pic: Inverted CIM Gearbox V3

Still CAD-ing away! I was going to stop after 15 gearboxes, but I came across a similar design and I just HAD to do it :stuck_out_tongue:

Third inverted CIM gearbox made, which allows for more electronics space, or for whatever needs to go inside the frame perimeter. Still iffy about the fact that there are three CIMs for this one due to brownouts, but I have some feeling that I could get away with it, given that number of motors for other parts are limited, and there is no motor quantity restriction.

Weight: 11.40 lbs
High Speed (adjusted): 18.06 fps
Low Speed (adjusted): 6.02 fps

Files, including STEP files can be found in my GrabCAD folder under GBX-116.
Link: https://workbench.grabcad.com/workbench/projects/gcOhfXOehYETdalk4HNJ_3A0HceTxrt6Dv4ytxedJppYkO#/space/gcy-RpuR7fueMhPxpys8KUh7yF1NwcTEWAlStxFm6c3ypj

Now I’m really going to halt working on gearboxes (with the exception of another one I’m working with someone else on the team) and do something else :smiley:

Have teams actually gotten away with cantilevering gearbox gears like this on competition machines? Maybe it’s just me, but it seems like a REALLY bad idea, and I’ve been seeing more and more of it lately. :ahh:

You will never get a perfect mesh all the way around a loop of gears. You may even mash teeth depending on tolerances and how things line up.

I would also be concerned about the cantilevered gears, since they do transmit torque, and it seems they are only riding on one bearing.

Design Comment

I think the previous poster has the same thought, but your center CIM gear between the two other gears is asking for trouble. There will be some manufacturing tolerances and it will mesh well on one side and poorly on the other side and probably shred the gear teeth. No Fun.

Since you seem to like designing gearboxes - what about one where you are trying to keep CG low by keeping the CIMs at or near the same distance from the ground that the drive axle is? Probably not the most useful as you are only lowering the CIM a few inches, but would be fun to see what you come up with!

SolidWorks Comment

I like that you included your SW files, but you did not include them all. This is a common enough problem that SolidWorks has a feature for it called “Pack and Go” under the File menu. It allows you to take an entire assembly and copy all the parts. You can flatten them to a single folder (recommended) and even ZIP them up, all in one step. Great for sharing whole assemblies while keeping them in SW formats (instead of STEP).


An easy fix (can’t tell if it’s already implemented) would be to make the gearbox slightly asymmetrical so the center CIM pinion only meshes with one of the two gears next to it.

Edit: Rather than relying on the screws to position the CIM, why not use the boss on the front of the CIM to position the motor with more precision by shrinking the current clearance hole?

What about it seems like a bad idea?
EDIT: I just noticed theres only one bearing. This is a bad idea. If there was a second bearing, it would likely be okay for the following reasons.

Cantilevering the gear has no effect on torsional loads, and the gear/shaft has no significant axial loading, so we only need to worry about radial loading on shaft and bearings.

Three CIMs have a combined stall torque of around 1000 in-oz, and pinions have a pitch radius of around .3 inches, so the tangential force at the pitch circle is around 3000 oz or 190 lbs force. FRC gears have a pressure angle of 14.5 degrees, so the radial load is 190 x tan 14.5, which is 50 lbs.

50 lbs is less than a cantilevered wheel will experience on a WCD and these shafts do not commonly bend.

I am more worried about the ~25 in-lb torque on the bearing (gear cantilevered .5" from the bearing). It appears there is not a second bearing to fix this, but I haven’t loaded up the CAD so I could be wrong. Perhaps one solution is to rigidly mount the shaft to the gearbox plate and run one of vexpro’s bearing bore gears directly on a bearing, so that the gear is centered on the bearing.

When calculating reaction loading of a shaft you can’t ignore the tangental load. Assuming the tangental and radial load is what you calculated, the gear meshes in only one place, and the gear is in a static condition, the reaction load on the shaft would be sqrt(190^2+50^2). However this is a simplified analysis and this gearbox has gears with multiple meshes that need to be considered. Also, some of these shafts are smaller than the .5in hex that is commonly used on WCD wheels.

I’m not saying the cantilever will be a problem necessarily, I haven’t done the math, just that the loading is potentially higher.

I also totally agree with everybody so far, that a single bearing (assuming it’s a common FRC bearing not rated for moment loads) is totally insufficient for any of the shafts.

I’m not sure if there’s a reasonable way to do this, but if you could move the pancake cylinder to the other side of the gearbox, you could get the wheels a lot closer to the edge of the robot.

How would that get the wheels closer to the edge of the robot? As far as I can tell this gearbox is intended to be used on a WCD set up which already is as close as the wheels can get to the edge of a robot.

Looking at the render, I thought the wheel was just between the gearbox plates. If the wheel is by the “wire” end of the CIMs, then consider this modified so that moving the cylinder would cause the module would use up less space inside the robot, which I understand to be the main point of inverted CIM.

Oh ok, makes more sense now. The way the gearbox is set up, I don’t think it’s feasible at all to move the cylinder. It would either have to be put on the output shaft of the wheel on the other side and That wouldn’t be great, or be put on the gear just above it and rework all of the gearing in the gearbox. With that option, It might be able to be done, but I don’t think you would be able to get the geometry to work out to avoid the side rail and the wheel on the output shaft. If that was done, I would be impressed. Your idea would save some additional room in the robot and would be a pretty cool gearbox.

I haven’t worked much with shifting gearboxes, so maybe this is a silly question, but could the shifting be moved up to the output of the cluster axle, rather than the input of the wheel axle? With the wheel all the way out by the ends of the CIMs, this looks like there would be enough room for the pancake on the other side. This would obviously be a higher-speed, lower torque shift, so perhaps a different shift mechanism would be in order.

A possible fix for the cantilever gear would be to have a bearing in the gear riding on a “dead axle”- a screw with a spacer on it going into a tapped plate.

The only issue with that like I said earlier would be interfering with anything on that side of the gearbox such as the tube and the wheel as well as the limited supply of COTS ball shifter gears and dog gears, unless you want to custom make your own dog gears or ball shifter gears.

That wouldn’t fix the fact that it is still cantilevered. There would need to be a supporting plate on the other side of the gear to reduce the kinds of loads a cantilevered gear sees vs a non cantilevered one.

It would reduce the torsional load on the bearing to nearly zero. A dead shaft rigidly mounted to the gearbox plate can handle high loads that are cantilevered half an inch or less. The bearing moves to the same plane as the gear, much like the gears engaged by the dog/balls in the shifting stage.