More things to VR trainer. I spent time with some people and it looks that we will likely proceed with bringing VR trainer to market. The main criteria for design are high quality, very competitive price, small size and range of mechanical and other parameters that should satisfy pros.

We intend to use electric motors for resistance unit since they allow realistic of simulation of uphill, downhill and coasting if motor control is properly implemented. The trainer will also have true power-meter (yes it looks like we've found cheap and good quality solution for that) so it could be used to realistically compare results while doing online/network races workouts, provide spin-scan analysis etc, etc.

Now I have a question for you people.

I have tried few trainers, including Tacx Fortius VR trainer which I own. They all have problem of tire slippage when emulating tough hills. So I plugged bunch of number and formulas into Excel spreadsheet and quickly discovered that combination of human and bicycle can generate insane amounts of torque when staying in low gears. Existing trainers use small diameter shaft and therefore have really small contact area between the shaft and and bicycle tire. Combination of high torque / small shaft diameter leads to tire slippage then

It is not that difficult to increase that contact area by using larger diameter shaft but then there are no existing motors on the market with acceptable combination of price, weight, power and torque.

So one of the potential solutions is electronic gearing. If cyclist uses toll gears on his/her bike then the amount of force generated at shaft - tire contact area is well within a reasonable bounds. It will be possible then to simulate up-shifts and down-shifts by using motor controllers. We can even add some little jerkiness and proper sounds at shift moment for fun and realistic effect. You would not have to tell the controlling software what real gear you are using since combination of bike / motor would have enough sensors to figure that out.

So the advantages of using this approach are:

1. Significant lowering of cost.
2. No tire slippage.
3. Low wear on training tire.
4. Ability to simulate various chainring / cog combinations without really buying and installing all those. For example you can upload elevation profile of particular course to trainer and then "ride it" and select which combination works best for you on that particular course.

The disadvantages are:

1. You would not be using your brifters but small buttons that could be placed on the handlebar in position of your liking.
2. Since you only be using few real gears on your bike it'll probably lead to excessive wear of those most used cogs but that is probably what happens in real life anyways.

Sounds to me like you should spend some of your investment money on pro-rider types as consultants rather than soliciting input from the BF fodder.

We do not want the trainer to be for pros only. Any potential trainer user is our market. I suspect that few people here have spent enough time riding bikes and trainers and they are probably quite capable of answering the question. No need to go to pros and waste money without having couple of working prototypes first.

C'mon people. Can't you give me your opinion if electronic gear simulation cold be acceptable in return for better quality of VR trainer feel?

Good point. Send us the money instead.

How about gear reduction between the motor and the larger diameter shaft?

But I think even with a very large roller, you will have a roller slippage problem.

I think the fake gearshift idea is a bad one.

I understand your points, but... I dunno. Mainly I think it would be difficult to break the habit of shifting. Most riders shift using the brake levers without really thinking. So if they had to use a different method that would be a problem.

One alternative would be to build the device into a dedicated bike. Like a "spin bike" for example; in that case the rider would not have the usual shifters and have no habits to fall into.

Would you kindly provide us with your personal account details so we can transfer the funds? Your humble friend from Nigeria

Yeah, I think you've got the point here. But I just thought that it could be used as an option. If you switch using real gears combination of cadence and speed sensors would steel be able to figure out your real gears and adjust resistance/acceleration appropriately. You would just run a risk of tire slippage on simulating big inclines while using very low gears on your bike

1. Gears are noisy, expensive and reduce the efficiency. If I went this route I'd probably use belt drive systems for motor RPM reduction. It is silent and cheap.

2. I am going to talk to an engineer but the basics here I think that increasing roller diameter will increase contact area. And if normalized pressure between roller and tire kept the same then the friction should grow as a square of shaft diameter. So if we triple roller diameter (easy) then we should get 9 times more friction between tire and the roller.

3. May be it is may be it is not. That is what I am trying to figure out.

Why go through the hassle of a tire-to-roller interface. Put a cassette on your device, remove the rear wheel from the bike and use the bike's chain. Problem solved.

I like this solution. With proper design of the motor coupling and controller, you should be able to match the feel of a real wheel while retaining real shifting.

You're hoping CycleOps, et al isn't reading this forum, right?

Open discussion of product ideas from someone who is (either in fantasy or reality) wanting to build the product is stupid without an NDA.

What ever you do, dependability and relative ease of set up is critical. You should look at CompuTrainer. It is a fine, robust computer trainer.

If I were going to bet the farm on a new product, I would hire Robbie Ventura as a consultant. If all that you can afford is Pcad, then at least you will get is entertaining advice.