Gunn Robotics Team 192’s 2016 competition gearbox. 3 stage reduction, 2 speed ball shifter with inverted pneumatic actuation. Gears from VexPro, belts and pulleys from SDP-SI. Adjusted speeds of 13.5 and 6.0 ft/sec. Width without CIM motors 2.67 inches.
Looks very compact! I have many questions:
- Why the use of the cim-cim pulley? Would that cause the cim shaft to bend sooner?
- How is your shifting cylinder so tiny/compact!?
- What is the weight without cims?
- Why 3 cims over just 2?
Amazing as usual!
The nylocks at the bottom on the red piece don’t appear to have any thread engagement with the actual nylon. I could be wrong but that’s what it looks like to me.
Other than that, amazing machining!
Wow that is looking quite intense, can’t wait to see it in action at the Wisconsin Regional!
Beautiful as always.
I didn’t realize you could use timing belts without the flanges like that. I knew they (SDP/SI, etc.) sold them without flanges, but I saw they also sold flanges, so I assumed you put them on yourself. Are there any design rules about not using flanges? I didn’t see anything in the GT3 design manual, but perhaps I missed it.
Also, I would argue that this isn’t a gearbox until we can see the other side.
The CIM to CIM pulley is the best compromise to achieve the most compact pulley cluster. That is to say, with that layout, the pulley reductions take up the width of 2 belts, while allowing sufficient wrap on all pulleys to transmit power. The layout does put a little bit more load on the CIM’s shafts, but its negligible given that the outer motor is being pulled in 2 directions.
The shifting cluster size is optimized by mounting the actuators (Bimba Flat-I) backwards, with the piston arm attaching to the back plate. When the pistons actuates, the bodies of the pistons come out and bring the linkage out with it to pull the shifter shaft.
Weight without CIMs is 2.01 lbs, about 913 grams.
3 CIMs of course have more power than 2 CIMs, and the big thing that that power helps with is capability during pushing matches/when approaching stall.
I have applied thread-locking compound to it, so they aren’t going anywhere. Just wanted a little purple accent!
So the belt issues are actually very interesting. We actually machined the larger pulley ourselves, out of pulley stock, so that we could optimize dimensions, counter bore it and such.
The lack of flanges comes from several rounds of testing. When the belts are run at a reduced tension, they tend to drift around (move laterally) quite a lot less. Because we are running GT3 instead of GT2 belts and in such an arrangement where they get very good wrap around the pulleys, they do not need to be run at very high tension to avoid skipping (something that we and many others have had to do). As a consequence of all of this, even without flanges, the belts stay generally where they are (also note the slot so that a flange can be fitted at least in between the belts).
I’ll get you some more pictures in a few days as well, if you need assurance that it really can drive the robot!
First, whoa! that’s some nice nice engineering.
Second, thanks for the GT3 tip. As I have said elsewhere, I have not made the transition to belts yet (from chain and gears – I’m old school like that), but I can see the writing on the wall. It’s nice to be able to learn on you dime.
Third, more questions.
I am always puzzled when folks tell me a fps and don’t give me the size of the wheel or the ratio from the motor to the wheel. What do I do with that? If you have 3ft drive wheels it means one thing, if you are going with 2" wheels it’s a completely different beast.
So… what size wheels to you have? And what are the ratio from motor to the wheel?
Also, is that hex shaft peeking out go to a direct drive live shaft (that drives the wheels) or do you have another belt/chain/gear stage between the wheel axle and the hex output shaft?
Also, can you just break it down to the gearbox I’m looking at? What are the ratios from the motor shafts to the output shaft?
Great job. Nice bit of work there.
Dr. Joe J.
The three major design updates of the gearbox, the inverse pneumatics, the new bearings, and the removal of the serpentine belt, have been in the works since mid '14. Mihir has done a great job figuring out the best way to do the pneumatics and incorporating the changes into a three stage for this years game.
Belts walk on the pulleys due to the spiral of tensile members, they’re not perpendicular to the pulleys axis. The higher the tension in the belts the more they walk. This use to be a problem but as Mihir mentioned the tension can now be much lower. For this design the critical area is where the belts would run against each other, flanges may be required if we were to move them closer together. There are design rules about flanges in one of gates manuals. I’d have to look though them again to find which one. Either way the cim pulleys are double flanged properly constraining the belt.
This isn’t actually the ideal belt configuration loading wise. However the ideal configuration has some manufacturing concerns that are currently being worked but not implemented yet. Things for the future.
Do you mean inside? You can already see all sides of the gearbox.
Alright. This years bot runs 6 inch wheels, so those speeds from earlier are for 6 inch wheels.
The reduction within the gearboxes is 18.7:1 and 8.4:1. There is an addition reduction in the pulleys out to the main drive wheels (1.2:1).
The hex shaft you see on the back fits the drive pulley of the transmission, which goes out to the main drive wheels via more belts.
So the reductions step by step: 18:50 in the pulley; 14:30 first stage; 24:34 high gear and 14:44 low gear; finally 18:22 in the pulleys of the wheels.