Has anyone noticed that the drive shaft gear is a loose fit on the drive shaft? That is a bad thing, correct? I guess it will wear over time but it should be a pretty tight fit to begin with, right?. We took a toughbox apart (we have been abusing for a few days testing software) and it was a noticeably loose. But when we went to replace it with a new piece, they are loose also (tried 3 separate gears and shafts). It seems to be designed that way!
Please help a poor EE trying to diagnose a mechanical problem.
I’d recommend stepping back, taking a deep breath, and then wandering off and working on more important parts of your robot. The small amount of backlash in that joint isn’t going to make it meaningfully less sound, nor is it likely to affect any feedback code you have. The toughboxes are relatively inexpensive, but reliable transmissions. The small amount of slop there would be of concern if you were running at very high loads that you were rapidly reversing, but it won’t be a problem for 99% of FRC teams. Any teams that would have problems with it would know that re-machining the shafts is your only option. “Roughing” up the shaft will only give you a false sense of accomplishing something. If you’re running the kind of loads where the slop is a problem, a few puny dents from a center punch are either going to get flattened or going to gouge out the inside of the gear. Neither helps you out. If you’re not running the kind of loads where slop is a problem, them you don’t have anything to worry about anyways.
So, short answer, the slop isn’t a problem, just a consequence of the low, low price of free that you got the trannies for. “Roughing” the surface won’t get you any benefit and might actually be harmful.
I was showing one of our mentors the workings of one of our Toughboxes when he noticed the incredible amount of slop in the gearboxes. He suggested shimming them. We used a 7 mil brass shim on one gearbox, but, strangely enough, the other gearbox wouldn’t even fit 4 mil.
This is frustrating because the tolerances on other parts were very tight (e.g., getting the bearing over the output shaft, inserting the roll pin).
OK, I reckon we’ll move on. And I understand that you get what you pay for, however the transmissions are anything but free. But we are talking about semi-ridiculous machining tolerances. It is not even close. It seems counter-intuitive that this is ‘OK’. Any slippage at all with a new mechanism will accelerate the wear on the surfaces, correct? Its akin to using the wrong size wrench on a bolt.
We have the capabilities to make new shafts but its a lot of mill and lathe time we’d rather spend on other things. The purpose of the query was to see if it mattered enough to try and fix it. Thanks for the all the replies. One team wrote “would be of concern if you were running at very high loads that you were rapidly reversing”, the default code (which we are not using) will reverse the robot very quickly and most robots are working with nearly the same load (<140lbs on X wheels) so I’m not sure I get the point. It seems to me that this gear should not be loose. It does not cost anymore to make something the correct size (within reason).
It does in fact cost more. 4140 hex shaft from McMaster has a -.008 tolerance. Any hex broach you’d care to find is going to have a +something to -.001 tolerance or so. So if you’re interested in inexpensive product, you buy the preformed hex shaft, round it off, and worry about something else. If you’re worried about fit issues that only matter at the extremes, you buy larger stock, machine a hex on it, and charge a heck of a lot more for the trouble of having to use 2 machines on a simple shaft.
As to the rapidly reversing heavy loads… I’m talking about a 140 lb robot with 2 CIMs per side and ultra grippy tires driving at a top speed of 15 fps across the field and then instantly throwing the motors into full reverse. Repeatedly. And that’s assuming the robot stays rubber side down. And even then, I think you’re more likely to shear a key, or snap a chain before you ever dent that gearbox. So like I said, not the kind of load commonly encountered by FRC robots, and anyone who’s planning on that kind of loading is probably already looking elsewhere for their trannies.
“I’m talking about a 140 lb robot with 2 CIMs per side and ultra grippy tires driving at a top speed of 15 fps across the field” … thats us we hope but more like 12fps!
We see your point about the tolerances. If we have -.008 on both surfaces, that gets sloppy pretty quickly. And it seems our keys/slots tolerances are contributing the overall slop as well. Adding in the chain tension slop, it is enough to drive the encoders nuts now and then. We have to filter it out.
In mass production tolerances can make a HUGE difference in the cost of the product. The kit gearboxes are probably designed to have only the tolerances that they NEED to have. This can explain why some of the tolerances are so tight and others are not.
I agree with what others have said; not worth worrying about it.
This is a good solid engineering principle. Tolerance errors can add up.
We put hubs from AndyMark onto 8" - Mecanum wheels and they still have slop, but not as much as with chain drive. Also, I have found that after you run the tough boxes for while they tend to mesh a little better and run quieter but the slop remains.
It very important to assemble the tough boxes correctly. There is a star lock washer that holds the little gear in place. Without that the toughboxes will eventually rattle apart inside. Our team forget this washer in ALL 4-toughboxes so we got to assemble/reassemble them multiple times.
