A quick recap

I thought before getting too far into the next steps of the build, I’d quickly describe what I’m aiming for..

As mentioned in a previous post here, the v1 build has been pretty successful – I’ve been using it now for the best part of a year with no real issues. Still, the nagging feeling remains that I’d like something with sufficient accuracy that I don’t need to rely on an external power meter, which is where the v2 build comes in.

Looking at the unit as a whole, I think it’s helpful to break the overall resistance into two separate areas. Firstly, there is the resistance of the un-braked roller assembly (rolling resistance of the tire, bearing drag etc). Secondly, there is the deliberately induced resistance from the eddy current brake.

Modelling the un-braked rolling resistance

I am assuming that it won’t be sufficient to simply take a series of measurements from test rides and use them as a basis for the power requirements. The rolling resistance is liable to change with temperature, tyre pressure, and clamping pressure (or rider weight, depending on the design of the trainer frame). Therefore, I believe I’ll need to add an element of dynamic calibration through a spin-down process.

I think the following steps are required;

  • Establish the moment of inertia for the roller assembly
  • Instrument the roller for speed (using a hall effect sensor)
  • Measure the deceleration through a series of spin down tests

From knowing the speed and the moment of inertia, I can derive the stored kinetic energy in the roller assembly. Measuring the rate of deceleration should then allow me to obtain the rolling resistance in watts, and calculate a speed/resistance curve.

Another limitation with the v1 trainer was that the resistance maps assumed a steady speed. By measuring the speed more accurately at the roller, I should also be able to factor any acceleration/deceleration into the power model.

Measuring the braking force

With the above (hopefully) taken care of, the next stage will be accurately measuring the braking force applied through the eddy current brake.

This would once again be susceptible to changes in temperature during operation. As well as affecting the electrical resistance & the current flow in the electromagnet coils, it would also alter the electrical resistance of the aluminium disk & therefore the strength of the induced eddy currents.

I think in this case, the best approach would be to actually measure the force applied through the use of a strain gauge on the electromagnet assembly – and yes, this would also be temperature dependant, but at least it’s only in one place.

Next steps

That’s the plan anyway, as with all these things, it generally becomes clearer to me as I’m working through the process!

Next post will cover calculating the moment of inertia..


Brake disk

The new aluminium eddy current brake disk is machined and mounted on the roller.


Just for a change, the next step will be some maths instead of machining! I want to model the power curve for the un-braked roller, which I think I can do by calculating the moment of inertia, and performing some spin down tests..


That’s the new flywheel machined and mounted on the roller. Have spun it up for a quick top-speed test, and all seems to be running smoothly ;)


Next job is a new aluminium brake disk, which I’m hoping to get completed within a week or two. The disk assembly will be very simple for now, just to keep things moving along, though I already have some grander ideas for a future version..

Roller finished


Well, that took me a bit longer than expected, but the roller is finally finished. (I made a bit of a hash of the first thread, but managed to save it).

Next job will be turning a flywheel with a matching taper. I have a couple of spare evenings next week, so fingers crossed will get it done then..

Work in progress

Since my last update, I’ve made up a new base, and made a start on the new roller..

The base was made by gluing up and shaping a stack of MDF. I’d had an issue with the previous version, in that the front bolts for the pillow block bearings overlapped with the position of the bolt for attaching the base to the rest of the trainer (if keeping the original geometry between trainer arms and roller position). To avoid this, I’ve tilted the whole assembly forward by 30 degrees.


I’ve also started machining the roller from 50mm EN1A leaded steel. So far, the ends have been turned down to 25.4mm for the bearings. Next job is to get on with machining the tapers and threads for attaching the flywheel and brake assemblies.


New year – new toy

It had become clear that replacing the entire roller/resistance unit would require more machining accuracy than I was capable of with the wood lathe and pillar drill. So I’ve recently treated myself to a small engineering lathe (please excuse the shoddy mobile phone picture).


Now, it hasn’t entirely escaped me that I’ve just spent more on a lathe than I would have done on an ergotrainer! So I guess now this project is more about scratching an itch than it is about budget..

Am just in the process of gluing up a new MDF mounting for the pillow block bearings, then will set about machining a new roller.

New testing platform – partial success

Since my last post, I’ve wound & fitted some new electromagnets, and made some test runs on the trainer..


These are very similar to the previous builds, made from 1/2″ mild steel rod. I have drilled and tapped them at one end for attaching to a bracket. I’ve also added some nylon disks, cut from 40mm rod, just to hold the wire in place a little better.


This time I used 28 swg magnet wire. Unfortunately I ran out of patience when trying to count the turns by hand, so in the end just used the lathe to spool the wire on quickly. Each magnet took just under half a 250 meter roll, and has a total resistance of approx 20 ohms.


The bracket was made from a section of rolled steel, bent into a U shape using a vice. This was then drilled to accept the magnets, and for screwing to the trainer base. The dimensions were then gently tweaked with a lump hammer, until I got a 4mm gap between the installed magnets (leaving 0.5mm each side of the aluminium disk).



As always, some good news and some bad news.

I must admit I was a bit anxious before the first test, seeing as how the previous testing on the Tacx platform had been such a damp squib. However, the good news was that this was unfounded – as I was achieving several hundred watts of resistance from these initial runs – result!

The bad news was twofold. Firstly the bracket has too much flex, and as I would increase the current, the attraction between the two magnets was pulling them closer together,  and into the brake disk (as you can see from the scoring on the disk in the last picture). This became the limiting factor for the maximum braking force I could dial in.

The second issue was vibration, courtesy of my junk flywheel. This was the limiting factor for the maximum speed I could ride. Although it was sort of bearable for quick test runs, no way I could actually train on this setup as it is.

But nevertheless, am treating this as mostly a success. To make further progress, I’m going to have to machine some parts with much better precision. Watch this space ;)