Every time I try to design a pivoting mechanism like an arm or an intake, I always run into the same question of how to actuate it. I’ve seen all of them work, but could never find a reason why one is better than the other in pivoting arms and other mechanisms. Generally, I’ve heard that belts and chains are “opposite” with gears being a solid middle ground.
Here’s what I’ve generally heard about each system:
Pulleys & Belt
Good for low torque, high speed applications. Most common on flywheels or other sorts of intake rollers.
Simple to replace and very good at their job. Not incredibly hard to tension when needed. However, it’s hard to get a good center-center to get perfect tensioning. They can also be painful to assemble, similar to chains.
Chain
Good for high torque, medium/low speed applications. Common on arms, intake pivots, elevators, and drivetrains.
When an inline tensioner is added, it is very strong and robust. However, it can be hard to maintain/fix, assemble, and fine-tune.
Gears
Commonly used for many applications. Very common on intakes and shooters (to reverse roller direction), drivetrains, and some lower-torque pivots.
Typically easier and simpler to maintain and fix when they break. However, it’s very easy to get backlash in the system if center-center distances aren’t perfect or have extra center. Backlash can also be found in hex-hex interfaces, which could make a system very inconsistent and unreliable.
How does your design team decide what to use for each application on your robot?
For our team we try to stick with gears plus chain as the final reduction. Another good version of this is planetary/cycloid plus chain as the final reduction.
The gears and belts have their place too but it’s just mechanism dependent. If it’s anything high load stick with the chain
In a more serious manner, I’d say you’re pretty on track with your information. The only thing I might add is I think belts can be used for some medium ish torque applications. You could use them to actuate a light arm and they’re also pretty good for elevators if you have some strong belts. Based on my personal intuition I’d stray away from any use of gears directly driving a pivot. That puts a whole lot of load on 1-2 teeth of a gear whereas with belts and chain the load is spread out a lot more evenly. However, if it’s a planetary gearbox direct driving a pivot then you can get away with a bit more (not too much though).
We have pretty much standardized on a final reduction of 25 chain for pivots that need position control, and #35 chain for pivots that need to lift most of the robot weight.
Then belts (and gears) for high speed rollers, intakes, launchers, etc.
As much as I hate dealing with chain because of how much it stretches, getting the grease on my hands just from looking at the chain, and cursed half-links, it’s the only option that lets you have high torque, keep the motor lower, have minimal slip, and can handle shock loads decently well (like without sheering off the teeth from gears).
Chain unless you know what you are doing or your pivot is extremely small. Pivots have extremely high dynamic loads which will damage underbuilt pivots during accelreations and impacts.
Chains spread the load throughout multiple teeth and you can buy #35 chain for extremely high loads. While annoying to use, having your mechanism break is more annoying and the extra strength is worth the hassle.
Gears only transfer torque through 1 tooth unless you by expensive gears and manufacture your robot to extremely tight tolerances, which makes them not good at taking FRC impact loads. There are ways to make gears stronger, but chain will do so but better for lighter.
Belts can be considered if you pick stronger belts than the FRC standards( GT-3, GT carbon, HTD in a pitch higher than 5mm, or extremely wide HTD-5). A standard HTD-5 in 9 or 15mm width can often be too weak and fail via tensile forces or creep and lose tension during usage.
The thrifty cycloidal gearbox is now a thing, it can be considered if 23:1 is enough reduction for your arm, but you will often want much more reduction which will require chain again.
We typically only gears and belts for reductions. Gears for reducers we may want to change the ratio on, and belts for everything else, including final reductions.
Does that include the climber on this year’s robot? The offseason robot I’m designing tries to have a similar climbing system with an algae ground intake. From what I saw in your guys’ CAD was a 25 normal and billet sprocket to pivot the whole climber.
Should this be a #35 chain?
We ended up running #35, since it’s two runs we probably could have gotten by with 25 but we were having chain skipping issues that we eventually found where in the gearbox plates flexing, so it’s possible we could have gone back to 25 after we stiffened things up but we had other places to pull weight so we never did.
Using belts on super lightweight arms can be a solid choice, especially if you go for a belt with a larger tooth profile. You’ll see them used sometimes on spartan arms like 4613 in 2023
If your pivot doesn’t go all the way around, you can use a turnbuckle as a chain tensioner. Additionally, if the run is long enough, you can use wire rope between sets of turnbuckles to reduce weight. @JamesCH95 is more likely to know the specific parts we used/have pictures, but the general idea is to bolt the turnbuckle to the last link of the chain, replacing the chain pin with the bolt.
To expand: the wire rope is stiffer, stronger, and lighter than the chain. The turnbuckle gives you excellent tensioning. It’s a wonderful system if you need to apply torque over a long distance.
we did a WCP CC wrist for this season, and instead of chain driving the rotation we went for belt driven so there was no backlash in the belt system. unfortunately that doesn’t stop backlash from the planetary gearbox so it was still an issue, but less than it would have been with chain.
so like others have said, belt is fine for some medium torque applications, like a wrist, but I wouldn’t trust it for higher loads like a pivoting elevator or something.
also, Vibes is a totally valid decision maker. if it doesn’t work, often you can kinda pivot your design a bit to make it more robust. like in this case there’s no reason we couldn’t have swapped belt and pulley for chain and sprocket. it would require no major changes, just a bit of disassembly and reassembly