Belt Drives

What size timing belt do you suggest using for a drive train? Can I get away with a 3/8" .2 pitch or do I need to use something larger?


depends on where in the drive train, our team used that from cim to gearbox, and after gearbox we used 5/8 .25 pitch.

5mm belt pitches (HTD, GT2, T5) and widths of 9mm-15mm seem to be the most common synchronous belts in FIRST. For reference, an average 5mm HTD belt that is 15mm wide has roughly the same working load rating as #35 chain.

What if ?.. you used a 7 in plastic pulley on an 8 inch wheel than installed a .75 pulley on the CIM. This would be a 10 to 1 ratio and You could eliminate the gear box and have a quick robot. How wide should the belt be for this and would it hold up under the stress?

Would you only be using one CIM per side? It seems like it would be difficult to couple two CIMs using this method.

that would also be waaaaaaaay too fast (20 fps).

manufacturers will provide documentation to easily calculate the kinds of loads their products can handle.

Math, it’s great stuff (the calculation for speed is also a quick one).

This would be for a mecanum system which uses one CIM For Each Wheel. I Though it would be a large weight savings but I am not used to working on belt systems.

keep in mind that pulley/sprocket/gear weight increases in a squared fashion with diameter. Even pocketing, a 7" pulley will be HEAVY.

Does that meen a 7 inch plastic spoked timming pully would not be avalable for this aplication?

a .75 pulley sounds awfully small. for our dribbler we had a 1 inch plastic pulley on a .5 output shaft of a fisher price in a gearbox and it broke because there was not enough materiel around the key-stock. a metal pulley might work better though.

Check the manufacturer web sites - there are minimum pulley diameters for each size of belt.


I had noticed this too - that the linear shaft speed of a CIM motor is just about what you want for the linear speed of a wheel’s tread (regardless of wheel diameter). I did spend some time on the McMaster-Carr website and other places and found that belts 5mm width and smaller can handle the small diameter pulley you’d need on the motor shaft. To save weight, you could use 6" wheels with ~5" or larger pulleys. You might have to go to custom pulleys, though, if you have the casting or machining capability. As AdamHeard pointed out, though, even going up to a .75" drive pulley gives a really high wheel speed, probably faster than you want.

It looks on the surface like 5mm wide belt might have the strength to do a 1 CIM/1 wheel combination like you suggest for a mecanum drive. I honestly have no experience with what stresses a drive train actually sees during competition, so I don’t know what the real strength requirement is.

  • Steve

This is great discussion, but the application that my team wants to know about isn’t for a mecanum drive or for directly running off a CIM - we’re looking into alternatives to roller chain that we won’t have to constantly keep re-tensioning and breaking and greasing and all that fun stuff.

Fun question: I’ve heard that belt drives never stretch. Does that mean if you CAD appropriately and use exactly the right belts, you don’t need tensioners other than to more fully engage pulleys? I’ve seen photos of 1717 robots with belts used without tensioners before… Are they a requirement, or just nice to have?

Theoretically, yes, you should be able to design the distance between centers to a correct interval and end up with a belt without tensioners. However, the belt drives I’ve done have all ended up with tensioners. This was just to bulletproof the setup should anything change throughout the course of the season. The nice thing about these tensioners was they did not need to be variable. It was literally a piece of delrin turned to the correct diameter and placed over a standoff that was inline with the belt path. We had little issues with the belts with the tensioners in place. Post-season we ran without tensioners to test the setup and did end up failing a belt (sheered teeth) after 3 or 4 post season events. Luckily we had replacements at the ready and did a quick swap and reinstalled the tensioners.

I would say if your considering a belt setup, order some samples and just play with them yourselves. Use the fall to develop some kind of a prototype using belts. It doesn’t need to be an entire drive system. Maybe even try to do some lifetime testing and see if they do stretch in your applicaation.


When I was in HS I designed two transmissions for a CNC plasma cutter using belts, without tensioning mechanisms. I recently visited my old HS and found that both belts are still tight 5 years later. Very encouraging.

While my team has never used belts for the drive system, we’ve used them on roller systems in 2001 and 2003 and in 2002 for swerve drive steering. None of these had tensioner systems and they worked well.

We were geared for 23 fps in 2008, but that game was suited for nearly-out-of-control speed.

Well I do know that Chris did upload a great white page about belts vs chains. Take a look at that for more information as well. They found that 15 mm belt is a bit better structure wise than chains but the area it takes up may limit a few things.