I was wondering why you are bothering to pocket the gearbox so much? Since you said this isn’t for a FIRST bot, weight should not be an issue. In which case you are just using machine time you don’t need to.
Also have you or anyone else with the models taken a look at the strength of those side plates? As your pocketing seems to remove quite a bit of material. Now I am not going to venture a guess as if those will hold or not as I don’t know the wall thickness but generally when you remove 80% + of the material from the side the strength is just not there. Either way you or someone who knows what they are doing should look at this design. The same thing goes for your face width of the gears.
Just remember that keeping gearboxes light is nice, but don’t do it at the sacrifice of gearbox performance. If the gearbox will fail it doesn’t matter how much weight you loose because you wont be able to drive.
Losing weight is always a good idea, be it FRC or the real world. But realize there is a trade off. What amount of material can you lose before your assembly is no longer strong enough to do what it needs to? Also, at what point does ease of manufacturing come into play? Im no machinist but I have a hunch that removing that much material will take some time.
What about going to a lighter material than Aluminum? I know that some teams use Lexan (least thats what I think based on my limited knowledge) Maybe look into that. I wouldn’t feel comfortable cutting that much stuff out and leaving it on an FRC bot. Now, this mammoth SHOULDN’T be getting in a pushing contest with another mammoth but the saying, “rather safe than sorry” applies here.
Plus from a safety standpoint covering the whole thing makes sense, yeah optimally its going to be an internal mechanism, but kids have a nasty knack for getting into places they arent supposed to be, I wouldn’t want to be the organization that has to explain to a sobbing mother why her son no longer has a finger. Covering the whole thing in a lighter material may be a good idea anyway.
Also the smaller cims are more powerful. Bigger is not always better.
If you are going to make those on a regular CNC milling machine, then it would be SO much faster to fabricate by making the outside of the plates square. That way, you can just stick the plate in a vise and mill out the bearing holes and pockets (as opposed to mounting it onto a fixture plate, and mounting that inside the milling machine, then having to mill out the pockets and holes, as well as go around the outside to make the curves…).
Also, if the aim is weight savings, I would lighten the gears well before I would start pocketing the side plates to that extent. All it takes is two operations on a milling machine (either CNC or manual with DRO and rotary table) to make these. Depending on the gear size these can save quite a bit of weight.
Theres some good advice on this thread, hope you are seriously considering what they said, rc_cola.
One item not yet covered - if you are going to use a sprocket on the output shaft to drive a chain drive, you should think about the shaft loads that will result from the chain when the motors are at full load and the wheels are stalled. The chain load can be pretty high for a drive with this output torque.
One other thing you might consider - the standoffs that you are using there look pretty skinny. Along with making sure you have adequate strength of the sideplates (as Greg already said), I would make sure you have enough contact area between the standoffs and the side plates.
-Pocketing the gearboxes was because we wanted to try out the CNC and with Cosmos it tells us that the plates are plenty strong enough. They are 7068 plates at 1/4 inch. So the plates are just for fun.
-The mentors were going over the specs and said that the 3 inch cims would be better for a 2 stage reduction.