This is a gearbox I’ve been working on, obviously without any spacers holding up the inner two plates. Total length (R face of R plate to L of center motor) 9.5", total weight 12.1lb with motors. Pictured with 28:30 and 18:40, with the gears on the output shaft being modified for the dog shifter on the Vex WCP DS gearbox. Major problems I can see are attaching motors to the plates (fixed by offsetting inner-mid and mid-outer spacers) and the use of 2 pneumatics (1 more than strictly necessary). Gears between the CIM pinions are 60T.
A few questions/suggestions:
- Is there a reason you made the idler gears live axle instead of dead axle? VexPro sells 60t bearing bore gears, and it would probably save you some weight and space.
- How thick is the plate connecting the pneumatics with the shifter shaft? With such a far distance between loads, it looks thin enough that it may bend under the force of the pistons. You may consider either thickening the plate or replacing it with a more rigid profile (e.g. L, C, rectangle). To check if it will work or bend, this is a great place to start with FEA modeling.
- What is that thing between the middle and outer plate to the left of the output shaft? It looks like it could be an encoder. If it is, I’d recommend not putting it there. Having a significant amount of your output torque gooing through a small pinion sounds like an unnecessary risk. You can put the encoder on a driven wheel or on one of the CIMs using a CIMcoder (either AndyMark’s or Munchskull’s).
- What is the white thing in front of the middle CIM? What is it’s purpose?
- Thanks, that would be helpful; would a hex/keyed shaft be bored or would a round just be press-fit?
- 1/16", could be moved up to 1/8" fairly trivially (this means, probably, deeper cuts on the non-visible face of the mid plate which the pneumatics sit in). I’d considered a rectangle, but opted for this shape for ease of assembly. What would an L or C shape look like here, just going up and over the hex shafts [edit: or is that L and C brackets?]?
- Are you talking about the thing that’s visually above the dog shifter? That’s a plate to hold a bearing that holds the shaft on which the low gear idler sits, this has to be done because the bearing (OD .875) can’t be fit in the outer plate for some reason of interference or another (a reason which may not exist anymore?). Both idlers sit on cantilevered .375 hex shafts.
- That’s a remnant of an earlier version of the design where there was a single piston aligned with the output shaft so the center CIM had to be moved back; it can be removed and I just didn’t notice.
Thanks for the comments.
• I’m not exactly sure what your question is here. A dead axle setup has a static round axle (often a bolt), and the gear sits on a bearing, which rotates around the axle.
• I meant to possibly use an L, C, or rectangular profile rather than sheet aluminum. Sheet metal is prone to flex when you put forces on it in the direction of the normal to the plane (like here). Replacing the sheet with a profile should help reduce bending if that is going to be an issue.
• Yeah that’s what I was talking about. Obviously you know the intricacies of your design better than I do, but it seems to me like you should be able to mirror the right side without the idler gear on the left. That would definitely make the design more efficient and remove a failure point. In any case, you don’t want to have cantilevered idlers. The gears can easily get slightly misaligned, which can seriously affect efficiency and increase the chance of destroying the gear.
I’ve realized I can’t use a dead axle here anyways as the 60T are coaxial with/drive the 28T and 18Ts. I’ll probably change to a C, yes; and the inclusion of those idlers seems to’ve just been a mistake in conceptualizing (I’d came to the conclusion that the output gears would jam without the idlers but now I see they won’t) so I’ll also remove those.
How much would a gearbox like this cost compared to a store bought gearbox such as an andy mark evo shifter.
The biggest variable here is what machining capability you have. If you can make those plates yourself from sheet, you’re probably going to come out fairly close to even - ahead on the plates, but losing some ground due to all the duplicated parts necessary to put the CIMs in a line, definitely falling behind on the second cylinder. If you have to pay a shop to produce it, expect the cost to go way up.
Yeah, this is intended to be made on a CNC mill. For cost there’s also the use of 2 pneumatics, I’ll have to try and figure out how I can check if just using one is viable.
Here it is now, mildly updated on AriMB’s advice (and still incomplete, of course):
The spacer on the center CIM was removed, then reintroduced, for reasons of clearance with whatever’s holding the dog pin. I think spacing of the gears around the dog is a quarter inch or so off; I may have to have a custom version of the dog pin made on the lathe if it’s just that that’s too short.
FEA would be nice, but this seems like a pretty simple 3-point bending problem to me. You could very easily do the Solid Mechanics to figure this out if you were so inclined. FEA is really nebulous sometimes and easy to mess up in my (limited) experience, so I always start with a hand calc anyway.
Also, I am curious if there is a better way to do this that only relies on one cylinder. I could picture some sort of lever doing the work, where the shifter shaft is in the middle between the fulcrum and the cylinder.
You could get away without a set of the green and pink gears. You may actually have trouble lining up the gear teeth and hex bores during assembly depending on if the relative rotational clocking of hex bores to gear teeth are consistent.
There are 3 bearings supporting your shafts with the yellow, green, and pink gears. If everything doesn’t line up just right or the shafts aren’t quite strait this can cause binding and excessive loads and wear on your bearings.
Sure, as a third year MechE undergrad student I could do the Solid Mechanics. But for high school students without the prerequisite knowledge, I figured it’s easier to just plug the numbers into a computer and let it do the work. Simple FEAs like this are pretty easy to set up and understand*, whereas calculating area moments of inertia for various profiles and understanding how to use them to calculate beam deflection is 2 semesters of courses at my college.
*At least from my experience with the Solidworks Simulation Add-On
Also, I am curious if there is a better way to do this that only relies on one cylinder. I could picture some sort of lever doing the work, where the shifter shaft is in the middle between the fulcrum and the cylinder.
This was considered, it was just figured that that would be too difficult manufacturing-wise.
You could get away without a set of the green and pink gears. You may actually have trouble lining up the gear teeth and hex bores during assembly depending on if the relative rotational clocking of hex bores to gear teeth are consistent.
There are 3 bearings supporting your shafts with the yellow, green, and pink gears. If everything doesn’t line up just right or the shafts aren’t quite strait this can cause binding and excessive loads and wear on your bearings.
How would I get all three CIMs to drive the output shaft in that case (or would the CIM on one side just help drive the mid CIM?)? I did see interference in Solidworks, but just figured we could manage to make it fit – now I know it’s angle of the hex bore, though, thanks.
I’ll probably just removed the mid-plate bearings, in that case (which lets us pocket more)?
–
I messed around a bit with one of the simulation packages, wasn’t really sse what I was looking for.
Since that last set of images, there’ve been a set of printed parts to hold the shifter rod’s bearing, as well as standoffs and new holes to hold them.
Now try this gearbox with the NEO motors!