It’s a WIP and is missing most hardware but the base idea is done.
Not quite all belt- the first stage is gears. 
In any case, this is a very interesting concept. I really like your compact execution of the idea.
The one thing that I’m worried about: why 9mm wide 3mm pitch GT3 over 15mm wide 5mm pitch? Even using 2 parallel 9mm belts (which may face issues with not lining up), this looks like it may not be strong enough.
I thought this type of gearbox would be unworkably large, but the first stage gear takes up most of the reduction so it works out.
The flipped cylinder arrangement looks a lot like 192’s gearbox this year.
What are the reduction ratios?
SDP doesn’t make belts smaller then 40 tech so using 5mm GT3 would mean spacing everything out too much
What do you mean by the belts not lining up?
And yeah, I took a bit of inspiration from 192 
I had the reductions written down somewhere but I think it’s something like 10:1 and 20:1.
Yeah, bit curious why you left the first stage gears when you’ve used them everywhere else. It seems to be the reverse of optimal.
Could you post the tooth counts of each stage and any math you’ve done to confirm you’ll have enough power capacity with the belts.
Also need to think lot harder about that inverse pneumatic. I know this is a rough CAD but what you have could potentially kill the cylinder, way to flimsy.
Otherwise I like how it’s turned out.
EDIT: by belts not lining up I think he means if you’re not carful how you join the pulleys together one of the two belts might not fit into the teeth correctly.
EDIT2: pneumatic is less flimsy if you have another one the other side. Still could be stronger.
For your second stage pulleys, they are constrained in rotation by the hex bore. Let’s say you accidentally broach one so that the tip of the hex lines up with a groove, and the one next to it such that the flat of the hex lines up with a groove.
At this point the two pulleys do not perfectly match up in their rotations. This wouldn’t be a problem, only for their mating pulleys you have the same problem- the teeth may not match up.
This means that you could face some major indexing issues that would screw up the meshes on at least 1 pulley at a time.
The way to fix that would be to use a jig to align the pulleys before broaching/machining.
Ah, I see.
Considering that there’ll be three holes for joining the two sets of large pulleys, I don’t see a problem with those. If I do end up having this made I’ll probably use a jig for the driving pulleys.
The spacing worked better to have a single large gear for the first stage. It’s thinner than pulleys and I’m not concerned with the width of the gearbox.
TBH I haven’t done any math for the strength of the belts. Again, I didn’t intend this to be a production gearbox and more an experiment/proof of concept/CAD practice. I’ll post the tooth totals though. I think it’s 18-50, 18-30, and then either 20-50 or 30-40.
What about the paramedics are flimsy? I can always add a support on the back of them.
Edit: I realize that the pictures pneumatic is orientated wrong. The fittings should be either vertical up or horizontal to the outside of the gearbox.
It’s less the pneumatic itself and more of the connection with the shifting shaft. This is helped a lot by having a second pneumatic on the other side, but thats not ideal. With only one pneumatic theres a huge moment on plate connecting the pneumatic to the shaft. That will want to bend both the plate and the pneumatic shaft and generally push everything out of alignment. It’s very important that that connection and whole assembly be very stiff. Once again though, if there are two pneumatics then that moment is balanced and it’s less of a problem.
The pneumatic layout on the GRT gearbox this year was designed for just one pneumatic and designed to deal with this moment. There are two on there because we were having a bit of trouble shifting in a few extreme conditions so added the second for safety, there is improvement to be made to get back to one.
The same problem still exists.
This has killed me before, so I’m always careful about overconstraining now lol.
To avoid this problem, you could try using timing belt pulley stock instead of just pulleys. That would be a guaranteed fit. As it is, you are machining them anyway.
Just curious, what’d you guys do to get it to work smoothly (for the most part) win one pneumatic?
We were actually thinking of that when I presented the design to the rest of the team. It would also let us run custom width belts if they’re not too expensive.
Oh wow, interesting design!
Apparently three days ago I went on and designed a similar gearbox: inverted, two-piston shifting using a dog, except that the thing runs on gears entirely. Here’s a picture:
Regarding how to reduce the number of pistons to one, I’m not so entirely sure, but to reduce the bending force to the shifting plate, the piston cannot be offset from the shaft. However, by doing so, you have to consider that the wheel is also going on the output shaft, and if the piston sticks out too much, it will hit. Also, the other issue that would come up would be connecting the piston to the shifting rod. Tried that once, and I hated it. You could add a third plate for spacing like what 254 did with their gearbox this year. Then piston-wheel collisions wouldn’t happen anymore.
Just some things that I thought of, having designed 3+ of these inverted gearbox. Hope this helps!
Not quite sure what you mean by a third plate or piston-wheel collisions. The cylinders are contained inside the gearbox so they won’t be hitting any wheels.
Anyways, I like your design! I want to try doing the ribbing between mounting/clearance holes but it’s never worked out well for me.
For it to work smoothy the goal is to have the moment supported by something other than the pneumatic shaft and shifting shaft. The strength of the connecting plate isn’t a huge concern, it can be made stiff, the joints are the problems. There are a few different ways to do this. The option chosen for the gearbox this year involved flipping the pneumatic, having the plate mounted to the back of the pneumatic and have the whole cylinder move inside a plastic bushing that took the load. This can be seen pretty clearly with more examination of the posted gearbox picture.
