We like belts. They’re light, quiet and don’t really require any special tools if you do the center to center distances correctly. Chain needs a chain break, messing with links, tensioning can be fun depending on the application. 15mm belts are the way to go, have never had to change a 15mm belt in a drive train. I’ve seen plenty of 9mm belts break though.
Do you use a master link in your drivetrain, or a solid link? We worry about popping master links when going backwards, so I’m curious if that is inherent to the link or indicative of other design issues.
I don’t think we will ever go back to chain.
1712 uses #35 chain.
In our early drivetrains, back when kitbots were relatively new and there were far more limited COTS options, we used “pillow block” bearings from McMaster-Carr. There were definitely issues with maintaining proper tension on these, and even occassional throwing of chains.
In the age of the IFI c-channel kitbots, we continued to use #35 chain, will less issues (although significant maintenance time should we have to remove/replace a wheel on that kitbot design).
Since the release of the VersaFrame system, we have continued to use #35 chain, with basically no issues whatsoever. We have also phased out (almost) all master links and half links during this time, which has helped improve reliability.
Until recently, we lacked the ability to create accurate c-to-c cuts, which limited our ability to design around belts in a drivetrain. Additionally, we prefer to have as much of our needed drivetrain inventory on-hand prior to the start of build season, which is very achievable with chain (adding/removing links to adjust for different wheel placements) and somewhat less so with belts (having to buy a variety of belt lengths for different wheel placements). For a long time, we were very hesitant to switch to #25 chain due to issues with alignment and spacing. In the age of 3D printed spacers and spacers available from VEX and AndyMark, we could probably make the switch rather easily today. But we also don’t really see a need to, as we’re comfortable allocating that tiny bit of more weight to our drive to support #35 chain.
2102, we use chain. If a belt needs to be replaced, we’d basically have to disassemble the robot to get a new one in there because it’s a closed loop. If a chain needs to be replaced, it’s still a pain, but the fact that you can disconnect the ends and relink with a master link means we don’t need to take the whole thing apart. Thankfully we’ve had a pretty solid drive train design for the past few years and we haven’t had to worry about it.
I will say, there are no good master link tools for size 25 chain, and often involves us fiddling with needle-nose pliers for a good 10 minutes if it’s in a tight spot.
However, we’ve had lots of issues with these types of pulleys. The flanges are press fit on, and after an event or two, they will just pop off. We can’t have that going on in our drive train.
However, we use belts in most arms/manipulators for weight reasons, and those pieces also tend to be easier to service.
I’ve used all three types of drives: Belts, Chains, and Gears.
This past year was my first experience with a custom belt-driven drive system. The main benefit to belts (and the reason we chose them) is weight savings compared to chains or gears. That said, the tradeoff is that your drive system will tend to be wider (leaving less room inside the robot for other things) as belts tend to take more space than either of the other two options (assuming you’re using 15mm belts, of course. It’s not as bad with 9mm, but you’ll tend to break them more). The other downside to belts is they can be a pain to do maintenance on, since the only way to replace them is generally to remove the sprockets from BOTH drive and driven shafts. You also have to keep stock of the EXACT length belts you need, since you can’t exactly make a custom length belt at an event (or rather, you probably shouldn’t try to).
Chains have the advantage of being easier to work with, and are generally more forgiving to slight spacing errors. This makes them great for teams trying to make a custom drive system or other mechanism, but that don’t have the expertise to get their hole spacing exact. Chains can also be removed and attached without necessarily having to remove any of the sprockets (if you have a master link or an appropriate chain tool), and generally take less space than a comparable belt system. The downsides to chain is it does stretch out over time, requiring either tensioning or replacement; it also (in my experience) tends to break more frequently than belts. As far as weight goes, it generally falls in the middle between belts and gears. I’m partial to using chain (#25 chain, specifically) for mechanisms, since it allows for easy and rapid design modification, but if your team is capable of precision machining, they really aren’t necessary for most drive systems anymore
Gear-drives have a few advantages over the other two; for starters, they’re significantly easier to do maintenance on (and require far less of it). The lack of any tension in a gear-drive means you can simply pull the dead-axle shafts on your wheels and they drop right out without affecting the rest of the system, it also gives them a slight efficiency advantage. Gears also allow you to build the narrowest drive systems by far, as the gears only require ~0.5" of space regardless of how many wheels you’re using (assuming you’re using COTS FRC gears and not some custom solution). The obvious downside to gear-drives is their weight, due to all of the wheel and idler gears (and their associated shafts and bearings); this can be mitigated to a large extent (as gears can often be pocketed dramatically without weakening them too much), but they’ll still tend to be heavier than the other two solutions. Also, because your wheel spacing is determined by the size of your gears (and the size of your wheel gears is limited by the size of the wheel), you either need a LOT of wheels, a lot of idlers, or a combination of large wheels and large idlers. In general I like these kinds of drive systems for defense-heavy games (that don’t require a lot of heavy mechanisms) , or for games with unusual floor geometry where you don’t want to use “Tank Treads”.
Only three? What about friction drives? Direct* drives? Shaft drives?
So many more options we can play with!
*for purposes of this discussion, I’m counting direct driving off of gearbox output shafts as direct driving
Ok, I’ll give you the direct drive one, I forgot about Mecanum drives for a second there (though technically, they are still “gear-drives” unless you’re running them straight off the motor).
I’m not sure I would count “friction drives” as something that would be commonly (or ever) used in FRC. I’m curious how you define “shaft drives” though; if we’re just talking something like a WCD where you have a live axle on all wheels, I would include that under the umbrella of whatever category of item is driving the shafts (IE, belts or chains).
A shaft drive would be a shaft running parallel to the plane of the wheels, linking the wheels on each side together. It would use bevel gears or similar. Here’s one of the more applicable examples in bicycles.
That’s pretty neat, though I suspect the reductions you typically get on those type of gears (not to mention the cost), would probably preclude them from FRC drive system use, no?
Though, that said, I actually had a concept drive system years back that utilized worm gears to transmit force between all of the wheels (the “unbackdriveable drive system”). Ended up ditching it it cause it was a dumb idea (totally impractical, inefficient, and expensive).
You don’t have to do the reduction through the shaft. You could simply have the shaft in a 1:1 ratio (or similar) with the wheels via bevel gears, the shaft effectively replacing the belt/chain. Reduction would still be accomplished in a gearbox.
Not saying it’s the best idea, but there are other options for power transmission out there.
Since finding the Dark Soul and similar tools for splicing (not just breaking) chain, I don’t ever want to go back to master links. No chain is stronger (or more reliable) than its weakest link. If you have a master link or a half link, one of them is almost certainly the least reliable.
We use belts for our drivetrain simply because if we order a kit bot we get those belts with it and after time you can use those belts for other purposes real simple, easy to put together
2017 - some type of black nanotube
2018 - AM14U3 it think
2019 - some kind of black nanotube
I personally am not a fan of the nanotube. The procedure our team went through to put (red stuff) on it was awful. We are getting the KoP. Not sure if we have the ability to go for chain, though, it sounds… reliable over belt? Any Einstein robots 2016- present?
Let’s say you push back against 1073’s monstrosity with a belt drive. Would your belts strip?
My understanding is the sprockets have more “grip” on the chain than a pulley does on a belt because of how much a belt tooth sinks into the pulley, VS chain and sprocket. Therefore, chains are more efficient, Correct?
Don’t combine these two, unless you have a tensioning system. The KoP spacings are all designed for 5mm pitch belts and same-size pulleys at both ends. #25 and #35 chain pitch is based on inches (1/4" and 3/8", and I don’t recommend odd multiples). Even before the chain is broken in (“stretches”), the chain won’t fit right.
Use the belts and pulleys that come with the drive base. The belts are 15mm wide and the pulleys are large 42T ones, they will serve you well. We used them on our 8" pneumatic tire 6WD last season and pushed plenty of teams around at IRI, no issues with anything.
Most importantly - get it together week 1!
We (3641) use belts inside of our drivetrain tubes. It’s great because they never need any servicing unless you let debris get into the belts. I think we have been doing it this way since 2013 (with the exception of 2016, when we still used belts but not in tubes) and have had only one belt failure during competition season since then.