pic: Banebot P60 Failure



One the right is the output shaft and carrier plate for the Final Stage of our Banebot P60 Transmission with a 256:1 Ratio that failed after some rough arm testing (whoops), on the left is a brand new unit. I noticed that there were very few pictures of the P60 failures so I figured I’d post this up.

I believe that this was broken when we were testing our arm modifications. We had switched from a PVC arm to an Aluminum arm with the hopes that the Aluminum would deflect less under load than the PVC. The aluminum arm didn’t deflect a noticable amount so I decided to test the full speed of our arm in both directions with about 5 or 6 rapid direction changes. I’d estimate that the transmission was subject to a minimum 40-45 ft-lbs of torque just prior to failure.

Another thing to look out for is the roll pins that hold the outer ring gear to the front plate. The wall on the roll pin holes bend and lets the flats on the ring gear slip which ends up rounding the whole front of the gear. I can see it already happening to this gearbox.

Install a torque limiting sprocket to take care of some of that shock load at ends of travel if your using chain.

This is copied directly from the Banebots P60 64:1 RS550 page.

Maximum torque recommendation
We recommend maximum torque not exceed 35 ft-lb for all P60 Series Gearboxes

Yep, that’s what happens when you stress things past their design spec.

We stripped our 256:1 today also…

That’s a shame considering how long they take to get…

we did that to a 144-1 p80 once, made quite a lot of rattle. we used 81-1 and then a 3-1 chain reduction to solve this problem (it will take 10 times what we expect.)

The 4 roll pins have already stripped out of the front bearing mount, but I believe that we may be able to press new roll pins in - I’m not sure yet.

Our mechanism is designed so that the gearboxes themselves should never “see” more than 15-20ft-lbs of torque. The majority of the damage done to these gearboxes was done when I was testing an arm fix and went from full speed up to full speed down, something that the operator does not do in competition.

Out of curiosity, did you have the end of the shaft supported?

And can we get a picture of the whole arm?

The end of the shaft is not supported right now, but the point where the sprocket attaches is only about 3/4" from the face of the gearbox. There isn’t really any way for us to support the end of the shaft, and we’re not sure if we need to yet.

I don’t have any pictures of the arm handy but here is a video: http://www.youtube.com/watch?v=s8SR93PZ_Ac

I believe that the gearbox failed during testing of our aluminum lower link fix. The arm is geared to approximately 90* per second max speed, and at that speed I abruptly reversed the arm causing one of the gearboxes to fail. We plan on limiting the speed of the arm once all of the sensors are installed so that this can not happen again. Strangely enough the arm was put through nearly 40 cycles after that test and there was some binding when un-powered, but powered the arm worked fine.

I would modify your system before ship, this isn’t the kind of failure you want to ignore.

High reduction banebots gearboxes are a bad idea, I’d solve it before ship.

If it can happen it will happen. I am with Adam on this one and would redesign it so you can’t see more torque than the transmission is spec’d for. Just because you normally wouldn’t operate in this fashion, if it does happen, I wouldn’t want to be the one begging a team for a spare transmission at competition.

What is the overall ratio? How heavy and long is the arm? Is it counterbalanced?

I’m just curious about the conditions prior to failure. We have a similar implementation using a BB 81:1 gearbox and have beat on it pretty hard over the last two days. We haven’t experienced any problems yet, but if our implementation is similar to your, I may tear the gearbox apart and take a peak inside.

Dustin,

Aren’t you already slowing down the arm in software? All you need is another stage of chain reduction and you’ll dramatically decrease the load on the gearbox, increase the output torque on the arm, and fix every problem you have.

The overall ratio is approximately 819.2:1. The Banebot Transmission itself is a 256:1 ratio, and there is a 3.2:1 ratio after the transmission which drives the arm. The arms themselves are approximately 57" long with a 7lb load on the end. We also have some surgical tubing on the arm so it’s effectively lifting a 1lb load at the end of the arm, and that load is being lifted by two RS775-18’s through two 256:1 transmissions.

We were planning on slowing the arm down in software if we found the need to. We may implement a 2nd reduction after the transmissions, but we’ll need to see if this is physically possible with our system. A more likely solution is to swap the 15 tooth drive sprocket on the transmission to a 10 tooth drive sprocket which would give us an effective 4.8:1 ratio after the transmission.

I just want to say that they did not fail under “normal” use, they failed during a test in which I tried to get them to fail. I cycled our arm at the highest load point from 100% forward to 100% backwards a half dozen times before getting any sort of play in either transmission. One transmission failed, while the other was fine. (The drive sprocket for the arm was not tightened properly during re-assembly which appears to have caused one side to see more force than the other, but I’m not sure).

Note: This picture wasn’t meant to start a discussion about our system, but to provide a visual record of a failed P60.

Our failure point was those spring pins getting torqued off the flats also. All the gears are fine. We removed the spring pins and drilled out the pin holes for larger pins going deeper into the face plate. Then put divots where the pins interface the flats for the pins to seat in. Seems better than new this way and running good now.

I am very interested to hear how this works out over the long run. I wonder if banebots designed these pins to shear off first to protect something more exspensive in the assembly.

I just want to say that they did not fail under “normal” use, they failed during a test in which I tried to get them to fail.

If they can fail in a test, there’s a decent chance they can fail in a real situation, too… Wouldn’t you rather fix it now, instead of hoping these same circumstances can never repeat themselves in competition?

Extra chain reduction and some surgical-tubing counterbalancing could solve your problem. BaneBots gearboxes work much better when you treat them nicely…

You’re right, we are planning on modifying a few things to make the mechanism more “nice” to the P60’s, though we don’t feel that a full on redesign is necessary. There’s already quite a bit of surgical tubing on the arm, and we plan on adding a bit more when we get our next McMaster-Carr order. We’re also considering swapping out the drive sprocket to a 10 tooth which should make the transmissions see about 33% less torque during operation - not to mention that the arm should slow down a bit too.

I’m also going to look into the modification that Swampdude suggested and see if we’re able to do it in our shop.