Banebot Transmission Issues

[Tom Bottiglieri]

Karthik,

Thanks for backing up our hesitation to use these gearboxes. It seems like everyone who's used these things before thinks they are garbage..

[Gdeaver]

Planetary gear boxes in similar form factor to the banes transmissions are used for motion control in industrial applications. They are extremely robust and typically last for years running under heavy continuous loads. The problem is this quality costs some serious money. I believe the entry price for the small ones is around 300$. Beyond what most teams want to invest in transmissions. The banesbot trans are affordable and something has to be given up to get in their price range. The first time I got my hands on the 56 trans I took it apart all the way. After that inspection, our team made the decision to spend part of our minuscule budget on 2 Andy mark transmissions. I don't think our team will regret it. I've seen some team with monster traction wheels and the banesbot transmissions. They are setting them selves up for problems. Used with the KOP wheel may be acceptable.

We may want precision planetary gear heads with electronically commuted brushless dc motors and fully programmable intelligent speed controllers and a micro itx robot controller with a high level development environment. But, our budgets allow for brushed dc motors, spur gear boxes, dumb speed controllers, PIC based microcontroller robot controller, and easy-c - mplab development environment.

[Joe Johnson]

Originally Posted by squirrel

The quick reversing of direction can be avoided with careful programming and control setup, and I have a feeling that this is much more likely the cause of the problem than the 5 ft lbs of torque.

It sure is nice to have this early failure report, so we know to be careful with the BB gearmotors, and have contingency plans in mind using either the Globe or window motors.

The last joint double D* on the Banebots 56mm Trans has given me heartburn since kickoff when the joint first caught my attention (actually I am worried about the same joint on the 42mm & 36mm too but only at the higher ratio -- read more here).

I have done some calculations, some FEA's and finally this past week some tests. I am still working on my "final report" but here is my initial report. The joint fails under static loading when the input to the 12:1 gearbox is about 3X the 12V stall torque of the CIM motor. The joint fails under back and forth cycling at about 2X CIMstall.

From my point of view, the joint as shipped is not strong enough to be used with the 2 CIM adaptor from BB. It seems like it should be okay for use with 1 CIM with the 12:1 ratio but I am not sure because I usually estimate the dynamic loading to be twice the static loading. That would make the 12:1 gearbox right on the edge of acceptable. BUT, in this case, with teams rapidly switching from forward to reverse, I am not sure that a factor of 2X is enough.

The 56mm carrier plates are pretty soft. They are Rockwell A hardness of about 46 -- corresponding to a yield stress of about 64,000psi. It is not too hard to get a material with a yield stress of about 2-3X that number.

I think that that is probably where I am going to go with my recommendation. More to come.

Joe J.

*In many ways, double D's are not very effective torque transmitters, they cause stress risers and concentrate stresses on small sections of the joint. Splines are better because the distribute the stress to more material but not as easy to make.

[MrForbes]

Thanks for the analysis....I was referring to applications for the small Banbots gearmotors, not the 56mm version. Manipulators can be set up to have gradual acceleration, there should not be much requirement for quick back-and-forth motion, hopefully.

I agree that in driving applications the back-and-forth motion is pretty much unavoidable, so your single CIM per transmission recommendation makes good sense.

[JB987]

We were going to use a FP with a BB 42mm, 256:1 mated by chain with an additional 6:1 reduction to lift our arm (less than 3 lbs.)...our prototype has worked well but??? Would addition of gas spring reduce some of the shock and help us reduce our concern over these BB's...or should we start over?

[MrForbes]

we were planning on 8:1 chain with a 64:1 or 128:1 BB gearbox, and either the BB motor or the weak Mabuchi motor. Arm weight is similar to yours. We are planning on using a gas spring to take most of the weight of the arm.....and we are also planning on being able to start over and put a window motor or two on it to replace the gas spring and BB stuff if it doesn't work out.

[Jonathan Norris]

Now you just had to scare me Karthik... We ordered a pair of these early last week and have yet to receive them. With our troubles last year of using the FP in the 36mm gearbox's made for only a RS-385 motor under a low load application, I should have expected this. I was hoping these larger gearbox's would be able to handle the power of a FP motor. We will do some testing when we get ours in and see what we come up with... but now I am scared of these gearbox's...

[Joe Johnson]

I am not saying teams should panic. I believe the BB transmissions are still a great bit of engineering. BUT, like all things they have limits.

I believe that the higher ratios are even useful, but more for speed reduction not torque increase. The output joint is the same for every ratio. If the torque gets too high it will break that joint.

Now to your questions:
Will a gas strut or other counter balance help? Yes it can but it is not a panacea. Counter balance is just flat out a good idea. Teams should do that regardless of what motors they choose.

Should you toss out the BB design and start over with a window motor? This is a tough question. The window motors are great motors too but they only have about 20W of power available. If you are doing some significant work with the motor, you will have to do it about 1/5 as fast with a window motor than with a BB motor. If you are using the higher ratio but you never stall the motor you are probably going to be just fine (try to limit impact loading if you can).

In rough numbers, I project that the 42mm gearbox will fail with repeated cycling loads of +/-350in-lbs. I project that the 36mm gearbox will fail with repeated cycling loads of +/-100in-lbs. If you can design your mechanism to keep your loads below these (with a safety margin that lets you sleep at night), then I think you should be okay.

Joe J.

[MrForbes]

Thanks again, you have pretty much mirrored my thoughts.

[ZZII 527]

Thanks for the warnings everyone. We have an FP motor with 42mm, 256:1 BB gearbox installed on our arm and an additional 7.2:1 chain reduction. So far, there have been no problems, but we are definitely going to look at the internals tomorrow.

I think many people have said it already: impulsive loading is the killer. One thing I can think of that might make a big difference: chain backlash. If on a direction reversal, your mechanism has significant momentum before the chain engages, the stresses will be enormous. Preloading your mechanism so that the chain is always in tension is always a good idea, but in light of the new gearbox info I think it will be even more important.

[DonRotolo]

Boy am I glad that we chose to use four KOP trannies instead of two with our 4 CIMs in the drivetrain. We also ran in each motor/trans combination without load for a few hours - they definitely loosened up after that and drew nearly half the current of when we first started.

Our design is such that a change now would be nearly impossible, so we're going with the KOP trannies - but we might just disassemble them, clean them out and refresh the grease.

The small 64:1 trans will be sued for a light load, but we ran that in as well.

Counterbalance was not somehting we considered, but in retrospect it should've been obvious - thanks Joe J for reminding us.

Don

[CrosbyAR]

Thanks all for warning everyone. Now everyone can spend time preemtively strengthening and modifying untill finding out later that we trashed perhaps an important component.

-Alex

[Joe Johnson]

One think I should have mentioned in my "DON'T PANIC" message above is that if you designed the gear ratio of your arm/mechanism right in the first place, you are probably never going to see any problems.

I have recommended that teams design the ratio such that the highest "normal" load the arm is expected to put on the motor would be 1/5 to 1/4 of the stall torque. This is good practice for a lot of reasons, but to my usual reasons I can now add one more: The Banebot Transmission will not see anything like the stall torque of the motor (assuming your don't run your arm into end stops or the ground or whatever).

For what it's worth.

Joe J.

[MrForbes]

Originally Posted by Joe Johnson

This is good practice for a lot of reasons, but to my usual reasons I can now add one more: The Banebot Transmission will not see anything like the stall torque of the motor (assuming your don't run your arm into end stops or the ground or whatever).

limit switches are a good thing!

[Joel J]

Originally Posted by Joe Johnson

One think I should have mentioned in my "DON'T PANIC" message above is that if you designed the gear ratio of your arm/mechanism right in the first place, you are probably never going to see any problems.

I have recommended that teams design the ratio such that the highest "normal" load the arm is expected to put on the motor would be 1/5 to 1/4 of the stall torque. This is good practice for a lot of reasons, but to my usual reasons I can now add one more: The Banebot Transmission will not see anything like the stall torque of the motor (assuming your don't run your arm into end stops or the ground or whatever).

For what it's worth.

Joe J.

I don't understand this.. you mean a Don't Panic for the fisher price, or for the BaneBots?