Hello! I’m wondering how teams are justifying using swerve during the upcoming “real” season. We saw heaps of teams trying swerve drive over the offseason competitions held this year but I’m wondering the practicality of that when it comes to budget constraints. For a team with lower cad/design resources there is no way that we could feasibly create a swerve drive competitive to that of swerve drive specialties. But if my math isn’t wrong the package for swerve drive specialties mk4 modules is $310 for the modules itself + 400 for the 2 falcons + 50 for the can coder totaling 760 dollars total cost per module. That’s a $3000+ drivetrain. Is the performance you get from swerve justified from a budget perspective?
If budget is a concern, then I would not attempt swerve. Really, I would strongly advise you don’t attempt to run it inseason at all until you have had the chance to try it in offseason.
With that said, some teams can justify spending that amount of money on their drivetrain. Swerve does have it’s advantages, and there are several threads discussing this already. Basically, it comes down to trying to squeeze any amount of performance out, regardless of cost. In my opinion, for 98% (literally) of teams out there, they would be better served putting their resources into other mechanisms and doing the Kit of Parts chassis, perhaps stepping up to a custom tank chassis. I would only consider running swerve when you can honestly asses your robot performance and conclude that your drivetrain is holding you back. If it isn’t, then focusing on your drive train over other mechanisms is getting some serious diminishing returns. Also, don’t forget that your drivetrain is the most important part of your robot - a failure there, and you’re left with a very expensive and heartbreaking paperweight.
IMO there is not a “simple answer” to this. For some the answer may be yes, but for others that is not the case.
At least speaking for my team, the year we began to develop our own swerve module we did have a begun spike in our performance. However, in line with this was the massive amount of time that some core students spent to develop the technical skills needed to do just that which translated into better designed robots.
Don’t get me wrong, to somewhat quote 2767, “why wouldn’t you want to go sideways”. The real benefit is what students and mentors can learn from the swerve experience.
In the end, I would say any quality robot does not come from a cool drive base, but of a complete robot that can be controlled well by the operators for the game.
Do you see the shift for top level teams(like 1678, 2910, 1690) to swerve as a tell tale sign that the rest of frc is moving in that direction? Do you have an opinion on if the game will remain “swerve-friendly” from now on?
I think I see the use of swerve by high profile teams as a fair notion for other teams to make these questions. Does that mean I believe the way to succeed is by having swerve? Not at all.
I think teams will need to look at their own circumstances, but there will never be a point in which you must have swerve to succeed in FRC. Only a desire to succeed.
If you look through last games. There are times when swerve teams didn’t use swerve (2767 in 2016 for instance) or swerve teams that adapted to make swerve work (16 in 2016).
I would not count on FIRST consciously making games swerve-friendly, although I do think swerve is well suited for most games. With that said, don’t exclusively base what your team does off what top-tier teams are doing. You won’t be disadvantaged by not doing swerve, however you will by having an ineffective or unreliable drivetrain. Swerve is growing in popularity, but even now, for every “top team” doing swerve, there are just as many that are sticking with good ol’ tank drive. I would be very surprised to see swerve grow to be more than 10% or so of teams globally
^+1 to this
We’re going to see a whole lot of barely-functional swerves this year.
Not only does it not make budgetary sense for most teams, but it’s also not trivial to get one driving well enough that it’s a benefit in the first place. We ship swerve code with WPILib, but it’s not exactly usable without a fair amount of controls knowledge, and there’s no way to change that.
That’s what I’m worried about!!
Then run the AM14UN and worry less.
You know who 1690 picked first at Chezy Champs? Not a fancy-pants swerve, but 973’s well-driven skid-steer setup that just drove angry, shot hard, and complemented their strategy well.
I’m sure the day is inevitable that swerve will be necessary to keep pace at the highest levels, but I do not believe that day comes in this half of the decade at the minimum. Until you’ve tried it in the off-season and you feel ready to make the jump, don’t.
*There are a few teams who have the mentor and financial resources to attempt swerve for the first time in-season if they decide it’s the way to play the game. If you’re asking these kinds of questions on CD, you are almost certainly not this team. And that’s completely okay.
I’m not sure who you’re left with, if you want to find (teams who’ve chosen swerve for competitive performance reasons) ⋂ (teams who are willing to lay out their decisionmaking for public comment).
I want to say I agree with everyone above when saying to make sure you are ready for swerve. It’s expensive and the programming is intense.
That being said…if I could be a devil on your shoulder…You keep seeing all these swerve robots and it’s look really fun. Well I can tell you, in fact, it really is fun. We’ve also noticed that a well programmed swerve is easier to drive than a WCD initially (this has several caveats). I highly suggest picking up swerve over the summer and testing it for comp (if it’s with in budget). When talking about high end teams and the functionality of their drive train I feel like there’s a disservice being done. When it comes to auto and pathfollowing albeit WCD or swerve, the programming is very complex regardless of the option you choose. On the mechanical side, COTS swerve modules have made it so that putting a frame together is about the same as a WCD.
I think we are in that intersection.
Reasons why we are did swerve in 2021 and are planning to do it in 2022 (unless the game encourages more ground clearance/wheels like 2016)
- Swerve has always been appealing but the drawbacks always out weighed the benefits for us until recently. I laid out all the changes to the rules and FRC ecosystem that have removed some of the problems with swerve in this post - Swerve implementation for teams doing swerve for the first time - #12 by AllenGregoryIV
- COTS Swerve will save us machining and assembly time. (This can also be accomplished by using the kit chassis).
- COTS swerve removes design decisions around the drive train. (This can also be accomplished with the kit chassis or if your team chooses a standard WCD or other drive system)
- Swerve allows our driver to recover from errors faster. A well driven tank drive will be close to as efficient as a swerve drive at aligning to targets or driving to a spot, but when a mistake does happen a swerve drive can correct the error much faster since they can slide sideways or rotate while translating.
- Keeping your drive train only in the corners of your robot frees up a lot of volume, allowing the robot’s foot print to get smaller or the space to be utilized for other things.
- I believe swerve will allow our autonomous performance to get better, we haven’t been able to do this one on the field yet so that is really just me hoping at this point.
All that being said if the cost of swerve is preventing your team from doing anything else in game, such as deciding between Swerve and doing the end game, or between Swerve and an intake, or any other such decision you shouldn’t choose swerve.
I have a swerve cost comparison spreadsheet here: Swerve Comparison - Google Sheets
The only swerves on there I would consider using are the new WCP X and the SDS MK4. So you’re looking at $2400 to $2520. However, note that modules are reusable, especially with steel gears and brushless motors. This makes the cost easier to stomach.
Most teams will not benefit from a swerve drive. The money could go towards retooling, organization, or Greyt mechanisms (which will do more for a team’s performance than any drivetrain). However, if you have new drivers and strong programmers, plus time in the offseason, a swerve drive can realize benefits for upper-middle tier teams that are struggling with building a competitive robot. You should really only use a swerve once most other aspects of your build season are figured out, including engineering expertise, finishing on time, and limiting scope.
We’ve not done swerve in a long time, but this is correct as I see it, the advantage to a COTS swerve is time & energy saved when time & energy are at their most premium. I suppose we might also eventually recoup the cost by not machining new gearboxes every year. However, we feel the learning our kids get by designing and building a (WCD) drivetrain from scratch each year is valuable.
I think you should ask yourself: what is your role on the field? When analyzing the game at hand, you should look at the typical cycle that you’ll be running on the field and optimize for that. We saw at the start that we want to play the short cycles intaking from the field, and designed our whole robot to make it best in that(swerve, wide intake and shooter optimized for target zone shooting). Swerve is only a part of that equation. From the start we also knew that if we wanted to run only loading station cycles then swerve is not going to give us an advantage - a well driven tank drive with good subsystems is just as good if not better for that goal (if taking into account shifting and easier crossing of the rendezvous). We chose 973 in CC because they were simply the fastest shooter-regardless of having swerve or not (plus the synergy we had with them was insane) . They built a simple, well designed tank drive robot and drove the hell out of it, and I would take that over a fancy swerve drive every day.
That being said, having already had a very long development time on our swerve drive and forming a good eco system around swerve for the past two years, we’ll probably stick to it if we can (ie the game doesn’t have 2016 style bumps). We’ve now reached the point(as a team) that building and programming a swerve drive is easier than tank for us, but I would never go for it in a season without significant off season experience.
As to justifying the cost, a year with no competition actually gave us some financial breathing room. Ah, the wonders of use it or lose it funding…
We did invest is some COTS swerve modules and have been tinkering with them. I doubt we’ll get to “bet the season on it” levels of comfort in time, so I suspect we will end up going more basic albeit with brushless motors for the first time.
So for us swerve capability might be another year off, and our work to date on it going towards a very clever demo bot.
We design and manufacture our swerve modules. Our cost are much less than a purchased module. However, for a first time team to do this for a first year is a major project with a ton of land mines to step on. Cots modules are allowing teams to get into swerve, but the cost of cots is substantial. Pay or sweat.
For how long? I mean this is the elephant in the room right? You can buy COTS swerve and use them forever, but if you build it, you have to “throw it away” every year and build a new one – even if all this entails “skill-building wise” is buying new stock and loading it into your CNC, loading your CAM, letting it go, and then assembling.
I doubt this, really. I’m sure there are examples here and there, but generally the teams that have access to the systems engineering knowledge required to run a swerve effectively are also the teams that could already afford to build a swerve.
The teams most likely to be convinced to try swerve by these products are not teams that will likely be able to use them successfully. I also don’t think they’re really the intended market, for reasons other people have covered above.