Belts for drive train

Belts don’t actually break in most FRC applications - stripping teeth from a belt is the failure mode that is seen most often. We stripped teeth off frequently last year using a 15mm belt on a 24 tooth pulley with a 6" wheels (6 wheel West Coast Drive). This year we used a 30 tooth pulley with a 15mm belt and 6" wheels and only stripped one belt on our P-Bot which sees about 100 hours of drive time. We also later realized that the drive axle was bent which probably did the most damage to the belt.

We have been using belts for a long time. I wouldn’t recommend using a 9mm belt unless your DT is really weak - but that kinda defeat the purpose doesn’t it :slight_smile:

8 teeth per inch of wheel seems reasonable for a 9mm belt - but for a 6 inch wheel that is a 48 tooth pulley and that is huge and you have the potential of beaching yourself on them, or just wearing them out on the floor.

2 Likes

3534 used 9 and 15mm belts this year. No problems at all, with no visible wear over 70+ matches. 124lbs, 8 wheel WCD, no drop. 6" vex traction wheels in centers. Omni’s on corners. 60t belt pulleys were the key with plenty of tooth engagement. Ran 15mm between traction wheels, 9 to Omni’s. Zero maintenance. Would go all 9mm next time.

Doubtful. 24t pulleys for 6’’ wheels is well outside of the rated loadings for 9mm belts. There’s some leeway there because the “official” load ratings are for industrial settings where thousands of hours of service life are expected, but you should not be surprised at belt failures with that setup.

The belts work out to be fairly close to the floor with that rule, yes; I don’t recommend 9mm belts in years where ground clearance might be an issue, for the reason. On flat fields, though (think 2017 or 2014), it’s fine.

We use 15mm HTD belts inside 2"x1.5" tube using custom 24T pulleys. Bearing holes are machined to the exact center-to-center distance for the belt. Pulleys are machined from pulley stock, then hex broached. This drive train has been absolutely bullet-proof all season, and runs buttery-smooth. Only maintenance has been replacing worn wheel tread periodically.

IMAG0901
image

4 Likes

I really want someone to make these pulleys as COTS. I’ve been wanting to run this style for years :frowning:

4 Likes

I’m a fan of the “outside-inside” belt arrangement, myself. Gearboxes on the end of the robot, first set of belts outside of the frame, second set of belts inside the frame. By far the easiest configuration I’ve ever used both to build and maintain.

We’ve been running this with 15mm belts for a couple years and haven’t seen any problems.

My team in 2014 ran 9MM belts on the drivetrain, and probably ripped 6 of them in half before we replaced with chain. That year we had 24 tooth pulleys, 4 inch grippy wheels, 100lbs and a 6 CIM drive geared for 10fps. It was just way too much power and very undersized. If you do belts on DT, definitely make sure they’re well within spec.

Another thing to think about is it is impossible to replace a belt without having full open access to where the belt goes. Unlike chain, which can be put in around bolts and other spacers, belts cannot. So if you’re going to use belts, make sure you are well within spec, otherwise if you have to replace one it will be a nightmare.

We had 4 belts on the drivetrain, but the only one we stripped teeth on was the loosest belt. Even after replacement, that one stripped again.

Even with the failure, the result wasn’t really catastrophic. We were still able to play aggressive defense with the one stripped belt and it didn’t have a noticeable impact for our driver.

Also, do you have a link to load ratings? That would be useful for future designs. I’ve seen tensile strength load ratings, but I haven’t seen tooth loading

548 had lots of fun with belts this season. Haha. But we’ll use them again in the future for sure.

Take a look here for some examples:

Pardon the ignorance on timing belt design as I’ve only worked with them maybe twice in my time but the only rating information I’m able to find and that I have used in the past has been horsepower based ratings. I am unable to find raw data on tensile strength for the Gates HTD 5mm equivalent belts. So I guess I have some questions. Other than a rule of thumb does someone have a source for this tensile strength information or with the HP rating information available how much power (at what RPM) does one rate their drive base. Let’s say three NEO’s per side?

Gates makes it easy when you are designing a 22kW, 125mm wide belt running continuously with a known duty cycle that pules between 5% load and 30% load. Not as applicable with FRC drive characteristics.

There is a catalog on the Gates website with load ratings, IIRC.

Maybe this is what you are looking for from SDP-SI. T15 refers to tooth profile workable tension, T29, 30 on belt tension in transmissions, and T38, 39 on bearing load and belt pull.

Team 395 didn’t fully use the provided equations, but they were a good guideline, along with the 8 tooth/inch of wheel rule of thumb. We ran 9mm belts on 42T plastic VersaPulleys on a flat 8WD (6" wheels) WCD, 4 corner omnis and 4 center blue nitrile traction wheels. 120T belts, two center pulleys powering two wheels. No tensioning beyond precise bearing holes in 2x1 tubing. We experienced no snapped belts or stripped teeth across 2 regionals of playing fairly rigorous defense. Good for us, because they would’ve been a real pain to switch out.

Vexpro has some specs on their drawings for the htd belts

Thank you for pointing that out! I’m so used to going through industrial sources and catalogs like Gates to look for their specs I never noticed VEXPro’s note before (albeit this thread is the first time I have considered looking into belts for FRC drivetrain application). I’m curious where VEXPro got their numbers from and what safety factor they are applying?

9mm defo too small, we broke a belt at every comp we went to with that size

1 Like

Team 4908 has used 2 18t pulleys for 15mm belts inside 2" square tube for the last two seasons without many issues.

I am curious as to the motors and gear ratio you were using. I know of a team that used them with no issues so I am trying to determine what the variables are.

I made a quick comparison based on our (team 48’s) numbers from this past season. We use #35 chain, we are the most comfortable with it and it very rarely fails us in the long run. Using #35 chain and it’s breaking strength as the point of reference and a few other assumptions I have a chart of comparable Factor of Safety ratings.

Assuming we went from 12T #35 sprockets to 36T/15mm wide belts we would lose about 1/3rd of our “comfort zone”. Again, this is all relative based on what we ran in 2019 and would be a topic of discussion if we were to make the switch. How much are we comfortable losing? I don’t know, without better assumptions about heavy defense/shock loads to the drivetrain we’ll have to guess. I’m going to say if the total load of sprockets and chain weigh considerably more (more than ~1.5x) than belts and pulleys on the same drive base I would consider the switch.

(If my numbers or assumptions are way off base please inform me of such, I’m using peak theoretical torque of the NEOs to get the most conservative estimate. Again it’s all relative.)

We used 2 neos per side with shifting GB geared at 8 and 14 ft/s

We used 9mm on our 2 CIM drivetrain with 24 tooth pulleys, we broke one belt in our second competition but much more frequently we tore the teeth off the belt, so it looked okay from the outside until you inspected it and realized it was really falling apart. We switched to #25 chain when we arrived at Champs and the drivers immediately realized how much more responsive the robot was because it wasn’t eating its own belt teeth. I’d recommend at least 15mm

1 Like