: The elevator has a foot which reaches the ground in a braking manner while the elevator bottom is still some toe-heights above the platform.

Spine Trains

We may now for a moment digress to alternative train designs which are somewhat different from our main recommendation, and take a closer look at the vertical chassis design mentioned under “Are Strong Beams needed?” above.

The first one is compatible with the 2C-beam design, but if trains can run unmanned, there is probably no point in connecting these cargo wagons to the passenger wagons/cabins.

Three vertical chassis cargo wagons.  They are connected with hinges in all four corners. The upper right hinge can pull and push.
Adding empty wagons will improve the load distribution.
One or more cargo containers can be replaced with a motor pack which connects to the nearest bogie – no locomotive needed.
The train can put a container down upon a shopowner's pickup – no station needed.
This beamway is reinforced with (carbon fiber) wires, although the strong spine should reduce the need for beam strength.

Suitable wagons, for giving the required spine strength in conjunction with a partly flexible movable beam, would be the vertical chassis wagon for cargo and the exoskeleton type for passengers. A separate article discusses a next generation beamless beamway.

The Cabins

A wagon = a set of bogies + a suspension unit + a cabin.  In a conventional train, this sequence goes upwards, but in the beamway train it goes downwards.  And normally, the suspension unit is not noticed as a separate unit, but when the passengers are in a separable cabin, things are different.  A wagon can have its suspension unit integrated in the roof, or it can be separable. If a bogie is motorized, it (and its suspension unit) can be regarded as a microlocomotive which may be programmed to run independently to another train needing it.

When the beamway can deliver its separable cabin by lowering it down upon the surface of the receiving vehicle, it interfaces nicely with various other means of transportation like boats, conventional trains and buses. They hold cabins from below – by simply letting cabins stand upon a surface, so a cabin can be transferred to and from the beamway with a very short transfer movement – a few centimeters.

Flying contraptions like planes and blimps are not so suitable for drop-down delivery, but the beamway is able to come close enough to them to facilitate simple cabin delivery horizontally.

This article is not the place for envisioning the potentials of such interaction, just the technical details.

First of all: The cabins must have light weight.  Forget about conventional train design.  Steel should be avoided.  We are talking fiber-reinforced plastic (FRP), or at least aluminum.

We will consider two ways of grabbing the cabin.  

The most obvious, and seemingly most elegant solution, is to make the roof strong and double.  The outer shell of the roof has a slit along the middle, from front to back.  It can be lifted by means of a steel rod with something like a row of small propellers which, when turned to point along the rod, can be inserted into the slit.  The "propellers" are then turned 90°,  and the cabin can be lifted.

The other method requires holes on the upper edge of the wall.  (Visible in this picture as black rectangles between the upper window corners.)  A large frame is lowered down upon the roof, and its edges go down to cover the upper part of the wall.  One steel claw will now be aligned at each hole, and together they get a secure grip.

If the row of "propellers" misses its slot, or if the slot is covered by snow, ice or dust, the first solution will fail.  Besides, the roof must be very stiff and strong for this method to work.  If there are two slots (and two rows of "propellers") near the walls, the forces would be more correct, as they would pull along the walls rather then bending the roof.  But the above mentioned problems will remain.

The frame, however, will more easily find its position around the roof, and holes in a vertical surface are less likely to be obstructed by e.g. snow.  Besides, the frame fits nicely around the suspension unit – a connection unit hanging under the suspension unit would not control the movements of the cabin so forcefully and precisely.  That is why the pictures here are favoring this method.  

The sides of the frame should be as narrow as possible, as this width will reduce the sideways travel of each suspension, and hence the curve negotiation capability.

But the suspension unit can also become flush with the wagon as shown here.

A suspension able to move the wagon e.g. ± 10 cm will suffice for placing an ordinary wagon upon a properly positioned immobile platform. This will come handy at a service platform, as the suspension can easily and safely be disconnected from a wagon that is held from below. Reconnecting a series of displaceable suspensions, however, will require manual work by service personnel. But a normally-looking wagon can be rapidly delivered to and fetched from a properly positioned ( ± 10 cm) platform (like a ship deck) if a connection mechanism (as depicted above) is used.

The Load-carrying Capability

As a train and beam design have now been recommended (here and in the introduction), we may evaluate it in comparison with the ideal vertical chassis design.

The beam: 40 m between poles, 20 m splints on poles, giving a 20 m weaker midbeam between splints.

The trains (described in the introductory article):

(Assuming a 3 or 4 m long center wagon.)

We may assume that the short version will never load the midbeam with more than half of its weight, provided its weight is being spread by the center wagon pulling both passenger cabins at floor level. The loading on the midbeam will probably be highest when the end of the train is there, so it may be possible to reduce this loading through intelligently controlling the proactive suspension.

The long train is so long that it needn't load the midbeam at all. It could put its weight on two poles, and the simple tensioning will relieve the midbeam of considerable weight. But this will cause a large midbeam load when an end of the train passes midbeam, so it may be better to reduce the simple tensioning for the long train, and use intelligently control of the proactive suspension instead.

A long freight train with vertical chassis would certainly give the highest degree of load control, as it could be programmed for bending vertically to follow the contour of the track, yet not follow the midbeam downwards.

The High-speed Train

Various high-speed technologies have now been discussed, so we may now try to choose the most promising and practical ones.

Two of the air-cushion schemes described seem to be most efficient:

A bogie design for both cases is described under A Universal Bogie.

The magnetic hovering and propulsion schemes may not be advantageous at all, and they will certainly make the beamway line much more expensive.

It may, for advertising purposes, be wise to aim at claiming two or three times the permitted car speeds. If the speed limit for cars is 80 km/h, it may be realistic to aim at 160 km/h for the wheel-based trains, and perhaps 240 for the hovering type. If there are 100 km/h roads, then perhaps a 2.5x advantage may be a suitable claim for hovering trains.

The train wagons described here were designed to be versatile – with that funny looking hat from which exchangeable cabins could be lowered. But the high-speed train will probably have conventional aerodynamic wagons without any cabin exchangeability. It should have the elevator built into the end of the front wagon – occupying e.g. the last 3-4 meters of a 24-28 meter long wagon. And it could be made aerodynamic by having a windowless bulge containing a toilet in front. Such a bulge should have an evacuation door in the end.

Alternatively, a compressor locomotive could be used before and/or after the train, and tubes for compressed air could be going through the roof of the train. The compressed air could be useful also if wheels are used, as the wheels could be turbine driven by the air. Mechanical energy from the locomotive can in this way easily be delivered to all the wheels of the train. But it would be more efficient to have compressors near the bogies – if high enough power and low enough air resistance can be obtained in this way.

If the train is for very high speed, also the rear wagon should be aerodynamic – perhaps with a toilet in the bulge. If compromise bogies are used, the rear wagon could be a versatile one, long or short – to be exchanged with a local train, or have a cabin which can be lowered down upon e.g. a barge or trailer.

High-speed wagons should have a set of stabilizers on the bottom – like anal fish fins, but at least five side by side. The outer ones should be able to detect side wind gusts, and the whole set should then be rotated sideways to counteract the roll torque from the side wind. This stabilizing roll torque will come in addition to the roll resisting torque from the bogies. With wheel bogies, this torque is limited by the danger of wheels being lifted on one side, but with compressed-air hovering, reversibility – the capability for suction (reducing the air pressure to below one atm) can give a higher roll resisting torque – particularly if there are (reversible) air cushion units also under the beam.

In case of a beam collapse, a high-speed train would be somewhat safer if all its bogies were tied together with one strong and slightly flexible cable on the left side, and one on the right side. These could be air tubes from a compressor loc. Falling down from a destroyed beam would now be a more gradual process – like a conventional train derailing, but more unlikely.

Beam Irregularities

A beam irregularity which just causes annoyance and material wear at low speed could cause a crash at high speeds, so special precautions will be required for fast trains. An irregularity could cause a small and sudden detour for parts of a bogie (a wheel or an air cushion unit), sideways or vertically. Detours sideways are unlikely to occur unexpectedly. They are probably caused by inaccurately built tracks, and the sideways-acting air cushions (or windsurfing) – used for both partial levitation (with wheels) and full air cushion levitation (without wheels) – should be able to deal with such irregularities quite well.

A vertical irregular detour could be caused by an object lying on the track, or the beam may be sagging down in the middle. A conventional railway line is very easy to sabotage by throwing an iron rod upon the track. A beamway is much harder to sabotage: It will require something like precision shooting with a slingshot, or climbing up a ladder or pole. If the beamway, as proposed above under The Control System, checks for a clear path by sending a microwave or ultrasonic signal through 1-2 beam lengths, it should be able to detect at least large obstructions. It will still be wise to let the train's first bogie have a scoop shovel in front, as the detector is unable to remove objects, just warn trains.

Wheels will be sensitive to objects in the path – particularly steel wheels hitting metal objects. Air cushion units could more easily be designed to deal with this. The trailing valve could simply yield by swinging up when hit by something hard. The leading valve could yield by retracting the valve in its own plane. The symmetry of the hovering bogie could be retained if all the valves could yield in both ways.

Another vertical irregular detour is: A beam is sagging down in the middle. This should not be confused with the (predictable) bending a beam does while it is loaded, as a beam will not have time to bend much when a train is passing at high speed, and at low speeds the train's suspension can easily adjust to height variations. (The last part of the train will experience some beam bending done by the first part, but this is predictable.)

A car should generally have as little unsprung weight as possible in order to avoid receiving hard knocks from an uneven ground. Wheels must imply a considerable amount of unsprung weight, while air cushions can seemingly give zero unsprung weight if they are thicker than the obstacles. This holds not only for reactive suspension, but also proactive suspension. The beamway system should have a database containing information on the extent of beams sagging of each beam, so that the proactive suspension control can increase the air flow and hence the air cushion thickness at the bottom of the sagging beam. This is more elegant than doing the proactive suspension control down in the suspension unit. If we look at the beam section pictures under Compressed Air Hovering and Propulsion, we see that the lid will remain quite closed even if the air cushion unit is raised considerably – and that the side walls of the beam should be vertical for this to function properly. (The middle part of the inner beam profile may bulge outwards – to give room for the axle of wheels which can slide inwards along their axle.) But it it is not enough to vary the airflow: The air cushion unit must be able to lower their valves considerably, so perhaps each air cushion unit should have a double roof. The lower one, to which the valves and dividing walls are attached, could be pushed down by an upper air cushion when increased hovering height is needed. This method will entail the problem of keeping the lower unit part level.

An alternative method for following the sagging beam could be to let the whole air cushion unit go down – by letting the last (vertical) part of each air supply tube be telescopic. This would simplify the air cushion unit design, but will increase the unsprung weight somewhat.

Both methods will permit independent movement for the four air cushion units. For a car, such an independent suspension for the four wheels is desirable, but for the beamway train, it is not. The left and right side height adjustments should go together – to ensure equal halfbeam sagging on the two sides. It may now seem sensible to propose method three: let the common (vertical) air supply tube for the whole bogie be telescopic, and we are indeed coming back to the old concept of suspension control at the suspension rods – those going down to the wagons.

So – to what extent could we give in to the urge to choose a simplifying centralization? The salient point is: How fast do the height adjustments have to be? A 40 meter long beam is traversed in half a second if the train speed is 288 km/h, and this might be a suitable design goal for a hovering train. A full air cushion height change might then have to be completed in about 200 milliseconds. (If air doesn't leak out this fast, the opening of valves needed for propulsion should mainly be done when fast height reduction is needed.) We must choose a height adjustment method for which the unsprung weight is small enough (and the correcting force is large enough), so that a full height change can be completed in less than 200 milliseconds. A good solution will permit high speeds, and permit some beam sagging before the halfbeams will have to be turned upside-down.

The correcting force needed will probably require another air supply giving an extra high air pressure, delivered by an extra compressor stage. If the first method (increasing the air cushion thickness) is used, the high pressure air is sent into the upper air cushion compartment. For the second method, with four telescoping tubes (each having an inner tube sliding within an outer tube), each inner tube could have a closed end (above the main air entrance), forming a piston which is pushed down by the high pressure air. Similarly, for the third method there will be a single piston in the suspension rod/tube. If the high pressure is needed infrequently, it may be generated by heating the air cushion – perhaps even by igniting an injected fuel.

For the first two methods, the extra compressor stage could be on the bogie. For the third method, an extra tube for high pressure air should come from the compressor locomotive or the train – so that the unsprung weight of each bogie could be low.

For the moment, before calculations for comparing the three methods have been done, the second method may seem a good compromise, having the advantage over the first one that the lowered part of the air cushion unit is easily kept level.

A Compromise Bogie

The compromise implies: We are aiming for quite high speeds, but not so high speeds that propulsion (linear motors) must be provided by the track. This means the train line will be as cheep and simple as when plain old-fashoned wheel bogies are used. Wheels are used for pulling (and regeneratively braking) the train, but only to the extent wheel forces are needed. Air cushions lift as much weight as possible, but are controlled to stop lifting before wheel traction is lost. At low speeds, air cushions are unimportant and may be dispensed with. (In case of e.g. compressor failure, the train could still roll along at well beyond 100 km/h, with only moderately annoying delays and mechanical wears.)

As the speed increases, the forces to be excerted by wheels are decreased, so the load carried by wheels may be reduced by perhaps 80 %. (Emergency braking is not done with wheels, but by clamping the beam.) As the air cushions take care of the side forces, the forces acting on the wheels are greatly reduced, so speeds a little above 200 km/h may be achievable. When movement is controlled with wheels, the speed can be accurately controlled while stopping in a hill or in strong winds.

Simple air-surfing vanes could be used instead of compressor-generated air cushions, but investing in air cushion equipment for the bogies gives some extra advantages:

Various wheel types

The flanged type to the left will not be interesting, as air cushions acting on the side walls will ensure proper tracking.

The steel wheel in the middle is simple and reliable. It shouldn't be too noisy when most of the forces against the beam are eliminated by air cushion operation. It has the advantage of being thin, giving low air resistance at high speeds.

The special rubber wheel to the right gives higher friction, so more of the train weight can be air cushion carried before the pulling power of the wheels becomes too low. If the rubber part of this wheel is destroyed, the wheel should be able to function well as a steel wheel – at least if the air cushion assistance is automatically increased. Well armored compact rubber should work well here.

The above picture doesn't show how the hub motor occupies most of the wheel, but this is shown in the next picture. The hub motor will in fact be a very thick stationary axle for the wheel, so there is no need for an axle connecting two wheels running side-by-side. This means the bogie can easily adapt to half-beam separation variations.

The suspension.
The part inside the train is approximately as shown in the first picture in the chapter A Universal Bogie above: The suspension can move sideways (passively or actively) by means of minibogies rolling along roof rails which follow the curvature of the train roof. But in this design variation the entire suspension control mechanism (except the motors in the minibogies) is moved up to a nacelle above the train.

This design variation doesn't use a centralized compressor, but has one compressor in each suspension. As the compressed air has to run only a short distance, it will lose little of its heat on the way. Such near-adiabatic operation gives energy efficiency as well as little or no problems with freezing at the air cushion pads.

The bogie, its suspension and compressor hang together as one unit which can easily be detached from the wagon below – because only simple suspension rods (each with a power or signal cable attached) go down to the wagon. (The contents of the nacelle could have been placed in the bogie in the beam, but this would have caused a much stronger air resistance.) This combined unit may function as a microloc, capable of moving by itself.

The small distance between the roof rails means that the lifting force from them can continue along one (carbon fiber) strap running around the wagon between windows. It is also easier to protect the train interior against precipitation etc. when there is a narrow slit in the roof.

The minibogies in the top of the wagon should be about half a meter long, and have notches under each end. Horizontal crossbars on the lower end of the two suspension rods are held in these notches. (One notch should be slightly displaceable – to correct for geometric inaccuracies when the minibogies go down the steeper part of the roof rails. ±80 cm displacement should be allowed.)

The picture seems to have one suspension rod going up to a piston in a cylinder marked Lifter. (A suitable travel length for this piston could be 20 cm – for dealing with slope changes.) This picture part should be regarded as containing two superimposed suspension rods going up to two lifter cylinders. (A signal cable runs along one of these rods, and a power cable along the other.) Above these cylinders, the upper suspension rods are attached in-line at the central part of the nacelle. These are depicted in two colors for distinguishing those going to the left and right side of the bogie. The short ones go to the air cushion pads working against the bottom of the beam. The four upper suspension rods are hollow (or supplemented with tubes) for transferring air to or from the air cushion pads. Three such rods/tubes may be needed for each bogie side – in case separate air cushions are needed against the side walls of a half-beams. Alternatively, sensors in the air cushion pads can monitor the distance to the side walls, and local valves can direct more air to the side where the wall is too close.

The upper suspension rods should be fixed in relation to the nacelle, so that train tilting is handled by only the lower part of the suspension. But there should be a little flexibility up here, so that these rods can adapt to some variation in the separation between the half-beams.

The mechanism in the nacelle is as follows: The compressor creates a high air pressure in the pressure tank, from which computer-controlled valves let air act in the lower or higher part of the system. Lifting or tilting the wagon is done by increasing the air pressure under the pistons in the lifter cylinders, and all the suspensions should co-operate in doing this in the same manner. If done equally for both cylinders, the wagon is simply lifted. If done differently for the left and right cylinders, the wagon is tilted. The train may be left to get its natural tilt, as determined by the gravity and the centrifugal forces, and this is what the passengers will feel is correct. This can be done by connecting the lower parts of the cylinders with a shunt tube which equalizes the pressures. There should be an adjustable valve in this tube – for controlling the air resistance in the shunt, so that penduling is sufficiently damped.

Air from the compressor is also sent up to the upper air cushion pads (three on each side), and the lower air cushion pads (one on each side) may be connected to the air intake of the compressor. A valve should be able to rapidly exchange these two airflows if a tilting force is threatening with lifting up the wheels on one side.

The two lower suspension rods can be rotated when they are being attached to or disconnected from a wagon, as they hang under pistons which can be rotated in circular cylinders. A wagon (or train) to be attached, is lifted by a bottom support so that the horizontal crossbars of all the (properly aligned) low suspension rods are swallowed by roof slits. A maintenance mechanic then goes along the roof and turns each suspension rod 90°, ensuring each crossbar is being attached to a minibogie in the roof. A single microloc – a bogie and its suspension unit – may be exchanged on a wagon hanging under a special (open) service beam.

Each wheel in the beam contains a hub motor, which (as shown by the inner circles in wheels in the picture) occupies most of the wheel volume. Such a motor (a Protean motor, measuring 42 cm in diameter, 11.5 cm thick) develops 80 hp at 2000 rpm. A wheel diameter of 55 cm will give 200 km/h.

This power gives 320 hp for one bogie, or 1920 hp for a 6-bogie front wagon. (This is a 28 meter long minimal train with a rear elevator. A short or long wagon may be added, but as this will bring along correspondingly more motors, we can check the power requirement by considering the single wagon train.)

If such a train, weighing perhaps 20 tons, were to negotiate a 10 % hill at 200 km/h, it would be ascending at 5.56 m/s, needing 20000x5.56/75 = 1482 hp for just lifting the train. Most of the rolling resistance will be removed by the air cushions created by means of additional motors, but the air resistance remains to be overcome. The air resistance at 200 km/h would require about 1000 hp. This is the air resistance for an ordinary train. The beamway train has a considerable additional air resistance from the bogies in that narrow beam. On the other hand, it hasn't that air-resisting underside, and the air can be displaced to four sides instead of three. We might guess if these differences cancel out.

1920-1000=920 hp remain for hill-climbing + the (strongly reduced) rolling resistance. This means an ascension of less than six percent can be managed at 200 km/h. At 140 km/h, the air resistance takes 500 hp and getting up the 10 % hill takes 1037 hp, leaving well below 800 hp for acceleration. If we want to avoid slowdowns in hills, there are several solutions:

As the miniloc is needed only above 120 km/h, it is unusually easy to use. This is because the higher the speed is, the lower is the pulling force needed for supplying a certain number of horsepowers. Low pulling force means we needn't worry about using a loc having a small fraction of the train's weight. A miniloc with e.g. 2000 hp should weight just a few tons, and if it hangs under eight wheels, the wheel-protecting air cushion needn't steal much traction, ensuring weight is put on the wheels only when really needed for traction.

Having an additional detachable microloc fore and/or aft gives little extra pulling power, but this gives some new possibilities. These microlocs may run away by themselves to deliver parcels in the vicinity. Or the one in front could run 1-2 kilometers ahead of the train to check that the track is ok.

A Lighter Beam?

We may now consider saving beam steel – by using the strategy described under Are Strong Beams Needed? above. Could we even try to do without the steel beam – as envisioned here?

This may not be wise. A continuous beam is valuable for:

Quite small beams (e.g. 50 centimeters high instead of 80) may suffice if the mini trains are limited to 3-4 tons. Air cushion units need little height in the beam – although the air resistance of their movement through the beam would be lower if they could have plenty of space. Wheels might only be needed as a supplement – for low speeds, or for use in conjunction with weight-reducing airfoils – and in these cases they could be quite small.

Small beam dimensions will be OK in tunnels, as they can there easily be fastened with small intervals. If the beam is hanging in a (cable-stayed) span, it is quite easy to have many suspension cables. If the beam is held by poles or racks, it will be inconvenient to have shorter intervals between them. The train should then lift itself at fewer suspension points by increasing the air cushion pressure at the poles and splints rather than letting the weight be evenly distributed between all the suspensions. When an air cushion unit follows a sagging beam downwards, it is not to exert pressure there, but to be ready to exert a stabilizing sub atmospheric pressure.


A point (switch in USA) is a place where more than two beams meet, enabling two lines to merge into one, or one line to split in two.

(General description in Wikipedia)

A special transition area for line switching

The inverse U bracket cover (with its beam-holding tongues and power line) is not shown here – to reveal the interior.

The gray area is traversed by wheels, so any additional support – vertically from the cover – must be in the undisturbed white area.
The vertical plate on the right side is one of two such supports shown. It holds up the “Manhattan” in the middle. This is the so-called frog, and it is strengthened by having extra steel on the underside.

(The discussion area of the above-mentioned Wikipedia article tells how the term frog came from 19th century horse hoof terminology. Modern readers are probably more familiar with Manhattan than with horse hoofs.)

The stiffness of our trains is well suited for crossing the weak and missing parts of the switch area, particularly if air cushions are used, and/or the proactive suspension is programmed for removing weight from a bogie while it passes here.

Now to the problem of choosing the path through the switch area. If the direction is from the multi-track side to the single-track side (called the trailing movement), this is no problem, but for movement in the opposite direction, a track choice has to be executed. This can be done by the vehicle, or by a switch mechanism. Altering the path in the switch would take a few seconds, and this is rather much if a stream of small vehicles is to pass through. Such a stream is what SwedeTrack deals with, so this system expects the vehicles to do the shunting (switching), like cars in a road crossing. But for our trains with several bogies, it is extremely important that all the bogies in a train be sent into the same track, so here it is natural to let the line do the switching. It would be foolhardy to expect computers to dispatch all bogies without a glitch, and impractical to have secure steering on all the bogies. When we are dealing with real trains, several switching operations will rarely be needed within one minute, so if high speed operation is valued, the method described under Movable Beams above, would be best: Make a track switch by pulling one end of the beam so much sideways that this end is aligned with another, adjacent track. It is for bus traffic it would be nice to have the transition area switches described in this chapter.

The switching mechanism could simply be a mini version of the moving beam, having two U-shaped steel rails. As shown in these two picture, the left (pivoted) ends should be attached near the end of the left beam, while the right ends should be movable between the two tracks. 

This arrangement would have the added advantage that the transition area would give continuous and quite firm support for the wheels. But the switch mechanism would have to adapt actively to trailing movement (from the right). As a safety mechanism, the pivoted ends of the U-rails should be able to be swept aside by wheels arriving from an unexpected track.

Another, and mechanically simpler solution, would be to have a single moving rail with the same U-profile, and pivoting fastened near the middle of the transition area, where the two middle wheel-paths (gray) meet. The moving end would be placed over either rail portion of the single track (to the left in the picture). 

This rail would be pushed to the correct position by any trailing wheels (from the right), but such pushing sideways with wheels would be a crude, low-speed mechanism which should not replace decent motorized switching.

The movable rails in these pictures are straight, as they are used by both the straight and curved track, and it should be possible to pass through the straight track at a quite high speed.  (The movable rails are of cause tapered to a minimal thickness in the ends, so they should not be experienced as bumpy at least if rubber wheels are used.)

Here is a different variety of the previous design.

When it is set for using the curved track, an appropriately curved rail in the movable double-rail is used.  But this conflicts with the first half of both the upper and lower of the four side walls.  (The two middle side walls must be absent in the switch area for all these three designs, but generally three of the four side walls will be available for the passing wheels.)

The movable end of each movable rail should (in its outer corner) have a vertical rod going up to the switching mechanics under the ceiling of the transition bracket. In case of system malfunction, it should be possible to operate the switch mechanically from a slowly approaching train.

The sideways movement of the bogies in the switch area could be regarded as corresponding to a very small turning radius, but this is not the relevant measure here, as the wagons may not really have to turn so abruptly. This is because the sideways movement, which may correspond to the width of a beam – about 80 cm – would largely be performed by suspension rods being displaced sideways across the wagon roofs. If the suspension system were designed to absorb the entire sideways bogie shifts in switch areas – ± 80 cm – even the largest trains could glide smoothly through a switch without tossing wagons to the side, albeit at a moderate speed. (Switches should be near stations.)

The trains depicted here were not designed with this smooth switching in mind, so they happen to have only something like a ± 40 cm suspension. This means that while it goes through a switch, the wagons have to be moved 40 cm to the side, so that the following suspension rods must be shifted 40 cm to the opposite side, and then the suspension system has absorbed the sideways movement without having to rotate wagons around a vertical axis during switching. This should work satisfactorily for the small (our yellow) bus/tram-trains, but large trains should be scheduled for passing only straight through such switches in normal operation.

A ± 80 cm suspension movement would also be useful in conjunction with stabilizing side wheels (described under The Suspension). If the suspension unit is made so wide, the cabin-grabbing frame around it must be correspondingly thinner, but this shouldn't be problematic. A ± 80 cm suspension movement is definitely a valuable design feature.

Closely related to the switch area is the simple crossing of two tracks in the same plane. Also this should be placed on a pole (or equivalent multipod stand). Crossing should occur at an angle of 45-60 degrees, so that not both wheels on an axle have to fly at the same time over the crossing gap. Power lines have to be absent here (like in a switch area), because they are below the wheel-top level. These powerless stretches are insignificant, as there are normally other powered bogies or locomotives in the train. (In the unlikely case of a full stop for a small train here, the emergency power supply in the center wagon can power far longer emergency trips.)

Wagon Exchange

When two beamway trains meet, it may be useful to transfer a wagon from one train to the other. This will be done mainly when a long distance train passes a city (or even a small town) having a local beamway line. It may be difficult for the long distance train to come sufficiently close to the center of the city, but even if it could, it is best for the passengers to be able to choose the city part for departure or arrival.

The wagon to be transferred should be the last one in the train, and it should be a short one in order to be manageable by a local train. A long distance train could have such a wagon as its third wagon, as passengers there will not need easy access to the elevator. They will go to this rear wagon while the train approaches the destination at which the wagon will be detached. And passengers arriving by an attaching wagon must immediately leave it.

Two possible track configurations for wagon exchange.
A small local train has delivered a wagon (shorter) for a long distance train expected to come along the straight track. The small train is also ready to receive (from behind) another wagon. (Track switching is not shown, but should be obvious.)

The track configuration to the left may save a few seconds for the long distance train, as the waiting wagon is waiting a little ahead of the position in the rightmost configuration. But this is likely to be unimportant, as the long distance train is likely to have a full stop for passengers to/from the ground there.

The incoming long distance train will detach its last wagon in good time before arrival, so that after this train has passed, the track can be switched for sending the rear wagon out to the local track. The loose wagons are supposed to be motorized and computer controlled. They might then be exchanged while the train passes at high speed, although this is unlikely to be the chosen procedure.

If the long distance train isn't met by a local train, it will just leave the delivered wagon alone. This should then be parked at a platform by which passengers may descend if they wish.

The shown track configurations are symmetric, and will work for traffic in both directions, provided the local train can go through its loop in the same direction as the long distance train has.

The Control System

The trains should detect the identity of the poles they pass, so GPS navigation should be unnecessary.  There should also be some sort of wireless communication from pole to pole between the half-beams, perhaps using a waveguide effect.  By detecting the signal signature characteristics of the received signal, it should be possible to detect irregularities in the beam.  The transmitter should send a complex waveform, and the receiver could Fourier analyze the signal, checking both the sine and cosine components. The reference signature, indicating a healthy beam (without birds and sabotage objects), could be stored locally or in the central database. Such a path check will be valuable to check if a moving beam track switch operation has been properly performed.

If ultrasonic signals could be used, birds may be scared away from the interior of the beam.  The train should also receive information about the speed limit at this place. This limit will be adjusted according to special conditions detected at that place.

In addition to this computer data exchange, there will of cause be at least voice communication with the train staff, as well as central monitoring of what happens on the train.

If poles stand on unstable ground, it will be useful to have strain gauges in beam holders. The poles should be so adjustable that incorrect position parameters can be corrected. The correction could be automatic, using motors in the pole, but it should generally be OK to have only screws for these adjustments in the poles, so that workers would go around with power tools for adjusting those screws. The above mentioned detection of signal signature characteristics should work between non-adjacent poles, to detect incorrect positioning of a pole between.

The front of the train should have a radar (or ultrasonic) obstruction detector. The interpretation and response choice for these detections will be far simpler than for road vehicles.

At some road crossings, it may be useful to detect the approach of extra high vehicles, by means of e.g. photocells. These detections could simply be brought to the attention of the conductor.

The elevator should detect obstacles below it. It will be useful to start the train while the elevator is going up, so also obstacles ahead of the elevator should be detected.

The detection of passengers and their payment status will be important for the prospect of fully unmanned operation.

Collecting Energy

The beamway is already an energy collection grid, as it is ready to collect energy from regeneratively braking trains, and if it gets too much energy in this way, it should be able to return the surplus to the external AC grid. This system should have no problems with working as a collection system for various types of energy it finds along its line.

Where the track lies in the east-west direction, it should be quite simple to attach solar panels to splints facing south. They could be vertical, but should preferrably be tilted somewhat upwards. The width of such a panel assembly could be 10 – perhaps 20 m – without stressing the beam significantly – at least if the panel assembly is stiff and balanced at the pole. The height could be 1-1.5 m. With an area of 30 m², an efficiency of 15 %, and the normal 1000 W/m² sunligth, it would give up to 4500 W. Such panel assemblies (with interface electronics) could be factory produced to be hooked onto the upper and lower edge of a splint, and could be plugged into a connector on the pole – just like the linear motor stator modules. It should then be possible to deploy a panel assembly in minutes from the external platform of a service train.

Solar panels could also be mounted down on the pole, from the most elegant type which is flush with the southern surface of the pole, to the most efficiently tilted (but ugly) flat plate.

A wave power plant could be combined with a tube or pontoon chain for a coastal beamway, although beamways will certainly be located mostly where there is little wave disturbance.

Building a Beamway

The most awkward parts to handle are 40 meter long half-beams.  Few roads can handle such beasts, but ships and conventional trains can handle them, so the building of a beamway line should start at a point near a harbor or a railway line, or at least a suitably connected bay, river or straight road.  We shall later see how the beamway can extend itself by transporting its own beams.  Normal poles and (20 meter long) splints should be transportable on normal roads, but in difficult terrain the beamway may have to transport also these, however only from a trailer parked on a road which is hopefully not too far away.  

In difficult terrain, the workers may have to go on foot to the places where they will build the sockets of the poles.  They may have to do some digging and concrete pouring at those places, but hopefully a terrain-going concrete transporter with a long hose can pump concrete into e.g. a plastic mold shaping a pole socket.  A connection mechanism on top of this concrete socket should enable an aluminum pole to be inserted quite rapidly by helicopter.

The simplest method is to use racks with two feet which can simply stand on the ground.

This is a cross-section of a beam carrier.

The upper wheels of the beam carrier roll on the beamway like an ordinary bogie, but they are not motor-driven.  The lower wheels carry the half-beams to be transported.

This little vehicle is operated by remote control, and can do these things:

Half-beams can be pushed to and pulled from carriers when their brakes are on and clamps are opened.  When one or two half-beams have each end clamped upon a carrier, a long wagon is formed, ready to be pushed by a locomotive to the place where the beamway is being extended.  The clamps should be quite long along the clamped half-beam (visible on the left side of the next picture) to dampen pendulation.

If the beam carriers have a long way to go, their transport capacity can be greatly increased through a relay race: Each carrier pair (with a miniloc) will then shuttle back and forth on a small part of the transport route.

This is a beam extender, parked in the normal working position.  Beam carriers (to the left) are delivering half-beams (here one, but normally two), and push a half-beam forward through the (reddish) rear clamp (which has internal rollers) of the beam extender, and on to the side-tilted end of the long forward boom.

The motorized wheel on the boom tip pulls the half-beam forward, while the rear boom ensures balance by pushing the beam from below.  The rear clamp then helps lifting to the correct position.

Here we see the half-beam being lifted up.

(In this case the splints had been mounted in advance.)

The end of the half-beam is inserted upon the tongue under the (U-shaped) transition bracket.

The beam extender then prepares for the other half-beam by turning both the boom tip and the rear clamp to the other side, and by lowering both to the level of the incoming half-beam.

The splints can then be securely fixed to the beams – by means of the curled brackets depicted in the beginning of this chapter – by workers standing on a high side platform of a service vehicle.

The next step is to mount some clamps about midways between the poles, and then cover the beam with the plastic, under which the power line can now be inserted from below.

This method should also work when the half-beams are staggered – displaced half a beam length relative to each other.  Each half-beam being mounted will then be pushed half-way through the last U-bracket – which had been affixed to the end of the previous (adjacent) half-beam.  This staggered mounting will normally be used in a suspension bridge, when each U-bracket is hanging in a wire.  It may, however, then be more convenient to assemble the beam below (e.g. on barges), and then hoist it up to its wires.  Perhaps the whole cable-wire-beam assembly will be assembled at the low site.

Splints will not be used with such a staggered beam, as each half-beam will act as a splint for the two half-beams joined at its middle.

In really difficult terrain – like a mountainside sloping 45 ° to the side – it will be useful to transport also the poles to their positions along the growing beam. This should be possible if a staggered beam is used. A lightweight, collapsed telescoping pole could hang under a single-wheeled roller which could roll out to the end of the last half-beam. There the pole (with its attached U-bracket) is fastened to the end of the half-beam, and the outer tube of the telescoping tube pair is lowered till its end stands on firm rock. The bracket foot of this tube is then bolted to the rock. The pole can be light-weight because the distance between poles will be half the normal separation when beam staggering is used, and the poles here are shorter than normal because no traffic taller than pedestrians (or skiers+snow) will occur in these steep hills.


Tunnels can be made quite cheaply with an automatized procedure if the beam is mounted as soon as possible during the process. A utility tunneling machine could be used for drilling a hole whose diameter should be <4 meters. Behind this comes a machine that cuts a slit in the roof, e.g. 90x90 cm. (Such a slit cutter is also useful for enabling a beamway line to us a road tunnel without making this tunnel too low for trucks.)

Immediately behind a special beamway tunnel interface machine should follow. It should have a telescoping beam whose length can vary with an amount corresponding to the length of each standard beam piece. And there should be a special wagon having a container able to receive the removed slurry material from the tunneling machine, and carry it out of the tunnel to a suitable dumping place. This wagon should also bring half-beams and their mounting brackets forward when needed (when the telescoping beam is able to contract enough to give room for the next two half-beams), and then mount the half-beams approximately as described above. (This mounting of half-beams could be simplified now when the tunnel floor can give support at a known level.) Attaching brackets in the slit by bolting, and then mounting J-profile halfbeams (adjustably) in the brackets, may also be automatized. The waterproofing needed in under-water tunnels may be remote controlled. (But if the tunnel is above the sea level, leakage may not be a problem, as a brook may be permitted to occupy the bottom of the tunnel.)

Measuring tunnel directions should also be simpler when the beam gives simple-to-use reference points.

This tunnel has a diameter of 3.7 meters, containing a train that is 2.35 m wide and 2.40 m tall. The blue contour shows a 20° tilt (with suspension rods on one side contracted).
The object in the bottom (50 cm thick) shows how a maintainance worker can be passed. Dumber creatures can be toppled over to this position without being really injured. The bottom has also room for accomodating e.g. power cables.
Stabilizing ailerons (rudders) should be on the lower part of the front and on the sides of the bottom.

This wagon shape, which fits so well in tunnels, is anatomically shaped, being widest at elbow height for sitting passengers. It is also more aerodynamic regarding cross-winds, giving less wind-induced tilt.

As tilting may be minimized in tunnels, the tunnel diameter could have been reduced to 3.4 meters with a 15° tilt. But narrow tunnels give higher air resistance.

A 2C Metropolitan System

We should distinguish between two types of beamway systems:

Previous beamway/monorail systems are clearly locked into a metropolitan orientation. The beamways use monolithic all-in-one beams (like SIPEM), and these are designed for metropolitan use, in areas criss-crossed by roads, so that heavy machinery may be used anywhere. The beamway designers may know they are designing a metropolitan system like SwedeTrack, but it would really be unwise of them to design for incompatibility with country-wide use.

The present 2C beam system was intended to be country-wide, and to need little ground machinery support during line construction. So we will now ask: Can the 2C beam be used in a metropolitan system? It was conceived to resemble the SIPEM beam (as described by SwedeTrack). The important system type difference is: Metropolitan systems may be designed to handle streams of small private cars (single wagon minitrains), while country-wide systems are unlikely to permit this, but rather deal with centrally controlled trains. When cars come in streams, track switching becomes too slow: The cars should do the switching.

This doesn't really concern the 2C beam, as track switching is done in the transition area between sequentially connected beams. When we are dealing with streams of small cars in a metropolitan system, levitation mechanisms are quite unlikely. We may assume the vehicles are rolling on wheels, so for pushing the bogies sideways, it is natural to think in terms of horizontal wheels meeting vertical walls in the transition area. Those vertical walls should follow the tracks going out from the switch area, approximately as if the upper vertical walls of each outgoing beam started at the incoming beam, except they should be so high up that the bogies and their wheels can pass and cross under them.

A bogie (with blue frame) with a selfswitcher module (white) mounted in front. It approaches a switch area with guiding walls hanging down from the ceiling, just above the wheel level.

Seen from the front: The selfswitcher raises the horizontal wheel on the side it should turn to, and this wheel engages the wall leading to the correct out-track. The wall leading straight ahead is here followed.

The front plate of the selfswitcher is here removed, so we can see four small wheels which can slide vertically, each in its own slider. This vertical movement is restricted by three levers (pivoting around their centers) which prevent certain dangerous situations. The two lower levers prevent a horizontal wheel from raising if the opposite vertical wheel is unable to go down as low as the large wheels. This prevention will occur in the beam, where the edge of the half-beam will prevent the vertical wheel from going down, and there is no room for the horizontal wheel to rise. It will also occur in a switch which determines the switching, and the bogie is running in a track with edges preventing the vertical wheel from going down. The third (upper) lever prevents the selfswitcher from trying to choose both paths at the same time. (Some detail about the lever movement: The bolts of the front lever (most horizontal here) don't block the movement of its companion lever because the pivot bolt goes forward to the front plate (invisible in the picture). And the right end of the front lever connects to its slider bolt via a V-shaped bracket into which the companion lever can go down.)

These safety mechanisms should of cause be implemented in the programming of the electronic control system, which receives its input from inductive or photoelectric detection of any edge below the vertical wheel, so those small vertical wheels should never be lowered to a track wall.

Towards the end of the switch area, track walls should rise gradually and finally blend into the track walls of the half-beams. The small vertical wheels will roll up these ramps and cause the horizontal wheels to be lowered – if this lowering had not been done by the electronic control system.

To turn the bogies into a certain side track is, however, only half the job. The other half is to support the bogie whose one side passes the weak frog area and the adjacent gap for the unchosen track. We have discussed this before, and seen how a real train with several bogies, and/or large bogies, can distribute its weight to avoid loading weak or missing parts of the track. But now we must deal with small vehicles, for which a rigid suspension may not be enough to keep the single little bogie level. They need a torque inducing a roll away from the onsupported side. The horizontal wheels, rolling on the vertical walls, can provide this torque. If a stronger torque is required, there could be selfswitchers on both ends of the bogie. The horizontal wheels resemble those described by SwedeTrak here (fig. 12&13), but are engaged only when they are used to control switching. In the present design, the bogies run on all its wheels all the time, so rubber wheels needn't be dimensioned for experiencing twice the weight because wheels on one side are lifted over a central track wall. These walls (at the suspension gap in the floor) are omitted in the present design, as there are continuous walls along the outside of the Y-shaped track area, and the walls in the ceiling prevent the bogie from moving away from the outer track walls.

When the track does the switching (and there are probably no ceiling walls), the movable rail has walls on both sides, so the wheels are properly guided. (Telescoping axles can increase their length somewhat, but not become shorter than normal, so they will not cause wheels to come into the gap.)

A switch area may be designed for doing the switching only periodically – by having both movable rails and the guiding walls in the ceiling. When movable rails are in place, their side walls will prevent selfswitchers from trying to follow the other route. When the movable rails are removed from the switch area, selfswitchers are free to choose route. (But the track may become too uneven for steel wheels if it were shaped to become even with movable rails in place.)


GTS (General Transport System) is a Swedish beamway transport concept, a successor of Swedetrack's Flyway concept. It is described here. (In Swedish here)

The word General in GTS seems to apply to the transport modes covered, and mainly tries to combine car and rail transport. Cars and the larger bus/tram/train vehicles seem to be combined, too (at least according to the illustrations), although small cabins (often called pods) are emphasized.

The presently described transport system should be compatible with the Swedish GTS, but is trying to be general in various (not independent) dimensions:

  1. Transport modes:
    The (optional) exchangeable cabins will give generality across modes like train, bus and boat. High speeds can to some extent replace air transport. Beamways can cross water stretches in tunnels under the sea bottom, in submerged floating tunnels, or on pontoon-carried beams.

  2. Wagon sizes:
    Bus/train formats are emphasized here, but also small, private vehicles may be accomodated.

  3. Cargo types:
    Passengers, equally light goods, and cars are accomodated. Cars are carried in large wagons, mainly for long distances and crossing difficult terrains. Putting cars inside wagons may seem inelegant, but passengers can now have their private cabins but still get access to toilets (and this fact of life
    is inelegant). Assuming people will buy compatible vehicles engenders elegant, but futile transport schemes.

  4. Ownership:
    Passengers may own (rent) a seat for a ride, a cabin for a ride, really own a cabin (renting bogies), or really own a whole wagon (with bogies).

  5. Personal mobility:
    Wheelchairs should be accomodated. The small GTS vehicles are, according to depictions, evidently not prepared for wheelchairs. And stairs can be seen to go up to station platforms, but not elevators.

  6. Traffic loads:
    This is described in the chapter Scalability. This point covers also a related generality: 24/7 operation. (Can stations with elevators be open all night?)

  7. Topography:
    This generality implies: Lines can be built also in difficult terrain, where heavy machinery may not be used for lifting in place the beam. The (self extendable) 2C beam has important advantages here.

  8. Line positioning:
    A beamway line can be positioned and easily repositioned for going in its own tunnels, in road tunnels (where trains behave like busses), immediately above the ground traffic (within elevator operation range), or high above the ground (up to perhaps 20 meters, where the elevator can only be used in unsteadied emergency mode).

  9. Vertical movement types (outside stations):
    Steady (elevator - for not needing elevated stations), unsteady (hanging in wires - mainly for emergency escapes), crash softening straps only, and no vertical cabin movements.

  10. Speed:
    Rural lines with long distances will need high speeds. Urban lines with frequent stops will not, but light trains with good acceleration can still use quite high speeds when going to and from suburbs. The bendable/twistable 2C beam will enable high speeds.

  11. Distances:
    Toilets must be available on long distance trips. Telling passengers to leave and pause at certain toilet-equipped stations is not good enough, particularly if families are expected to use the transport system.

This was about beamways. But if we widen the scope to include conventional railway, the load weight can be included as an additional generality. The distribution of the load weight is quite peculiar: Most of the traffic will be passengers and equally light goods, at only 7-8 % of the full track load capacity. When high speeds are emphasized, the traffic percentage will approach 100. When the very lowest loads dominate in this manner, it is really unwise to sacrifice generalities 1, 2 and 4-8 in order to maintain weight generality and its old technical standard – particularly when the many other advantages mentioned in other chapters here are considered.

Regulatory authorities, such as the department of transport in the country, will discourage or prohibit construction of transport systems lacking important generalities. Certain generalities, like ensuring personal mobility by admitting wheelchairs, may simply become mandatory. (Ensuring this by means of elevators in station buildings may not be an acceptable alternative to elevators in trains, which may become mandatory for safe and rapid escape anywhere along the line.) Other generality deficiencies may be discouraged through economical mechanisms: Systems with such deficiencies may be disqualified from receiving governmental financial support. This is presently a quite important incentive for those building conventional light rail systems: A national rail standard is often chosen - not because it is needed for technical reasons, but because it is needed for receiving governmental financial support. When beamway beams are mounted over public roads and streets, the conformity pressures must be expected to become stronger than they are now for conventional rails.

A local transport system may not need certain generalities, but permitting such deficient systems to be built, will cause a technical fragmentation of the infrastructure in the country or state, and it may make it impossible to merge or extend the local system into an extended system later. That is: Interoperability may become important in the future.


A beamway train will have an "upper floor" for itself, in which other objects will rarely come in the way. It can have this privilege (right of way) even in city areas with heavy traffic. Others may occasionally use this high space in the beamway's path, e.g. for construction activities, but will then have to look out for trains.

A 4.5 meter safety zone is announced here in Dortmund

High-reaching activities are subject to approval here in Wuppertal.

The probability for collision might be approximately like the probability for a train hitting vehicles at a level road crossing, but without the devastating consequences of people getting hit. It will rather be about light equipment getting hit by a light train. A beamway train should have something like a U-shaped aluminium beam along its front floor for protection against e.g. ladders and boat masts.

A bus can easily crash due to blunders done by its driver or others on the road. A tram is safer against blunders from its driver, but correspondingly less able to prevent accidents caused by others in the traffic. A beamway only has to look out for objects coming into its path, and this is such a simple task that it can generally be dealt with by a computer analyzing the signal from a simple radar looking ahead from the train front. When an obstruction is detected, the train will automatically respond with braking strongly. If the train is going in 200 km/h, and the emergency brake gives a deceleration of 1 G, the train will stop within 160 meters. At this speed, the track curvature radius will be at least 630 meters (assuming a .5 G centrifugal acceleration is tolerated), so those 160 meters ought to be within sight. If they are not, reduced speed may routinely be used there. Or other indicators may be used. At road crossings, photocell detectors in the vicinity may be used for reporting tall vehicles approaching.

In order to obtain really high run safety, a small unmanned vehicle can run a few hundred meters ahead of a train. If it hits an obstacle, a signal is sent back to the train (or a series of signals is interrupted), which then turns on the emergency brake. This vehicle may have a video camera and be an extra eye of the train operator. Such vehicles may also be used for important parcel deliveries.

A conventional train can easily derail if something falls or is thrown in its track. A small snow avalanche is enough, and it is simply assumed in railway operations that nothing happens to the track after a several minutes old line check. A beamway, however, will not be affected by moderate avalanches. If snow or dirt heaps up to train heights, it should be detectable by the radar. The poles in such areas should be well secured by means of wires or rods anchored up in the hill, or the beam would jump in long spans. If a pole collapses, a steel beam will not break, but will go in an awkward bow. Several poles will have to collapse for the beam to sink the critical four meters or so. Such a disaster will require a major landslide.

If such a destruction occurs, the power line and any (fiber optic) signal cable will probably break, and this should be detectable from the train. If it is not, the beam will probably be bent so much that any beam integrity detection system (based on e.g. sending a microwave signal through the beam and checking for signal signature changes), should be able to detect it. If it doesn't, it is probably because the beam approaches the ground too gradually. But then the ground should be detectable by the radar looking along the beam.

A most serious condition will be a full beam blockage, most likely a partial beam breakage, which makes it impossible for the bogies to pass. The suspensions will then break. The last parts to break will be the emergency brake straps, which cannot be surprised by forces coming at an awkward angle. They can now give extra safety if they have some meters extra length in one or both ends. This length should slide out with a great counterforce, and finally tear apart something with its end. A train collapse at high speed will be a series of such strap breaks. The beam should not be affected by this, as it can withstand greater forces along its length when they can be distributed among several poles which are locked at a fixed angle to the beam.

It may seem risky to sit five meters above the ground, but a beamway train is quite unable to simply fall down. It has a series of suspensions that have to fail, and this would entail a gradual lowering process even if the brake straps were not ready to yield perhaps those five meters.

The train may occasionally pass a deep gorge at greater height, from which falling down would be catastophic. This could be compared to a train derailing or a bus leaving the road at the brink of a precipice. Both these events are probably much more probable then a beam collapse.

If these prospects look grim, it is simply because we are considering scenarios graver than the point at which the conventional railway simply gives up. A heavy train can give some extra protection through its strength and weight, but if it hits rock, these properties are likely to cause dangers from forceful rubble. A light cabin (perhaps even with a streamlined, upward-bent nose), able to bounce off the rock, may then be safer.

If a train gets stuck and evacuation is required, the elevator should be useful for heights up to at least 20 meters, although only the upper five meter or so will have stabilizing tubes around the elevator wires. The train should have backup batteries able to power several elevator trips. (The elevator motors should have regenerative braking, and then passenger evacuation will not really require power.)

Evacuation from a tunnel or other awkward place can be done to a closely approaching train, as all trains will have doors in both ends.

Small material destructions will mainly be regarded as a cause of delays. Damages to the overhead power lines (and icing on them) frequently cause delays for railway operations. This will not be a problem with the well protected power line in the beamway. Rail switches on the ground can easily get jammed by debris and snow. Not so with the elevated beam movements. The nineteenth century signalling of trains is disturbed by various ground conditions, but anybody with a little knowledge of modern data transmission will regret the adherence to such old traditions.


The term scalability is used (mostly in connection with information systems) to indicates a system's ability to either handle varying amounts of work in a graceful manner or to be readily enlarged. A design with high scalability will be a more valuable design, because it has a larger market (with both small and large population densities), it is economical under low work loads, and it can work reliably when the work load increases.

Our beamway design is highly scalable, because it can operate in several ways with increasing work load:

  1. A separate elevator module can run alone, unmanned. Each passenger presses a button to indicate the destination, but the elevator moves horizontally to the destination between an ascent and a descent.

  2. As 1, but a (small) wagon is attached for increased capacity.

  3. As 2 (or 1) but manned with a guard/operator/driver. This will generally be required for public transport.

  4. A second wagon, and/or longer wagon(s) may be used.

  5. More is invested in the line for increasing the capacity – the weight capacity and/or the speed. The line may be reinforced with e.g. cables holding up the beam. It may be equipped with maglev or linear motor accelerators for increasing the speed.

Using elevated stations (or side tracks going down towards the ground level) is one solution for increasing the throughput, but this elaboration comes with two caveats:

A bi-directional line, with one beam for each direction, is, as for railways in general, an investment which can increase the line capacity dramatically. But the beamway has an important possibility here: If poles for two beams are used, the second beam may later be mounted and removed again – without leaving any trace in the landscape.

Many monorail designs have been introduced for serving high traffic loads only. Common scalability-reducing designer attitudes are:

Assuming the line has a reserved land area is a serious scalability impediment.


Many different solutions have been accumulated here over several years. It is difficult to state now which are valuable, as research and testing are needed for drawing such conclusions. Some solutions implied in old illustrations are probably outdated – like the use of flanged train wheels. Others are possibilities for future refinement – like the designs requiring special paraphernalia along the track, or whatever is done for obtaining the very highest speeds or getting rid of the beam.

The important points of departure are:

The efficiency of the (partial and full) air cushion hovering schemes should be investigated.


July 23rd 2006: New chapter: The Wheels.

August 3rd 2006: New chapter: Switching.  Contents table with internal links. Emphasis on flangeless wheels.

August 7th 2006: New chapter: Collecting Energy.  Agriculture note under Movable Beams.

September 16th: 2006: New chapter: A 2C Metropolitan System.

February 4th 2008: Modifications in The Beam: turning and twisting half-beams.

February 26th 2008: New chapter: Transversal force control.

December 3rd 2008: New chapters: Crossing Water and The Submerged Floating Tunnel.
Under The Beam: how half-beams can be twisted in the opposite direction, and how brake pads can act upon the interior of a half-beam.

December 18th 2008: Profile picture with description of the double-walled half-beam. Description of placing a motor in each wheel.

December 31st 2008: New chapter: Compressed Air Hovering and Propulsion.

January 9th 2009: At the end of Hovering Train: About partial levitation.

January 19th 2009: New chapters: Introduction and The High-speed Train. (+ small additions many places)

January 26th 2009: New chapter: Beam Irregularities

January 27th 2009: In Compressed Air Hovering and Propulsion: About how the same bogie can be used with wheels or for air cushion use

February 3rd 2009: New chapter: A Universal Bogie

March 1st 2009: New chapter: A lighter Beam?

June 23rd 2009: New suspension design under Suspension without Poles, and drawing of “invisible” suspension unit under The cabins.

January 15th 2010: More internal links & argumentation for a continuous beam.

May 14th 2010: New chapter: No Wheels at all?

May 18th 2010: An additional paragraph appended to No Wheels at all?

May 23rd 2010: Rewritten and expanded Transversal force control.

May 27th 2010: New chapter: Double Action Air Cushion. A paragraph appended in The Linear Motor

August 8th 2010: New chapter: A Compromise Bogie. Removed chapter: Also a Straddling Train? Changes in Double Action Air Cushion: The 4H beam is rejected. The stability of the Double Action Air Cushion is listed as the 6th major system feature in the introduction.

August 20th 2010: Expanded A Compromise Bogie. (Mostly suspension details)

August 30th 2010: New version of the chapter The Beam. New chapter: Scalability.

November 18th 2010: The chapter Building a Beamway is extended.

December 3rd 2010: Revised – many details rewritten. Obsolete features removed or deemphasized.

January 4th 2011: Picture of The Submerged Floating Tunnel included.

January 11th 2011: Important new picture in The Beam. The Beamways Recommendations article is now fully absorbed here. A Conclusion chapter appended. Rewriting various places.

January 31st 2011: New chapter: Tunnels, with picture showing preferred wagon profile and tilt-controlling suspension. In The Suspension: Supplementary suspension (straps/rods).

February 19th 2011: New chapters: Along Roads and Safety. Also various details added/updated. New drawing of a streamlined train.

February 28th 2011: New chapter: Wagon Exchange.

March 9th 2011: New pictures in Along Roads (Beamway positioning) and Suspension without Poles (A-racks).

April 2011: New chapters: Scalability and GTS. Modified power computations in A Compromise Bogie.

October 27th 2011: Two new generalities in the GTS chapter. (Ownership and Vertical movement.)

Copyleft   Olav Næss 2006-2013

These ideas may be used freely by anyone – without patenting, of cause.
But please report important dimensions in such a manner that standards can be established!