Just some quick and dirty CAD that would probably need a little work. Idea is to be able to drive a mechanism (intake) forward, but when hit (i.e. run into feeder station wall ) there is some give in the reverse direction before the motor. Keeps things from exploding.
This is a reduction on a jackshaft. Magenta hub on the gear side is the input ( or the hex shaft, or the gear itself (which can be omitted and replaced with a short piece of maxspline)) is the input - sprocket output. When viewed from the gear side end-on clockwise would be forward.
There are a pair of driving “dog” pins (orange) that are held preloaded clockwise to the driven surface by a torsion spring (green). Forward the pins are already engaged (purple curved slot). In reverse (counterclockwise) if there is enough “fighting” between the motor/gearbox control loops & inertia vs the force from ramming the intake into the wall the torsion spring will be overcome and the pins slide in the curved slots to the back>
Omitted in the model are some fasteners and a shaft collar/clip on either side to capture the hubs (magenta). There would probably need to be a little more work done stabilizing this thing on the shaft (no symmetric spacers (purple) unfortunately )
This can be made with a drill press, dremel, and a 3dprinter
Similar in concept, two years ago we improvised a friction clutch for our intake. Driven by a Bosch Seat Motor, all we had to do was cut a groove for an e-clip and drill out the center of a sprocket - everything else is COTS. The bolt on the end could be tightened to apply more pressure, and the acetal spacer acted as a “spring”, with the two shaft collars providing the friction against the sprocket to transfer the torque. This let the intake move between its set points without slipping, but could then easily slip up when pushed by a ball coming in or slip down if run into a wall. Add a through bore encoder on the driven shaft, and a PID loop takes care of the rest, making it set and forget!
Yeah, i was toying with how to do a friction clutch on the cheap.
Just a note, what I have above is just a way to add some give to a the system. I would probably prefer a polycarbonate intake and add the give in on the arms themselves. But sometimes packaging is a concern, thus the cad above.
In the past we have used a ball clutch. Was team made on a drill press. We used stack wave springs. A search will find that many use a compression spring. I prefer the wave springs. pre-tensioning of the wave can adjust the slip torque.
I love this stuff where you can modify a COTS item or two with common shop equipment, add in a few things from McMaster or similar, and have a functional piece of hardware for $40 in parts and an hour or two of time.
So while I have been doing robotics for a little bit now, I just can’t wrap my head around this. Are you able to explain a little more on how it works and where this is applicable? This seems like a cool concept and I can sorta understand how it works and just would like to understand it more.
The sprocket and gear can spin relative to each other, for about 120° of motion (angle of the curved slot)
Each side has a driver pin (orange ) that corresponded with the curved slot in the other side (purple).
The torsion spring (green) pushes against both purple pieces making them want to keep the driver pins at the end of the slot.
When spinning clockwise (viewed end on from the gear side) the pins are already at the end of the slot, therefore supply all of the torque to the output sprocket.
When back driven (sprocket side is spun counter clockwise (when viewed from the gear side end on) ) the pins are not pushing against the end of the slot and can travel the length of the slot (120°). The torsion spring is wound up as this happens.
This allows the output the take a hit up to 120° of travel before the pins travel the full length of the slot and hit the far end. Only at this point is there a major shock load traveling back through the gear train to the motor.
When the load that wound the spring up is gone the pieces spin relative to each other and return to the starting configuration as the torsion spring returns to minimum energy.
I love the idea! If you only need a little bit of compliance with relatively low loads or just early in the gear train I bet a printed or cast elastomer spider coupling thing would work.
I was curious on what an implementation would look like with JUSTCOTS, no modification at all, not even drilling out holes. This obviousally turned out physically larger than the Rev counterpart.
Andymark had the best COTS for this,
just a pair of their hubs (~$20 total)
some bearings (complete overkill in this situation there are other options; a few bucks),
an input sproket/gear/pulley ($10ish),
fasteners (a few dollars),
a piece of hex shaft (should be sitting around)
and a pair of springs (cost varies).
So total could be as little as $40.
Same general idea as before, input is the sprocket, out is the intake actuation arm (just a piece of 3/16" polycarb). If intake arm is pushed backwards the springs extend until the long # 10 bolts make contact. If intake input pushes forward the force goes in compression on the extension springs.
The arm has a lot of mechanical advantage, so likely you would need much stronger springs than pictured. The through bore REV encoder I tossed on there as the hex shaft in this example directly measures that arm location, in controls you wound need to compare the motor encoder position vs the arm to see of there is deflection and then choose what to do in that scenario (drive backwards, coast, etc).
This was the gear used to lift our 2023 intake with the intent described by @Skyehawk. It drove positive in one direction and springy in the other. This is what it looks like inside:
In 2022 we designed our intake so the pneumatic pistons were retracted when the intake was deployed and extended when retracted. Seems backward a bit but when you run the intake into a wall the pistons will extend and not get damaged. With the pistons opposite the piston rods bend before compressing into the cylinders. We had damaged quite a few pistons in past years past but not a one in 2022
Oh for sure, I am a big fan of using pneumatics here. Designed correctly they offer a lot without being particularly vulnerable (designed so that retracted= intake extended and hitting a wall doesn’t result in over centering the mechanism).
We seem to be moving further and further away from pneumatics after FIRST outlawed off board compressors and we got the extra PDP ports - that loss is a shame.
I’m hoping we can get away from pneumatics this year if possible. Our team is still kinda young experience wise. We lost all our experienced builders and programmers in 2022 and basically had to start from scratch last year
Different tools that excel in different ways. Pneumatics trade packaging in one area for packaging in another.
That new REV linear actuator is interesting, but it is geared so aggressively I struggle to see how it can outperform pneumatics for 2-position use-cases once packaging comes into play with a heavy motor on it.
it’s basically a DART and those were used very rarely. Some of that came down to what the games of that era required (wrong place wrong time).
