Box tubing inconsistenies for WCDs

[Travis Covington]

You need to accommodate manufacturing tolerances in all of the parts involved (both extruded tubing, as well as any parts you make yourself). As such, lengthening the shaft some to accommodate tubing width variations, as well as bearing housing manufacturing tolerances is necessary and not really a big deal. This is pretty common. If you want to get really picky, you can add precision shims later to make up for the difference, but again, this is rarely ever worth the effort. Making things "line to line" never works well in the real world as you have already figured out.

[AdamHeard]

For a more thorough answer, we also have to account for tolerance stackup... we just have one less item to add. this really isn't a huge difference.

[DampRobot]

The most I've seen box tubing run over in the 2" dimension is around .003", and about .002" for the 1" dimension. (Inside dimensions are another story.) That's pretty high precision, I doubt most students could get similar precision on a manual lathe between two of those snap ring grooves.

While the box tubing per se is fairly high tolerance, once you start adding in other parts on the shaft, the tolerances really do start to stack up. As Travis said, line to line with something like snap ring grooves tends not to work very well. Instead, it's a good idea to add in a bit of wiggle room in the form of a small adder on the groove-groove dimension.

If you don't need to do two snap ring grooves, my team's used a slightly different solution which soaks up tolerances on the shaft stackup quite well. We put one snap ring on the shaft (this year, it was on the inside) and on the other end of the shaft, we tap a 10-32 hole. A screw in this hole with a few washers retains the wheel, and nicely soaks up any inaccuracies in the box tubing, bearing blocks, etc.

[KrazyCarl92]

Quite honestly, until you're chasing after the last few available percentage points of efficiency in your drive system, tolerances down to the 3rd decimal place digits are ridiculous to make a fuss about. Precision to the hundredths of an inch or order of 0.005" at best would a more realistic goal for most teams when aiming to line up their snap ring grooves and drive components. It's better to have things exact, but at the same time it is not entirely critical to the drive system working.

While tolerances do stack up, adding on a 0.003" MAXIMUM variation (based on what I have seen and heard from others) is not going to have a profound influence when you aim to line up components over the length of your robot to within 0.020". If that 0.003" is the last straw, then you have another 0.017" somewhere in your system that you likely have more control over and would be more easily addressed. Go for the lowest hanging fruit first, then work toward a more elegant and optimized solution later as you chase excellence.

[sanddrag]

We typically guestimate to account for powder coat thickness, then add another 5 thou planned allowance (slop), then add another 20 thou or so per side with the anticipation of using a shimming washer between sprocket and bearing and between wheel and bearing. Also, it won't hurt anything if there's a little extra space between the grooves. If you want to get it just perfect, you can custom machine spacers to exact length to go between the wheels and bearings.

[Chris is me]

Easiest way we have found to account for this in our shaft design without having to think about it at all in advance is to use a snap ring on one side, and a screw + washer on the other (the wheel side).

[AdamHeard]

We've seen deviation of +/-.005" easy on tubing, possibly more.

Originally Posted by Chris is me

Easiest way we have found to account for this in our shaft design without having to think about it at all in advance is to use a snap ring on one side, and a screw + washer on the other (the wheel side).

The issue with this is it will compress the bearings and add friction. The bolt will also loosen over time.

[T^2]

Originally Posted by AdamHeard

The issue with this is it will compress the bearings and add friction. The bolt will also loosen over time.

Would a wave spring rectify this issue?

[Chris is me]

Originally Posted by T^2

Would a wave spring rectify this issue?

If you're careful when you do it, it's really negligible. Just backdrive the shaft before by hand to get a feel for it, line everything up, red loctite the bolt, and tighten down carefully until it's just snug. In our case, it helps that our Colson hubs have a ridge on the edge such that the part of the wheel contacting the bearing just touches the inner race. Get it just tight, then turn by hand and be sure it feels the same resistance wise. It may add a little friction, I guess, but subjectively it hasn't been much.

Adam's right in that it does loosen - check it before every day of competition, more often if you don't use red loctite.

[AdamHeard]

Originally Posted by T^2

Would a wave spring rectify this issue?

No. By the time you get any reasonable amount of preload to keep the bolt from loosening, you're clamping the bearing. Even if you have a spring, the spring will transmit this same force.

A wavespring could be used to take up slop in a bolted system for where the shaft is too long, but not when the shaft is too short as discussed.

The bolt also takes more time to install and remove than other methods.

[sanddrag]

While we're on the topic of retaining wheels, does anyone have a favorite brand/model of snap ring pliers? Ours all slip out of the hole in the ring too easily.

[Cory]

Originally Posted by sanddrag

While we're on the topic of retaining wheels, does anyone have a favorite brand/model of snap ring pliers? Ours all slip out of the hole in the ring too easily.

McMaster 5449A65

[sanddrag]

Originally Posted by Cory

McMaster 5449A65

Not to hijack the thread, but what actual brand is that plier?