My team is currently having a discussion about whether swerve is right for us this year. We are a 3rd-year team in a rural community and are finally getting our footing in FIRST. We have the funds to support swerve and many of my team members are convinced swerve is the only way to go. I feel that we will not have enough time to figure swerve out before our season. Even if we were to get swerve running I’m not sure if it would be a worthy investment.
So I guess my real question is, how difficult is swerve to set up? We have a skilled software engineer at our disposal but I don’t want to lean too heavily into our mentors. Is swerve worth it? or should we wait until we have more time to nail it down?
7 Likes
See the thread I made titled “Swerve is a distraction”. My main takeaway remains that “Successful teams are that way primarily because of their organization and structure. That defines their training, their leadership, their planning, their mentors, their resources.” If those pieces aren’t strong I’d focus on those first.
Regional matches are won on speed. A tank bot with a simple mechanism that is done early with lots of drive practice is far better than a swerve bot done late with a marginal mechanism.
This seems very late to start swerve. A strong software mentor helps a ton. Definitely taking a lot of risk to start now.
The strongest vote for swerve is the excitement it brings the students. I feel its a distraction for many mid-ranking teams but the boost in enthusiasm does have real value to it.
6 Likes
Just looking at one of your matches, it looks like that robot with swerve would not be much better and you’d be better served having autos, and making a simpler robot with tank drive instead of putting all of your effort into swerve.
See: 5586
2 Likes
I’m not going to tell you if you should or shouldn’t do it, but I will say my team got swerve at the beginning of the summer and we’re still figuring it out. If you do get it, your experience might be different though.
2 Likes
To contrast the rest of the thread, the first time we had a fully wired swerve chassis was 12 days into build season and we were able to use (most) of the features and advantages of swerve.
There are a lot of public and open-source tutorials for swerve and you can get a pretty decent swerve experience with little extra time. If you choose to go this way, you need a lot of programming time, and a big enough team that a person or two can spend a lot of time working on swerve. You also need a simple enough robot that you are done with extra time to spare to work on swerve. We felt like the ceiling of swerve is much higher than the ceiling of a tank drive so we wanted to go for it and constantly keep updating it rather than sticking with tank, especially because of the very open field and the need for precise horizontal accuracy.
TLDR: It can be done, but you need to invest in it with time and money. Look at the game to see if it’s even worth it.
5 Likes
Building swerve modules is the easy part and even then expect to screw it up at least once (most commonly glue the magnet!!). Programming is more difficult but I have been trying to lower that floor with my library YAGSL. If you are able to make your own subsystems that use DifferentialDrive you should be able to create a SwerveDrive object which acts similar to the former. There are ALOT of configurations you have to do for swerve and it is very important you understand all of them!
In short- yes you can still do swerve but don’t set your expectations too high unless you’re able to meet as if it were build season for the next month with a programmable bot built and wired correctly by next week. It may be more fun to do but you might just be better off with pathplanner + differential drive.
2 Likes
i would personally recommend getting into swerve because it’s a great experience and allows for students to learn a lot more for programming. if most of your team wants to do it and you think you have the capacity for it, i’d say do it. however, i would strongly discourage you doing it now, it’s great to do in the offseason but doing it right now before competition season will very likely not go well.
2 Likes
Swerve and building a robot on time/having good drivers are not mutually exclusive. Just because a team chooses to run swerve does not mean that some other part of their robot will suffer. And for that matter, swerve allows you to build simpler mechanisms due to to using swerve as alignment and caring less about passthrough vs sameside and is significantly easier to learn to drive than 6wd, meaning that a swerve driver with an equal or even less practice time than a 6wd bot will be a better driver on field.
3 Likes
It isn’t too hard but requires attention to details to build and a bit of skilled programing
Yes
Depends, when will you be able to get into your shop and do you have enough students to build the swerve while designing, building and programing scoring mechanism(s)?
2 Likes
Our experience was we purchased swerve modules in December of 2021. We built our first off-season swerve bot in April-May of 2022 and then built a swerve robot for competition for the 2023 season. This timeline worked really well for us as we spent months tweaking and learning. As others mentioned, what mechanisms you have on top of the robot are really important. Doesn’t matter if you can fly around the field but can’t manipulate game pieces effectively.
I personally don’t think that building a swerve would be the best for you to get better. before our team started swerve my mentor said that we must be able to “pick and place” game pieces. Pretty much where we needed to build an intake that could grab the game piece the moment it touches it, and during the offseason we did that, so in august we started on our own swerve drive. This ideology has helped our entire team understand the importance of moving game pieces rather than running around the field, pushing cubes around rather than picking it up.
Saying that i think your teams best action moving forward is to focus a lot more on intakes and intakes designs. having a really good intake would make you a better team, better than having swerve and an okay/not great intake.
5 Likes
We built our first custom swerve drive in 2016 as a learning experiment. Our first FRCFIRST Robotics Competition-style full robot with swerve (Thrifty Swerve modules, in fact) was built in the winter of 2020, during Covid, and was used for training and practice. We chose not to use swerve in 2022 and did pretty well with just a standard WCD tank setup. We purchased several sets of modules in the weeks after the 2022 game ended, and spent our summer and fall building and programming robots that used the SDS MK4i modules. We used this exact setup on our 2023 competition robot and did well again. So there’s my personal journey with swerve, lasting at this point nearly a decade before we used it in competition. Hopefully this helps in understanding my answers to your questions.
Mechanically it’s as easy as most tank drive designs these days, especially if you use the REV, SDS or WCPWest Coast Products modules. All three are easy to assemble and only require four box tubes and a pan to be a drive base. That doesn’t mean it won’t take work or that you won’t make mistakes.
I admire this. I think my coding mentors and students are quite skilled this year and we have a slew of them, but even with our prior code they still run into difficulties on our off-season robots that they’ve built this fall. I wouldn’t want the pinch point of your team’s ability to play to be one coding mentor.
Maybe. I don’t know more about your situation than this post and the videos of your 2023 matches. My opinion and recommendation is that you do not take the leap this year, and instead focus on mechanisms and especially driver experience and practice. Build the KOPKit of Parts and drive it for a month before you compete, try to give your drivers realistic challenges with the joystick, and you will be more likely to be picked than you would be if you have swerve modules and no driver experience.
I would say yes, but whatever you choose I wish you the best of luck.
Mr. N
5 Likes
If it was August/September I’d recommend you’d invest in swerve and test it at an offseason, given it is almost December I’d recommend you roll the dice that next year’s game isn’t as biased toward swerve being the optimal drive type as this one was (the performance discrepancy normally isn’t as big as we saw this year). You can save a few $ and spend that debugging time that will inevitably partially happen during the season instead making your mechanisms really good.
2 Likes
Ignoring all the stuff about how “swerve doesn’t make you better,” other people have already given you much better advice on that than I can.
However, considering it is the end of November, and you have your doubts about finishing in time, I would wager that it’s better to wait until next off season.
Only a month to ship, assemble, code, test, and troubleshoot swerve? Sure, you have the season, but other priorities with the mechanisms take over, and the last thing you want is to be worrying about your drivetrain not being ready midseason.
3 Likes
As a fairly established team, we decided to take the swerve plunge 2 years ago. We did it pre-season and worked on it for several months prior to the season. We thought we had a working swerve going into that first season, but we learned so much during the season that if you compare our swerve season 1 to our swerve season 2, there’s just no comparison in how they performed.
A lot of that has changed. If you go with CTRE controls, they will quite literally write the drive software for you in Java. If you go non-CTRE, there are a number of libraries that are pre-written for you to draw from.
Keep in mind that just being able to write a drivetrain for swerve is one thing. Your drivers will need practice, and you’ll also need to develop a working Auto system.
If you haven’t started all these things already, it’s unlikely you’ll be successful doing it between now and your first competition unless you have experienced outside help.
1 Like
This.
OP, you are doing a good job looking at this objectively and evaluating resources.
Punt to next season, use this season to talk to teams that seem to have there stuff together, let them share their pitfalls, advice, etc. Take your own notes, what you do like for the 2024 style of game, what you don’t, how team xxxx tackled such and such swerve issues (packaging, obstacles on the field, etc)
Use this upcoming season to Learn and document, save a few bucks and have time to work on stuff, and approach next season’s fall off-season events with a swerve bot under you (i.e. lots of teams did small cube bots this year)
5 Likes
If you haven’t started yet, at the very least have a fall back. If you do want to pursue swerve for this year, here’s my suggestion:
Once you figure out your dimensions, build a KoP chassis to those dimensions, and build your robot on top of that, reserving ports on your PDP/PDH for the steering motors. Then, you can begin drive practice quickly, and improve your manipulators. Then, if swerve gets better than skid steer more than about a week before competition, decide whether to move it over. (Or build two robots like many teams did before bag and tag went away.)
2 Likes
After using Swerve for a couple of years now, I would STRONGLY recommend waiting until next season, in my personal opinion.
We built our first Swerve drive in late 2020/early 2021 and spent basically an entire year, and hundreds of hours fine tuning it. In spite of that, when we built our 2022 robot as a swerve robot and took it to a week 0 competition to try it out, it did not drive, at all, that entire day.
I would consider our team to be (both when we started swerve and today) fairly competent in both mechanical/electrical and programming areas, and we struggled to get it working to an acceptable level even with a whole year of practice, and still strive to improve it to this day, two seasons and three robots later. I’ve heard almost identical feedback from virtually every experienced team we spoke to about switching over to swerve as well.
I will grant you, swerve has become significantly easier to get into in the past few years than it was when we started (both with lots of COTS module options and programming resources), but it’s still not something for the faint of heart to jump into without preparation.
That all being said, assuming FIRST doesn’t make any significant changes to how games are designed/played, Swerve is pretty clearly the future of FRC and should be pursued for teams with the resources for it, as a long-term project if nothing else. If you insist on building one this year, GeeTwo’s comment above is probably the best option. Build a simple tank chassis and design the rest of your robot on top of it, and then, if you have the resources, build an interchangeable swerve chassis along side it. If you can get it working, swap the chassis over and you’re good to go, if not, you’ve got a decent working tank bot to compete with and a practice chassis for the off-season.
3 Likes
If this is the chosen path a team takes remember swerve acceleration is 2D, so structure has to be able to take forces from strafing (and collision while strafing) .
Wire lengths and electronics layout need also be a high level consideration, you could put some of the electrical and control system on the lower part of the super structure to make this a non issue.
Best answer is it depends on the game.
Having said that it also depends on your goals winning or learning.
Doing swerve is a great learning experience, even if you don’t get it working well it will be a great foundation for next year.
We did out first swerve during the covid season and never really got it working well but that laid the foundation for a district championship the next season.
The good thing about swerve is with todays available solutions you can get a swerve chassis up and running in a week giving your programmers the rest of the season to make it work while you continue to work on your scoring mechanisms.
2 Likes