Looking for some advice. Our team has yet to incorporate swerve drive into our designs. For reference, we’ve been doing a basic tank drive with a KOP chassis and 4 CIMs for several years.
We’re finally talking about making the advancement to swerve drive, but find ourselves asking “is this a bad time to get into swerve drive?”
We know swerve is worth it, there are just a lot of obstacles in the way.
Ideally our first choice would be an SDS MK4I set up.
We’ve never used falcons, and the few seasons we used Neos, we managed to have a lot of burn out issues.
However, the ongoing supply chain issues obviously are limiting us with product availability for either.
We started looking into some alternative solutions, like DIY swerve. Team 5006 has a pretty good “intro to swerve” project, but it’s from 2017, and some of the parts are discontinued. But using that as a starting point, I’m not sure our team would compete with it. I think it would be a fantastic side project for our students to learn more about swerve, but I’m worried about reliability for a competition robot.
Going back to the first choice, even if the motors that are out of stock became available, I’m worried about ship times being 4-8 weeks, and not having enough time to test a system in the off season and know we are capable of getting it working before build season.
As an alternative to swerve, we’ve thought about just doing a tank drive with a custom chassis, so at least it’s some advancement from the KOP chassis, and something new for our students to learn.
What would you suggest?
If the mk4i (or mk4) models aren’t available, is it worth it to attempt to compete with a DIY swerve?
if the MK4i or mk4 models are available, would you recommend Neos or Falcons
If we hold off on swerve, what would you recommend as something better than the typical kop drive train?
Swerve drive is complicated, luckily there are projects that alleviate sone of the issues (like YAGSL ) which help with programming. Definitely build one or 2 in the offseason to get used and figure out “what not to do” and learn! That’s really all you can ask of your team. Enjoy it and have fun!
Now is a great time to get into swerve for a few reasons:
School has barely started; you have time to learn how it works before the season starts
Swerve is more affordable than ever, in that COTS is far cheaper than anything you’ll develop in house
Swerve is more accessible than ever, in that there are dozens of fantastic resources for building and programming swerve robots
It’s absolutely not worth going DIY if you’re asking the question. If you have a few experienced mechanical engineering mentors, a well kitted shop, and two years for R&D before you implement it into your robot, you’ll be able to match COTS pods.
We use all NEOs, all the time. Drive, intake, arm actuation, etc, are all NEOs on our robot, and we rarely have issues. I think we burned 1-2 out in the last two years, but they were connected to other failures.
Go with MK4is. Order them now. Make sure you buy at least 5, because not having a hot spare means that as soon as a pod breaks, you’re immobile. Right now there’s “only” a 4 week lead time on new SDS pods, but that was 10-12 a few months ago, iirc. I don’t know what the typical led time on other options are, but stock is wonky across the board right now.
I’ll put in a plug for two alternatives to SDS modules.
REV swerve modules have some advantages in that they are less expensive by a few hundred dollars for a set of five, are dead simple to assemble, and are made to work with Neos since those are also a REV product. They have a tiny footprint and good support from the company. Disadvantages include smaller 3" wheels which don’t go over bumps as well as 4", and their wheels were manufactured poorly for the first season. It seems like that issue has been addressed though.
WCP swerve-X modules are similar to SDS and those who used them love them. They’re only sold by WCP and are not as popular so it’s unlikely you’d find spare parts at competition.
Edit: we’re very happy with our SDS modules, which we ran in the 2023 competition without issue, and they would also be a fine choice. Just want to make sure you’re aware of your options, OP.
3620 had several successful pair-ups with Stryke Force at district level events early on, which inspired us to consider swerve. We spent two years looking at DIY alternatives, when in hindsight we might have made the jump to swerve sooner by choosing one of the early COTS platforms. Today I would not advise any team to do what we did – just pick the best COTS option for your team and SEND IT! Don’t delay, start today. The whole team needs to be on board, from strategy to CAD to fabrication to electrical to coders to drive team. And the more overlapping roles your team has for individuals who do more than one of the above, the better.
Don’t go DIY. Either order SDS and prep code and chassis while you wait, borrow a chassis from a local team, or get a set of REV or WCP swerve modules (which are in stock right now). They are all easy to use. You will want a minimum of 5 modules, although I recommend 10 so that you can have a practice chassis.
Falcons are mega OOS so your only choice is to hope someone releases a new motor or go with NEOs. Many teams have used NEOs successfully. I hate the tiny encoder cable, but it’s not as bad on the full size NEOs as it is on the 550. If you’re burning out motors often, make sure your current limits are set up properly and potentially read the temperature back from the motor as a safety. NEOs are absolutely viable as a drivetrain motor.
Right now I would stick with KoP or go to swerve. Any money you spend on an intermediary like a classic WCD is basically wasted, because you’ll probably want to go to swerve next year anyway. Custom chassis take a while to machine, potentially longer than a swerve drive, and provide little competitive benefit relative to other things you could be doing.
It’s a bad time to get into custom tank.
It’s a fine time to get into COTS swerve.
Keep the kit chassis as an option on the table going into the season… it can take two or three years to make the jump.
Yup, thanks Patrick. And to be clear, SDS MK4i and MK4 are the units my team has decided to invest in. We currently have two robots running MK4i, one on falcons and one running Neos. Fantastic product! Just want to make sure that OP is aware of all their options.
Mentor for local team 177 here. If your mentors are interested in assistance in design of an SDS based swerve chassis, or even software for another swerve vendor, feel free to have them reach out to me via Chief Delphi
I’d agree with above suggestions to try Neos again as brushless motors will gain you the ability to use many newer control features and built in encoders that are more robust than adding encoders to the KOP chassis gearboxes. Otherwise, trying swerve now is definitely worthwhile. Ultimately that decision falls on your own team and your evaluation of your time, money, and skillsets.
Edited for a typo, originally accidentally said “aren’t as robust as external encoders in the KOP”
If you weren’t aware, our 2023 code is public for our falcon based sds mk4i setup. This isn’t super cleaned up as it’s been public all season and we haven’t had a chance to clean things up in the summer.