Concerning Proliferation of COTS Swerve

It’s something that I forgot to touch on in the original post, COTS gear boxes don’t fundamentally solve a game challenge. Yes there is some lose learning in buying a solution but the rest of the mechanism can make up for that. A intake spun by a VP gearbox can still teach the value of compliancy, belt drive, differing wheel types and sizes, linkage geometries etc. A climber in a box, COTS swerve or similar complete mechanism leaves alot less room to gain experience in implementation over a simple gearbox that would have to be used in tandem with many other elements COTS and custom alike. It also stands to reason why teams don’t cut their own gears, or cast their own belts. Theres a gradient and I’m not sure how it fits exactly, but COTS swerve is so far on one end I don’t think the impact is apparent.

Our 2019 robot was very nearly a swerve, and would have no doubt been better (and waaaay less complex) had it been.

30 Likes

This is where my teams at right now. We’re seeing a lot of teams of similar skill to us making the switch to swerve, but we don’t really think it makes sense for us, and in some cases for them either. We just won our first event this weekend against several swerve robots using our normal WCD, and honestly I think we were better for it. Don’t get me wrong some teams are doing fantastic things with swerve, but it’s not some sort of automatic improvement button and in some cases I think teams would’ve been better off diverting resources elsewhere. I hope teams are least decide to do it in the off-season first. Because doing swerve in season for the first time seems like a big risk, cots or no cots

4 Likes

That’s your prospective. To me, a Swerve so no different than any other driving solution. This year, I have seen COTS swerves do well and others that act like lumps on the field. Are you saying that swerve is OP and should be banned unless a team makes it themselves and THEN it’s acceptable? if so, that seems…odd.

Here is the beauty of FIRST: each team can do what they want. If you want to buy a swerve, go right ahead! If you don’t, that just as ok. This is the same argument as “mentor built” robots vs. “white gloves” or the student coach vs adult coach. Each is equal in the eyes of FIRST. It only becomes a problem when folks start to impose their own values on other folk’s choices.

my suggestion is to stop worrying about what other teams do, and focus on creating the best season you can for your team. I know that’s what I’m going to do.

Best of luck to ya.

2 Likes

Teams and mentors that are on chief delphi are largely of the more experienced and more well resourced type. So I’m not surprised to see that there is a bit of well reasoned pushback against this post.

That being said, I agree with this post completely. The educational value of a fully featured swerve drive arriving on your doorstep for the right price is negligible, and comparing these plug and play upgrade modules to COTS gearboxes and kitbots rings hollow to me due to a variety of factors.

COTS products have almost exclusively raised the floor of capabilities for less resourced teams to compete with consistent and reliable robots, whereas the new swerve modules are a way for wealthier teams to increase the gap between themselves and others, without having to use the normal development time and effort associated with swerve.

A single fully kitted out set of swerve modules can cost more than an entire competition ready robot without swerve. Factoring in replacement parts and a practice robot the extra costs over a tank drive base really pile up. Some teams have this money to burn and they are rewarded with a nearly instant bump in competitiveness and capability.

It leaves a bad taste in my mouth, not because the swerve drives are good, but because they are so obviously leaving less well resourced teams behind. Sure you could argue that this has been happening for a long time already, but this year specifically I have felt like it’s more obvious than ever.

4 Likes

Linked below is the same sentiment from 2004. I would encourage everyone to read this ancient thread from when AndyMark.biz was a fledgling start up. This has all come to pass (well, with EveryBot and Ri3D the game advice is free!) , and we’re still here inspiring students with robots.

EDIT: Well, I guess the ship dates are still not guaranteed too. :slight_smile:

I admire the entrepreneurial spirit, and I’m tempted to use the products myself, but I’m afraid that this could eventually lead to a situation that would compromise FIRST’s objectives.

Now I can buy a transmission component, or an entire transmission. Why not a chassis, or a chassis with transmissions, wheels, axles, chains, sprockets? It might come in four wheel, six wheel, or omni wheel versions. All I would have to do is drop in the kit parts and I’d have a box that moves. Then I could focus on designing and building the appendages. Or maybe I could buy those too, with a guarantee that they’ll get to me before the ship date. The deluxe package would include two sets of everything, one to ship and one to practice with. If I were on a tighter budget, at least I could buy a “Game Analysis Report” a week after kickoff, to save me the trouble of coming up with my own strategies.

Maybe I’m stretching the possibilities too far, but maybe not. Might the focus of FIRST move away from engineering? Is it possible that teams would spend all their time raising money to buy the best “standard” components, at the expense of the design experience?

7 Likes

This does not address the core concern I brought up though, which is that previous components have lifted the floor for teams with less capabilities and these swerve modules have instead served as a way with teams that already have lots of resources to tear away from the rest of the competition.

I’m not making a slippery slope argument here like the post in 2004, although I will point out that they have proved correct.

I wouldn’t have an issue if these modules were cheap(er) and plentiful, in fact I look forward to swerve modules becoming cheaper and more common for teams to use. That’s not the current situation though, instead the COTS options are prohibitively expensive and give a significant advantage in competition over other cheaper drive systems.

If the COTS option didn’t exist, the only option would be to manufacture them yourselves. The machining capability and manufacture time would be out of reach for all but the top couple of teams in the world. Big name teams with massive budgets would shrug off the tens of thousands of dollars of fab capability in the name of competitive advantage and everyone else would be left in the dust.

14 Likes

This is literally what people said in 2004. The only difference is now the $360 component is a swerve module instead of a shifting 2 speed transmission. Look at what the COTS vendors have done in the interim to improve robot goodness! It is truly awe inspiring.

9 Likes

I agree with the first half of this statement, but not the second half.

The spiraling cost of participating in FRC is a huge problem and I hope FIRST takes steps to address it. But this is a much broader problem than swerve drives, and is largely orthogonal to competitive balance. From a purely competitive standpoint, most of the teams trying out swerve drives for the first time this year are not seeing a “significant advantage.”

9 Likes

Inspiration comes in many forms. This makes most sides of the common discussions that come up here valid based on the view point of the users. Some view inspiration internally (how do I feel about what I’ve done) and other view it externally (how do I feel about what someone else has done). I hope some people can do both. I’m personally very proud of what our team has managed to accomplish this season with a mostly machined in house robot. I also adore one one of our local teams who is using a COTS swerve drive (their implementation of everything above the drive is also just gorgeous). While I’m inspired and I think they’re inspired, I can see how inspiration can be thought of as zero-sum based on your view.

It seems like part of the thought isn’t that the technology exists or that’s purchasable but that it’s expensive. I totally agree it’s expensive. Thankfully the cost is going down from what it was (a revolution swerve module with no drive/steering system was $330 each back when I was in high school) vs some modules now costing about that much for a fully working unit. If demand keeps up, I’m sure with future innovation costs can become more affordable.

5 Likes

I disagree with this. Using COTS Swerve is something you have to learn a lot about in order to use it well. Just because WPILib includes code to help control your swerve modules doesn’t mean setting it up is easy. You are going to learn a lot about encoders and PID in the process. Unless you copy another team’s code line for line and use their exact same motor controllers and encoders, you are gonna learn something about the swerve drive you install on your robot.

From 2018-2020 our team used the AndyMark swerve modules. They are not plug and play! No COTS swerve modules is. Getting the swerve modules to work my first year on the team was an awesome experience. Sure, no one created the design for these swerve modules we were using in CAD, but there were plenty of other things for our team to CAD. We may eventually decide that we want to create our own modules at some point, but I think COTS swerve modules are the way to go for any team wanting to start out with swerve.

6 Likes

I do not believe COTS swerve should be outlawed, as there are very justifyable reasons for selling and purchasing it. What I propose is that the large scale purchasing of COTS swerve is on average to the disadvantage to how I percieve the spirit of FIRST competition. There’s lots of good insight and differing perspectives from this thread that gives me plenty to consider, and I don’t want to put down teams that use these solutions. Out of my team I’d say I’m the most proud of what we’ve achieved, despite all i’ve stated.

Best of luck to you as well

Giving our students experiences like this is why we have deliberately made the choice not to purchase a COTS Swerve even though it puts us at a competitive disadvantage. Perhaps someday that disadvantage will grow so great that we must reconsider, I dunno.

1 Like

I agree that to properly utilise COTS swerve a team must put in the effort to properly understand it, a good thing as this is a great learning process. My concern was the commonality of the complete modules that for example came with Falcon type motors and a specific encoder, that the vendor produced code can almost run out of box. Teams may learn from implementing COTS swerve, but nowhere near as much as a full development process which in my opinion is a significant loss for the significant cost, the gain being a more competative machine.

Interesting to hear that you recommend a team start with COTS swerve based on your experience with the Andymark swerve. This isn’t something that i’ve done, so what about the process would make it worthwhile for a potentially new team? When my team started designing our swerve we 3D an entire module from a open source cad on CD, and while playing with and slowly improving it we gained the insight into it’s fundamental mechanical operation. The code was done from scratch by breaking everything into basic vectors and working up. More specifically, compared to this approach of learning swerve what is the advantage to purchasing a COTS one and working from there given the significant cost increase? Aside from having a working swerve for the next comp.

1 Like

Our swerve was manufactured with basic 3D printer, borrowed lathe time and a cheap chinese 6040 CNC router. While this is probably outside of alot of teams but I would not consider it ‘top couple of teams’ as these resources already existed to make our machines.

5 Likes

I know you said you don’t hold it against teams in the slightest for going the cots route, but this thread is giving me serious flashbacks to the KOP drivetrain arguments well over a decade ago. IMO there is this type of energy in this thread: “If you don’t do it the hard way how are you gonna appreciate what you have?”

Frankly I detest this subtle gatekeeping, intentional or not. call me a suborn old alum. whatever. /fin

14 Likes

I think something else to consider is the different strengths of each team. For example, some teams might not have the strongest mechanical team, but have a really strong programming team. A COTS swerve module would allow the programming team to experience programing a more advanced drive train and do more advanced autos that otherwise their mechanical team would’ve struggled to make.

19 Likes

The pay to play aspect of COTS swerve is a concern, but if you don’t think your custom designed and machined swerve modules are worth at least $500 each, you are vastly discounting your labour and opportunity costs. Just being able to build your own is a mark of great privilege.

25 Likes

Yep, I agree that this is a very good reason to obtain COTS swerve. It is then a tool to gain experience and skills otherwise unattainable.

1 Like