Our team is looking into using belts, as opposed to chain or gears, for our drivetrain and we are prototyping it this summer. After a quick look I did not see anything pop out at me on the search that seemed to have what I wanted so I decided that a new thread was in order.
If your team used a belt drive this year, tell us a little bit about it. Any helpful tips or a general description of your drive system would be very appreciated. We understand the basic concepts but if there are any hidden details please let us know.
We did a belt drive this year for our six wheel drive. We set it up with the wheels cantilevered, and the belts captured between two sheet metal C channels. Worked out well. We used 36 tooth 5mm HTD pulleys and 15mm wide belts, which was probably overkill. Could have definitely gone with smaller pulleys if we’d used GT2, but our drive ended up sized perfectly to use one of the FIRST choice HTD belts that was avaliable, so we went with that. Always use the documentation on belt loading from Gates to size your pulleys.
We used timing belt pulley stock from SDP/SI and custom turned our sprockets, which was a great move. Saved us about $100, and allowed us to custom size and broach the pulleys to fit our needs. The center wheel was a double pulley, which was simply an extended length of the pulley stock. We used pieces of laser cut delrin screwed onto the faces of the pulleys to build custom flanges.
Center center distances were calculated using the SDP/SI’s calculator. We didn’t add anything, though we’ve heard of teams adding a few thousandths to this number. Worked great, belts were tight, and we never threw or broke one all season.
One key thing to remember when switching to belts from chain is that you cannot open up a belt like you can a piece of chain. Therefore, you have to think through any and all structural members that run through the belt’s loop, and make sure they can be installed after the belts are in place. Also, if you go with an exact center-center design rather than a tensioner, make sure you think through how you will pull the belt tight for installation. In our design, we installed the corner wheels first, and allowed the center pulley to float with its axle absent. The tension on each side of this pulley cancelled out, allowing us to then slip the center axle in and lock everything together.
Our team has successfully used belts in a variety of applications, including three CIMS on a climber, and 1 CIM on a swerve drive. We used HTD 5mm pitch belts from SDP/SI, but I think McMaster might sell them now. You can save a lot of money making your own pulleys if you have a lot of them, and for low torque applications (like our shooter this year), they can even be 3d printed.
It is useful to have a method of tensioning, even if it is as simple as adding spacers beneath a gearbox. Also, be sure to choose a belt length that is actually made, so don’t use some weird number like 73 teeth. For center to center distances, you can use Inventor, but they don’t have the HTD belt profile in the design accelerator files. I have used SDP/SI for two pulley setups, and Gates Design IQ (free download) for setups with more than 2 pulleys.
We used the base and the kitbot on steroids tips and we had a excellent drive base. I like how the belts were center to center distance and the belts were self centering. The Andy Mark folks did a outstanding job designing the KOP drive base.
As long as you know the ratio you want and the distance between wheels, it’ll tell you all the possible options. As long as the number is with .01" of your goal, you’ll be fine (though it’s better to be .01 under than .01 over). For a for-sure fit effect, couple this with a tensioning system, and you’ll get a belt drive that’ll never fail.
For drive systems there are two ways of going: c-c and tensioning. Tensioning will give you perfect tension. On the other hand c-c can be a lot easier, as you don’t have to make bearing blocks etc. If you do c-c you want to add a couple thou to the c-c to artificially tension the belts. Based on the length of the run somewhere between .003 and .008 thou extra seem adequate.
Looks like it doesn’t work for fractional C-C distances. I don’t use the belt length calculator anyways, as I prefer to calculate C-C based on the pulleys and belt I want to use(Center Distance calculator). The gt2 belts are metric anyways so it is probably better to use the center distance calculator and then plug the metric center distance into CAD.