Train Lift Mechanics

Bearings, Drive and Sensors

The Preview

In the preview I have already presented an uncut lifting process to you in advance:

The train lift concept has become a functional operating element of the track layout. What is still referred to as the fiddleyard in the track layout for 1 scale will now be replaced by the lift. How things will continue “under the surface” can already be guessed in the preview: At least there is a turnout in front of the lift… more about this in a later post. But now let’s have a look at the details of the lift. If you already know why and why, you can skip the next two sections.

Why a train lift and what is it good for?

Who hasn’t read other posts on the subject yet: This lift is used to lower or raise a model railway train from the ‘rail level’ of the model railway (normally table height) to a lower level. This means that a so-called →station yard can be realized in the lower, invisible area. This would not be possible without the lift, because especially at 1 scale →a very large helix would be required. The lift offers a very space-saving and elegant solution here. As a result, several trains can be invisibly parked in the staging yard and a varied operation is also possible on a small layout. At this point please note again the reference to the contribution linked above (track layout).

What should the lift be able to do?

In the post Concept of an elevator, some requirements for the design have already been listed. These include, for example, the required load capacity, quiet operation and so on. As a reminder here again the sketch with the essential elements:

Concept sketch of the lift with linear guides (blue), threaded spindles (orange), toothbelt (dashed), motor and support structure. The thick black lines at the top and left represent the tracks. A train arrives on the upper track, the track section (carriage) travels down and the train leaves the lift on the lower track level.

I can now proudly say that everything has been successfully implemented. So the lift (→with the corresponding control electronics) can easily be integrated into the automation of the layout. In the end, it is just another special element in the set of automized articles and in the series of landscape modules, respectively.

Technical Data

The following concrete properties have now emerged from the partially only fuzzy requirements placed on the lift:

  • Liftes rail length: 1410 mm
  • Lifting height: 400 mm
  • Drive: 12V wiper motor, approx. 120 W
  • Stroke done by two trapezoidal threaded spindles
  • Transmission and power distribution by toothbelt
  • Sensor technology:
    • Two switch-buttons for the end positions
    • One 10-turn potentiometer for position measurement
    • Two emergency switch-buttons as end stops
  • Duration of a lifting operation: 30 s
  • For a quiet operation: rubber dampers and low speeds

This is what the lift looked like at the time it was installed in the layout:

Installation of the lift

Very nice: Also as a size comparison here the DHG 500 with a few cars can be seen. On the lower right you can recognize an industrial power supply, which supplies the lift with 12V (silver housing). Since 10 A can flow loosely, the power supply is dimensioned accordingly. The lift on this somewhat older picture is not completely finished yet: The rail power supply is not done so far. The cables of the different blocks are still hanging down freely.

The Realization

  • The development time for this mechanical stuff, the →electronics & →software was about 180 hours. This corresponds to about half a year for me. Additional time was spent in the apron with research and some preliminary tests. It should be mentioned that commercial solutions with these requirements practicaly are not being offered.
  • Originally, the mechanics were to be based on wire rope hoists instead of threaded spindles. The basis for this is the fact that ropes can transmit large forces very quietly. However, this turned out to be too yielding: When a heavy train moves onto the carriage, it should not give way in the millimeter range. Although 2 mm thick, Dyneema-like and very little stretchable material was used, the overall elongation was too high due to the rope length. Various required deflection pulleys also contributed to this elasticity. Attempts to actively compensate for the yielding failed due to instabilities in the control system.

The Video

Now a detailed video with all functions is presented. I hope that the function becomes clear. I am always happy about interesting comments, questions of all kinds or even “Likes” on Youtube:

Details about the mechanics

Mechanical guidance

The carriage contains the track section in the lift and is the part which is vertically adjusted by the lifting height. It consists of some chipboard residues, which were processed as a T-profile. For the guidance of this carriage DIY store drawer guides were used as linear bearings. They are not really ideal for fastening… In the end the price was the deciding factor :-). And since the guides are not exposed to any great loads, this is completely sufficient. A few additional holes and nuts that are screwed exactly at the right angle now connect the carriage with the wall via L-angles. Square aluminum profiles serve as intermediate elements. The profiles are connected to the wooden base plate. The limit switches are also fitted here. Two screws bolted to the wall at the upper ends serve as additional support points. Thus also the wall is not pulled too much into damage. That is so far already quite stable, since everything brings along together also a certain own weight.

The Drive

Wiper motors are popular for tasks like this one because they are high-torque, robust and inexpensive. Efficiency does not play a role here. So an →exemplar of this category was allowed to take over the driving role. 12 V and up to 22 Amps are the characteristic data. In practical operation, the current consumption is between 5 and 8 Amps.

The Trapezoidal Threaded Spindles

Trapezoidal threaded spindles are the first choice here. Conventional threaded rods do not run clean enough and friction is very high despite lubrication. This also increases wear. TR 12×3 was chosen as a suitable dimension. The dimensions are still easy to handle and the price is manageable (the required two threaded bushes must also be included here). At the same time it is stable enough to withstand heavy loads (buckling load in fully extended position; vibration stability).

Bearing and drive of one of the two threaded spindles: The belt pulley (black) is glued directly to the bronze-coloured threaded bushing. The bush in turn is placed on the inner ring of a ball bearing. In contrast to many other concepts, the spindle does not rotate here!

The spindles run in the threaded bush and through the ball bearing. In this way, the vertical load is neatly transferred from the dead weight of the carriage and the train including load into the wooden structure. The latter transfers the load to the ground via adjustable feet.

Toothbelt Drive

A toothbelt T5 is used. The large output wheel on the motor shaft with a diameter of almost 80 mm has 48 teeth. The two belt pulleys on the spindles have 22 teeth with a diameter of approx. 32 mm. This results in a transmission ratio of 1 : 2.182.

Drive dimensioning

Together with the pitch 3 (TR 12×3) of the spindles, this means that one revolution of the motor shaft leads to a height difference of approx. 6.5 mm. As you can see in the video, a stroke takes about 30 seconds. With a lifting height of 400 mm this is 13.3 mm/s. So the motor has to do almost exactly two revolutions per second. Essentially, the friction torque of the spindles has to be overcome. This is clearly noticeable despite lubrication. Unfortunately, the required torque multiplies with the mentioned transmission factor and must be taken times two (two spindles are driven at the same time). So the motor has to produce a lot of torque. This explains the choice of the wiper motor and the observed current consumption: Purely mathematically, at 12 V and 8 A 96 W input power result. With an efficiency of perhaps 60 %, we are at 57.6 W pure shaft power. With a speed of 120 rpm, the output torque is 4.6 Nm (cf. calculation →Torque <=> Power). If this value is divided by two (value per spindle) and the belt transmission is taken into account, the result is approx. 1 Nm friction torque per spindle.

Noise Reduction

A basic rule in machine acoustics is that slow-running drives are quieter than fast-running ones. So it is a great advantage that the drive manages with a low speed. Together with the acoustically advantageous toothbelt and the silent spindles, this results in an extremely low noise level. This is →in the video also nice not to hear 🙂 Sometimes the quiet ‘clicks’ of the limit switches are louder than the drive at low speed. So it happens in demonstration mode that a spectator doesn’t even notice the lift and is surprised that after two minutes suddenly a completely different train ‘appears’ from the same corner. That’s how it should be, mission accomplished!

Construction Details

Spindle Distance

The distance between the spindles here is approx. 1000 mm. The value is not critical. I can even imagine a greater support distance. In extreme cases even another spindle would be conceivable. Then the belt run design would be even more challenging 🙂

Wood, Wood, Wood …

For some elements profane chipboard was used. The pictures say it all. Of course, you can make it visually more beautiful and more “sophisticated” in terms of price… Ultimately, the lift is not a “visible component”, I find the present solution to be fully functional and therefore completely sufficient.


End Switches

Two push-buttons provide electrical feedback as to the current position of the slide. Here is an example of an illustration of the upper limit push-button:

Push-button for the upper end stop (blue). The bend in the tab was intentionally introduced.

Each push-button (including the emergency stop push-buttons) is mounted on H-shaped plastic adapters, which allow precise alignment to the desired position. At this point a toast to the invention of the scroll saw 🙂

Emergency Stop Push-Buttons

The emergency stop push-buttons act as the last “lifeline” if there is a malfunction. Analogously to the end push-buttons there are two of them: one for each end position. The emergency stop push-buttons do not differentiate between top and bottom. If one of the two triggers, the entire lift stops and must be reset manually. This safety measure did not yet occur in practical operation (knock on wood!).

Position Sensor

A multi-turn potentiometer with an extra toothbelt serves here as an analog sensor:

Multi-turn potentiometer as position sensor. A separate toothbelt is driven by the carriage and guided over the potentiometer axis. In this way, the position can be measured with sufficient repeat accuracy. The toothbelt runs downwards from the potentiometer and is guided over two deflection rollers before being pulled upwards again. A spring (in the picture center above) tensions the toothbelt. In the lift’s video the complete process is shown

As the name suggests, a multi-turn potentiometer has several turns. This one has 10, which means that it allows 10 revolutions. The reduction of the toothbelt must now take place in such a way that the entire rotation range of the potentiometer is used with one full stroke of the carriage. This means that with a stroke of approx. 400 mm, 10 revolutions must take place. This in turn leads to a belt piece of 40 mm for one revolution and then to a belt wheel diameter of approx. 12.7 mm (D = circumference / pi). The next larger available diameter was used. All further details can be found in →the lift electronics post. The following already here: With an analog-digital resolution of 1024 bits, the position of the slide can be measured with an accuracy of approx. 0.4 mm. This is more than enough as a basis for speed control and would theoretically even be sufficient as a limit switch.

What does the fun cost?

The short answer: Amazingly little 🙂

The somewhat more detailed list (partly estimated, without track material):

  • Chipboard/wood scraps: 10 €
  • Drawer guides 25 €
  • Trapezoidal threaded spindles + 2 bushes: 50 €
  • 2 bearings, 2 belts and 5 pulleys: 50 €.
  • Geared motor 12 V: 10 €
  • Aluminium profiles and standard parts: 20 €
  • Push button and potentiometer, toothbelt and pulley: 15 €.
  • Electronics incl. power amplifier: 30 €
  • Switching power supply 150 W: 80 €

Thus one arrives at a total budget of 290 €. If you look at some of the offers on the Internet, it’s a real bargain 🙂

Operating Experiences

After more than two and a half years of operation and an estimated number of 750 lifts, there were only two incidents:

  • Once a connection on the potentiometer toothbelt had come loose. However, the software immediately recognized this as an error. The lift stopped and went into emergency stop mode (that’s how I first noticed the problem). After the belt was fastened again and the emergency stop was acknowledged, the operation could continue immediately.
  • The centering between the track ends shown in the video were only retrofitted after about one year of operation. Before that, there were occasional derailments, because the carriage was laterally offset relative to the fixed tracks. This led to ‘climbing’ of the loco wheels and in extreme cases to derailment. The problem has not occurred again since the centerings have been added.

Centering: At the top the guide element of the lower siding (aluminum part), at the bottom the centering bolt on the carriage (center of picture, black). The guide ‘catches’ any lateral offset of the bolt and guides it cleanly to the middle. The tracks are then aligned exactly with each other.


If you look at the main objectives of this project:

  • Good function: Sufficient load capacity and accuracy
  • can be integrated into the automatic system (Rocrail)
  • reliable operation
  • quiet operation
  • acceptable price
  • sufficiently fast lifting process

then I can only say: A complete success! Like →in the preview and →also in the video: The lift works very well, I am very satisfied and the work was not in vain. Model building is fun!

Leave a Reply

Your email address will not be published. Required fields are marked *