I’ve been getting some experience with CADding gearboxes in preparation for the upcoming season. You can check out my most recent one HERE. I have some questions regarding the design and manufacturing of these custom gearboxes. Just for some background, it is designed to be used with a WCP bearing block and bolted into some 1"x2" aluminum extrusion via the holes without any screws in them in the render.
Design:
Do you see any obvious flaws with this design?
What are the little clips that you use to retain the shaft called?
Manufacturing:
How would you go about manufacturing this part on a CNC mill without damaging the vice and given the irregular shape? Soft jaws? Tabs? We have access to a CNC mill but not a router.
What dimension do y’all use for your bearing holes and do you cut them directly on the mill or do you ream them afterward?
The clips can be called many things. I believe they are called retaining rings in McMaster’s catalog. I only see one sprocket. Are you planning to mount this on the back or front wheels? Also, you can clean up mounting by having standoffs coming off of the CIM mounting holes, making the other ones not necessary.
Also, just noticed you are missing holes for the WCP gearbox bearing block. I believe it’s two #10-32 threaded holes horizontal to the output bearing on a 1.875 bolt circle. Otherwise, nice simple gearbox.
Sorry I keep on making additional posts, but every time I look at it, I find new things. One thing I would double check on is the chain run. Right now it looks like the standoffs at the very right and left will interfere with the chain run. If you haven’t already, try modeling your chain clearance somehow in the sketch level.
Yep, they clear. I had to make the standoffs a little thinner to make it fit when I was first designing it. I might even consider going for a bigger sprocket to make sure, though.
For holding something like this in a mill (as opposed to a router), we made ourselves a mini table with tapped holes and a waste board on top that goes in our vice. This allows us to use bolts or step blocks to hold down our plate during machining. It is comprised of a 1/4in aluminum plate with plenty of tapped holes, a waste board (~1/8th) on top, and a 1x1.5 (I think) steel bar underneath, for the vice to hold onto. You can see it in the opening shot of our reveal video this year.
First off, I’m a fan of the 2 CIM + 1 MiniCIM. 4.5:1 is pretty fast though, and you might run into very high current draw: not necessarily a problem so long as you’re aware with it, and deal with it in code.
Can’t tell if you’ve done this yet, but it’s good practice to add .003" or so to the c-c distance between your gears, to account for your machining tolerances and to let the gears turn smoothly.
Sounds like you’ve checked the chain clearing the spacers, but it doesn’t hurt to give yourself some insurance there.
Also, looking at the plates, some of the pockets seem pretty small. Overall the plates are definitely millable though. Tabs could do it, but check what works with your setup.
All in all, great gearbox design! Straightforward and effective. It is worth considering whether custom gearboxes are the best use of your time and resources during build season. Many COTS gearboxes will get the job done. But if you have the resources and machining ability - and importantly, practice doing custom gearboxes beforehand - then go for it.
I personally would support more of the face of the MiniCIM with the gearbox plate. I know there is some debate as to how much this matters, but better safe than sorry, especially with aluminum gear. (I don’t want to turn this thread into another aluminum vs steel gears thread, but I would highly recommend at least using a steel pinion gear as well)
I would also see if you can squeeze a CTRE mag encoder between the CIMs on a 3D printed mount to read off the output shaft.
First post covers it.
To make plates on a mill I have used the method shown at 41 seconds in this video. Its reasonably effective. https://youtu.be/b7qBlwgEZyU?t=41
AndyMark has standardized on the following dimensions, if you can hit that range with your mill that should be all you need. When we water jet plates we ream them afterwards. CIM boss: 0.751 to 0.758
1 1/8 bearing: 1.1238 to 1.1249
7/8 bearing: 0.8738 to 0.8749
All the plates we make in house are either done on our Haas or our router, both of which can hold those tolerances so we don’t do any reaming for production in house. I have not yet visited our water jet parts supplier so I can’t tell you the specifics of how they fixture them. On MOE we made a custom swerve drive plate that we reamed by fixturing and indicating it in a 4 jaw chuck. Maybe someone else on here has some experience with post reaming a gearbox plate.
Nathan oversimplified a bit. We have a Haas VF-4 with the Renishaw probe system. This allows us to get tool measurements within .0001” which I then use to measure our router bits as well. The router is a bit temperamental and is a Shop Sabre brand router. I don’t remember the exact model number off the top of my head.
I said this in another thread for your team - come visit us. We’re like right down the road. I’m happy to help with some of this machining and can answer questions. Plus I need a good excuse to hang out with Leyf and Forrest. Seriously, stop being strangers.
You have three motors there, 2 CIMs and one Mini CIM. So, three motor controllers. But, those motors have different designs and, thus, respond differently to their input from the motor controllers. How do you ensure that they’re all going the same speed?
(Or is that just something you don’t have to worry about? If not, can somebody explain why?)
You can control this somewhat at a software or mechanical level, or you can just let it happen. If you let it go, the RPM difference will cause one of the motors to start being pushed, i.e. acting as a generator. By nature of conservation of energy you won’t get enough current back to blow any fuses.
At the very worst, you’ll lose efficiency if you’re at a free speed. At stall or high load, the slower motor will still be able to help out. In addition, the Mini CIM is essentially designed to be able to be swapped out for a CIM with minimal headaches. The major difference is power.
(At least that’s how I believe it is, let me know if I’m absolutely wrong)
I noticed that your fillet radius is pretty small and in some places there aren’t fillets. I usually prefer a radius of 0.13" so I can have a bit of playroom with a 1/4" cutter. You can certainly afford to leave a bit more material on, it’s not that much weight.