Forgive me for the bad pictures, please. I was in a hurry.
A bit of backgroud –
Each of our collectors is raised and lowered by a motor/gearbox/timing belt setup. Each collector has two RS-775-12 motors, 2 70:1 Versaplanetary gearboxes and two 24:42 timing pulley reductions. The belt is 90 tooth, 9mm wide from Vex.
One of each of these assemblies (a motor, VP and timing belt) is on each side of the collector. The collector is 11 lbs.
Despite locating the pulleys at the distance suggested by this calculator – http://www.wcproducts.net/how-to-belts/ – there was a pretty notable amount of slack in the belts. Our first tests result in the belt ratcheting heavily.
We added a tensioning system and that seemed to be working well on our practice robot, however the same system isn’t working at all on our competition robot. Eventually, a belt on our practice robot broke.
So, basically, what gives? This is our first time using belts; I have some ideas about what we could have done better, but I’m a bit surprised with how much slack we had in the belts from the get go.
At this point, I think we’re going to switch to chain and see how that goes for us. We’re also using belts in the drive and those have been okay so far, but we haven’t used the drive very much. I’m nervous.
This is our first year using a 9mm belt, no problems so far. Hopefully someone with experience can weigh in and suggest a reason why this could happen? That looks concerning, as we have something very similar.
That failure doesn’t look like a ratcheting failure, it looks more like a pure tension failure, or like something got caught in the belt…
Keep in mind, I’m speaking with very limited belt experience. I hope someone that has more experience will weight in, as this doesn’t look like a place where a belt should fail.
We were using a 15mm wide belt on our launcher that we eventually destroyed due to ratcheting. It looks different than yours and we were a bit surprised how loose they were at exact center distances as well.
We ended up moving to #35 chain and haven’t had any problems, we bend metal now when the arm goes past it’s limit but the chain never has an issue. Sometimes it’s just better to go with what you know, I guess. We still have 9mm wide belts on our butterfly modules and those have been working great, we haven’t been able to really test them since we push the carpet out from under the robot when we try to stall them into a wall.
We’ve checked our calculator against SDP-SI multiple times, it is always within in .0001 before adding tolerance. When done properly, the belt should be tight enough that there is some difficultly in assembly. In all of testing where both pulleys are located on the same plate, the calculator yields these results (9.5305")
It’s difficult to see exactly in the pictures, but it looks like the pulleys are not being located on the same part, but across multiple plates. This can cause issue’s depending on your manufacturing tolerances. On 1323 this year we had two belts that were located across three plates, and we did notice the belt was looser than normal. However, it is still tight enough not to ratchet. I would recommend to first try to measure the actual center distance on the robot. You could do this with a big caliper, or if you have the ability, you could make one plate with the exact center distance and compare it to the robot.
As far as the belt breaking, the loading on this belt does seem rather high for a 9mm belt. Your reduction has a max torque of 932in/lb, whereas an average 2 CIM drivetrain in low gear has about 650in/lb max torque. I would definitely recommend switching to a 15mm belt if you can. Some more info about how/when the belt broke would be helpful here.
We’ve been using HTD belts and the belt calculator on 1323 for two years now and have been very happy with them. The only time we have had a belt failure was when we cut a 9mm belt in half and then used screws to join it back together. Since then we avoid cutting belts in half, and try use 15mm wherever we know its going to be a high load.
Hopefully this helps you out, let us know if you find anything else out. As always you can contact us a [email protected] or [email protected].
The pulleys are all on the same part, though it’s probably flexible enough to explain where some of the slack is coming from; but we’re seeing more slack than I would’ve expected, even then. Tomorrow, we’ll try to laser cut a piece to force the versaplanetary shaft and larger pulley to the correct center-center distance and see what that looks like.
What am I missing with respect to the maximum loading?
RS-775-12 has a stall torque of .4315 Nm and free speed of 7300 RPM. 70:1 gearbox reduction followed by a 24:42 pulley reduction is 122.5:1 overall. Stall torque is 52.86 Nm or 467.85 in*lbf. Twice that is 935.7, but that is distributed over two 9mm belts (one on each side). There’s one motor/gearbox/belt per side. Sorry if that was unclear. Only one of the belts failed.
It’s pretty clear that the belt failed due to tension – but the tension we put onto the belt was the only thing we could do to prevent it from ratcheting; that’s what is most frustrating.
Thanks for the insight. Hopefully chain will be a bit more forgiving in our design and we can worry about more interesting problems.
Has the intake been moved several times by hand? Does it take a significant amount of force to break the internal friction of the VP that drives the tilt? That’s a huge lever arm, I would not be surprised to see much higher loads being put on that intake when it’s being moved by hand, or even changing direction quickly.
9mm belts might be a little small here on their own, but one on each side oughta be enough. Is it possible that you’re driving one side of the intake tilt faster than the other, and it’s torsionally binding briefly?
In the past, we have had far less success with belt reductions, especially near the end of a gear train, than we have had with 1:1 exact center setups. We had a belt fail in a similar fashion on our 2011 drive gearbox when we tried to use it for additional reduction. Our drive belts have held up much better.
Not saying that’s the source of your problems, but it certainly makes things harder. As long as you used a reasonable sized pulley and a reasonable width of belt on your drivetrain, I wouldn’t worry about it.
You also want to check and make sure in a 2-belt system like you have on each side of your intake - both sides need the exact same tension. Any tension difference will exacerbate your problem. The fact that just one belt failed seems to indicate that this could be the problem.
If you’re sure you’re at the proper center and your belts are ratcheting, then the logical next step to me is to step up to 15mm belts for extra engagement.
The collectors are quite easy to move by hand and, under power, are being moved with something like 50 or 60% of full power. One side of the mechanism is, on its own, capable of producing enough torque to lift the entire thing, so it could be possible that one of them is doing significantly more work than the other. We’ll try to measure current draw on each motor while we operate it.
Unfortunately, we can’t go to 15mm belts in this space. There’s not enough room for us to bump up the thickness of one of the pulleys.
Thanks, everyone. Lessons learned – now we’ll do what we can to salvage the mechanism.
Out of further curiosity, is there somewhere I can find data bout the maximum tensile load for these belts? It wasn’t immediately apparent on the Vex or WCP websites.
