pic: Belt Drive Chassis Take 2

I have reconsidered how to run the belt to get more contact area with the drive pulley. I am trying to get the belt inside the box beam. Which someone noted will be harder to change the belt because of the drive axle. Which is a good point. Even still I would like your input. Now one idler is fixed and the other is sliding. I still am wondering if it should step up to a belt with a larger pitch (other then .2) Thanks.

This path looks strange but I don’t see why it wouldn’t work. If you’re really pushing it, you can fit 2 separate timing belts in a 2x1 tube. Here’s 1625’s scarily similar prototype: http://www.chiefdelphi.com/forums/showthread.php?t=86668

I don’t know how you have modeled your belt in inventor (whether it includes the teeth in the width or not). If you have not included the teeth in the thickness of your belt I’d caution how close the teeth would be to each other near the idler left of the center wheel.

Also worth noting here that specing a belt to the exact length needed may be a bit more difficult because the belt is following an irregular path. Shouldn’t be too hard to figure out, but you may have to end up rounding your belt size up or down and relying on the idlers to tension.


That’s a 2"x1.5" tube Chris

With a toothed belt (or chain), I see problems arising when there are multiple, non-continuous contact patches with the pulley (like with the center wheel in the configuration shown).

For the sake of argument, say the top part of the belt is fully engaged with the teeth at the top of the center pulley. Clearly, you also want the other contact patch to be fully engaged with the teeth of the center pulley. Let’s say your pully has 40 teeth, and 10 of them are engaged with the top of the belt. Then there are 10 unengaged teeth (the gap between where the belts contact the pulley), then 10 teeth engaged with the bottom of the belt. In this configuration, there must be an integer number of teeth in the belt between where it leaves the top of the center pulley and returns to the bottom (since there is an integer number of teeth on the pulley between the same two points*).

How do you ensure this? You either design for the exact right belt length, or you add adjustment to your system. You already have a tensioner, but, depending on belt pitch, you may need more than one dimension of adjustment in order to find both a belt length and belt tension that work. Failing that, I could see (best case) uneven loading of the belt and (worst case) belts skipping out of their pulleys.

I hope that all made sense.

  • There could of course be, say, 9.5 teeth between contact patches, in which case the requirement generalizes to X.5 teeth in the belt, etc.

The other mentor is concerned about the same thing. I am not sure that I completely follow you however. I am under the impression that if a belt has a 5mm pitch it is only a matter of turning a pulley (also with a 5mm pitch) to the correct orientation. And if that pulley has a 5mm pitch then the belt should match at all points, regardless of the path in between. But who knows, you answer sounds a lot better then mine. I have never been much of a math wiz.

Have you experienced a issue like this or are you just making a well educated guess (theory)?

I am not sure either, I just specified my belt in the Design Accelerator. Do you know if that takes the teeth into account? Also I have a belt specified for the application already. The length works with the tensioner.

Measure the thickness of the belt as Inventor is showing it and cross reference that with the thickness of the belt w/ and w/o teeth as per the manufacturer. You should get your answer pretty quickly that way.


This is a real issue with any toothed belt or chain if you attempt to engage different locations of the same sprocket.

Ok. That is a good image. Thank you. I have a idea to work around that. Thanks.

I have been trying to figure out how to do that. What size belt were you using?
Did you have to make custom pulleys?

My concern is the moving tensioner. Any time the load on the belt gets high enough, the whole belt will move the movable tensioner, and there will be slack somewhere in the belt.

Somewhat like chains, enough slack and the belt will slip over the sprockets. Very UNlike chain, teeth will pop off the belt when this happens.

Don’t ask how we know this. In Aim High, we used a belt that was subject to high and variable loads, and had a spring-loaded tensioner. You do the math.

as long as the belt has an even number of teeth, and the distance between the center to the two outer pulleys is exactly equal, the teeth will line up on top and bottom.

oh, and the number of teeth on the center pully must also be even

We used 172 tooth 9mm wide Gates 5mm GT2 belts. We did not use custom pulleys. Our pulleys were 19 tooth GT2 pulleys.

Were you happy withe the 9mm wide belts? I have been thinking about going to the 15mm wide 5mm pitch belt, but they are huge.

The TENSION required to transmit the driving force will be a function of degrees of wrap around the cog not the number of teeth engaged. This is because driving force creates a force that lifts the belt away from the cog. It will take less belt tension to transmit a given force for a single segment wrapped twice as far then the 2 segments shown. This is assuming the tooth phasing problem mentioned above is solved. One more note. I have never seen this configuration used or recommended. It is a clever thought, but likely wont yield the hoped for result.

When you evaluate variations like this always look at the forces generated and if EXTRA structure has to be added to deal with those forces.

I believe that this holds only as long as the center pulley has an even number of teeth.

that was in the edit, i realized that moments after submitting the reply.

Have you looked at the loading on your drive, to know what size belt you need? The Gates site has several design tools to help you with the evaluation. You might be surprised at the capability of the 9mm belts.

Also, is there a reason that you don’t want to use two belts, instead of a single belt? It is a much simpler design, but I don’t know what design constraints you are working with. Just curious.

The main constraint is that the team wants to keep the belts inside the AL tube - right now it’s sized at 3x1.5 inches I think? I’m pretty sure it can be done with a 9mm belt.