pic: Team 221 LLC. - Wild Swerve Module



A pre-production Wild Swerve steerable transmission under development by Team 221 LLC.

These units are based on designs used by the legendary frc111, Wildstang.

You can see them in person at IRI in July! Production units should be available for sale before the fall.

What a day it would be when every FIRST team would have access to a reliable, dependable swerve drive system, especially one similar to Wildstang’s. Cool product!

I can see this design being a problem with any surface that’s not perfectly flat though.

(in before is this legal)

I’m going to catch flak for this. I don’t care.

What a day it will be when instead of engineering and designing our robots, we’ll all go out and buy our parts from our favorite manufacturer, and bolt the pieces together. If we wanted an erector set competition with everything simply to build, we’d be doing Vex. Remind me how buying a pre-designed machine will teach my students how to CAD, design around many variables, and fabricate/test a system? Oh wait, it won’t.

As an engineer if you are assigned to design X product, which requires components A, B, C, D etc… and you choose to design, prototype, test, pay for manufacture, etc… of all the components when suitable COTS items are available, you will be fired.

The cost of buying component B could literally be 1/10th to 1/100th the cost of designing and making it (or outsourcing production) in house.

It’s the system that matters, not the individual components.

1 Like

I actually was thinking this was the main reason there were no pre-built crab drives. And I, to a large extent, feel the same way (despite what my post above might imply, that was more of a “realization that FIRST has changed” rather than an opinion). “Crab drive” has always seemed to be one of those things every team approaches after awhile, then they make and master their own style and build of it.

Then again, we’ve had a similar debate before when the AM shifter came out, if my knowledge of the spotlit posts recall correctly.

This raises a question of where the line is for FRC in terms of part standardization. Surely no one wants to be able to pick a full, working, manipulator off of a shelf, plop it on their kitbot, and win Einstein. At the same time, I doubt anyone wants to get to a point where teams have to manufacture their own gears, sprockets, and wheels (talk about a high-resource advantage there, you’d need C&C or laser cutting just to get out the door).

Is a crab drive for any team too far? A snippet of old rules (though standard disclaimer that rules are always subject to change and that in 2010 we could all be allowed only LEGO):

However, COTS items that have been specifically designed as a solution to portion of the FIRST Robotics Competition challenge may or may not fit within the FRC intent, and must be carefully considered. If the item provides general functionality that can be utilized in any of several possible configurations or applications, then it is acceptable (as the teams will still have to design their particular application of the item). However, COTS items that provide a complete solution for a major ROBOT function (e.g. a complete manipulator assembly, pre-built pneumatics circuit, or full mobility system) that require no effort other than just bolting it on to the ROBOT are against the intent of the competition, and will not be permitted.

Full mobility system is the obvious place where “premade crab module” would hold, really. So does FIRST think this is okay? (Don’t answer this)

Relevant threads:

Hardware alone does not a robot make.

As Yoda once said, “Control, control, you must learn control.”

I am a hardware person, but I couln’t agree more.

Chris, I’m going to have to go with Craig on this one. It’s not really about making compettions fair, it’s not about equal opportunity or teams doing similar designs. It’s about learning and teaching…so to the point…is swerve cool? I guess so. But let me warn teams out there, swerve isn’t always as big an advantage as it seems to be. There’s a reason only 3 different teams have on nationals with a swerve drive. Swerve is a weight commitment that can severely take away from other parts of the robot if not done properly. I’m not saying don’t do swerve, I’m just saying it isn’t as big an advantage as people think.(heck my teams done swerve the past two years). Skid steer teams are at pretty much no diadvantage to swerve teams. It’s hard to tell on regolith, but in 2008 the two most manuverable teams were 1114 and 968, both skid steer teams.

Swerve is cool to do and it’s a good learning experience, but don’t think you’re at any sort of dissadvantage if you can’t do it.

That being said, that swerve module looks really nice and it’s great to see teams experimenting with things like this, good work.

Thanks for all of the comments so far, the debate over COTS items is an old one and will probably continue for a long time.

That said, the pictured module is an assembly of many components and does not fully represent how the final kit style product will be sold.

Some cool stuff going on with the current design…

-it uses the KIT transmission components, including gears and shafts
-it can use the KIT wheel as pictured
-it is constructed of over 50% currently available COTS components, including many Andymark items, like hubs, sprockets and spacers
-many gear ratios are achievable using Toughbox change gears from Andymark
-has feedback points for pivot rotation and wheel speed/location
-can utilize the US Digital Encoder utilized on KIT transmission
-has been adapted from highly successful designs used by Wildstang

This last point is my favorite. Just like Universal Chassis, which was developed by students and mentors from frc27, Team RUSH…over ten years of work by students and mentors has gone into the Wildstang swerve modules and many of those features have been directly included in this product.

If this module is available at a reasonable price, I could certainly see our team purchasing a set for offseason testing, research, and development use. The students have fallen in love with the idea of swerve drive, but as a team we’re struggling with the concept. This could serve as a training tool as one way to make a swerve. I’d be willing to put down money that says if we do purchase a 221 set, it will not go on our 2010 FRC robot as is - the kids would want to modify/improve it to suit our needs. Furthermore, it would probably take us a couple years to get our heads wrapped around the system that WildStang has been refining since 1868.

Remember the immortal words of Mr. Beatty: The three most important parts of any robot are drive system, drive system, and drive system.

I, like many of you I suspect, felt uncomfortable when I saw this for some reason. I knew it was just a matter of time until somebody offered a swerve kit. But that still didn’t prepare me for actually seeing it.

My first reaction was, this has gone too far. Traditionally in FRC, mastering the swerve was a result of years of iteration, piles of aluminum shavings, and many sleepless nights for programmers. Effective swerve teams were like an exclusive club. Now anybody who can afford it can just order a working swerve system* from a catalog!

  • it was at this point in my train of thought that I realized the flaw in my thinking.

This is not a working swerve system that is for sale. It is a single wheel pod. I know from experience and observation that this is just the tip of the iceberg when it comes to making a 3DOF drive train. It is no more a “full mobility system” than an AndyMark shifter:

  1. Effectively mounting the pods, steering motors, and steering sensors is a nontrivial exercise.
  2. Writing effective control software is a very challenging task, especially with non-coaxial systems that need to limit pod rotations.
  3. This is by no means the only way to do swerve.
  4. Many of the teams who successfully utilize swerve year after year (71, 111, 118, 1625 come to mind) have another hidden talent: they are masters of weight reduction. Swerve drives represent a huge weight (not to mention motor) investment, and require special attention to drive, chassis, and scoring mechanism weight in order to be good choices.

But of course, maybe somebody will offer a kit that does #1. And maybe WildStang (or anybody else) decides to offer their swerve code up to anyone. And maybe in a couple years ultra-light weight versions (with coaxial options) show up. And maybe…

The point is, even though this is not a “full mobility system”, the mechanisms available to FIRST teams are certainly moving in that direction. How will we, as a community, react to this change?

Personally, I embrace it. For many reasons. One, the FIRST component price limit will keep things from getting too out of hand. Simply put, vendors won’t sell what isn’t profitable. Two, we’re always just a GDC curveball away from forgetting about this stuff anyhow. And finally (and most importantly), are students really less inspired when they use a COTS item? If they want to succeed, they will still need to understand it, assemble it, control it, and maintain it. The fact that it was machined using resources not normally available to some of them is irrelevant.

A lot of people on here talk about how FIRST is expanding too quickly and hanging a lot of under supported teams out to dry. For these same people to critique the availability of new components that said teams could otherwise not fabricate strikes me as hypocritical.

:frowning:

Consumer priced swerve modules are also in the works from myself, Sean and RC. Guess we’ll continue that even after seeing this.

Competition is not a bad thing. For anyone.

I agree, competition makes products BETTER and usually winds up making the companies involved gather a larger collective market share than they could individually without competition.

For the somewhat moral issue of having a swerve module that we can just purchase rather than build, I can go either way on this one. First pass at the design looks great though! Once it hits production and is for sale, will you post the CAD files somewhere?

(Pro) We really should be able to push the limits of the cRIO given better mechanical systems relative to what a team can build with a less than perfect shop. Any robotics group these days is all about ‘autonomous this’ and ‘software that’. The group I meet with in Fairfax once a month is all about line following, maze solving, and swerving around cones … forget anything that’s technically complicated to build. Even the robots that win awards at FRC competitions these days seem to do alot with automation, regardless of what the award is for. With a COTS swerve module, I expect that in the longer term we’ll have more capability to attract software and systems engineering mentors to teams who may otherwise not have them due to limited mechanical capabilities.

(Con) In some ways it feels like something’s missing when one doesn’t understand the sweat and toil that goes into designing something if it’s just given as a present for Christmas. Then, 3 days later when it’s broken one may wonder how he/she is going to go get another one rather than fixing what is already in front of them…after all, it wasn’t designed in house, none of the design decisions are understood, and we may very well have hot glue holding on the sprockets if it’s given to the programmers to fix come competition time (;)) . This metaphor is an extremely common example of what happens in industry with COTS items, especially software. I just spent 260 hours debugging a problem that’s plagued us for months in our 6million+ lines of code, and every piece of it had to do with COTS software. 260 hours, plus what others have spent on other problems with it over the last several years … at our equivalent hourly rate, would it have been less expensive to make our own implementation of this software? It is truly hard to tell at this point, but at least I wouldn’t have spent so much time stressing over it at work :ahh: .

Along with progress comes change. I used to have to wait over an hour for a baked potato, but the microwave changed all of that. However, sometimes I still use the oven and wait that hour because I like what I get from doing things “old school”.

Make your own choices for your own team. I ranted on in more detail somewhere else in a similar thread. There in ZERO educational loss here, you just need to think differently about HOW to teach/learn with your team.

Can we also be careful about defining “issue” types? There is no moral issue here at all. There are potential rule, team, and design implications, but nothing at all here suggests any realistic moral issue. So let’s all think carefully when choosing our words.

Bottom line for us is this…
By elevating what ANY team can do, we also push those outstanding teams to go even further. Isn’t that a very healthy thing? Whining on and on about how things used to be is a lost cause in today’s world. Go ahead and spend a moment reflecting on what “was” and honor it for what it was, but don’t waste too much time because it’d be far more productive to look squarely at the present and best decide how you want to proceed.

With all of this said, just because something is made commercially available doesn’t mean it will be legal in future rules. Right? [insert discussion about carts, horses, and patience here as we’re not even sure about a few things yet, are we?]

I also find it a little funny that the whiners in these threads almost never include those outstanding teams and individuals who are/were the pioneers in these areas. Heck, if my team ever designs something so cool that someone else sees a need to mass produce, I’ll dance in the street … then go try and figure out what is next for us.

By the way, isn’t this the way industry and real-world engineering really work as well? I also think that 221 LLC’s honoring of the design origin here is a pretty classy act.

Craig, I see your point. I agree that we all learn a lot more from completing the design process than just bolting on and going… and I understand your concern.

However, it’s not as if teams don’t already use bolt on products. I’d venture to say over half of FRC teams use the KOP chassis and gearboxes every year. Is this really that different of a concept than purchasing pre-designed crab modules? At least with 221’s modules, teams are still forced to design a frame and steering system (or so I’m assuming).

I like the idea of making these available to teams who lack machining capabilities… just how I appreciate the availability of pre-built spur gearboxes, planetary gearboxes, 2-speed gearboxes, and mecanum wheels.

It does sound like this product will likely come in pieces and require assembly, and I’d imagine that students would still gain quite a bit of knowledge from working with the system. Personally, I’ve probably learned more by looking at other teams’ designs than any other process.

Keep in mind that not all teams have a Craig Hickman at their disposal. :wink:

+1

This. Is. Awesome. I would love to see one of these in person (hopefully at IRI)

It’s also good to see that the timeless debate of using off the shelf solutions is still alive and well. I think it’s a great thing to level the playing field for teams that may (or may not) understand the concepts of a swerve drive but do not have the machining ability to create one. As Mr. Taylor said, it’s a great opportunity for teams to acquire these, learn from them and potentially modify them to meet their own needs.

I actually found a picture of this 1868 early Wildstang concept -

http://i2.photobucket.com/albums/y17/Evilstanley/YeOldeStang.jpg

For the somewhat moral issue of having a swerve module that we can just purchase rather than build, I can go either way on this one. First pass at the design looks great though! Once it hits production and is for sale, will you post the CAD files somewhere?

Yes. Team 221 LLC. considers itself a limited open-source organization. We currently offer drawings and solid models for most Universal Chassis products on our website. We will continue this practice with the Wild Swerve.

Consumer priced swerve modules are also in the works from myself, Sean and RC. Guess we’ll continue that even after seeing this.

Please do. Also, If you’re not ready to start a company, source components, build prototypes, invest and endure lots of criticism contact Team 221 LLC. offline. We’re continually designing and considering new product concepts for all types of robot/engineering applications.

Anthony, glad to see this. You do like to cause debates don’t you? Here is my take on the whole issue, no one says you have to use it. If FIRST came out and said that we can’t manufacture anything anymore it all has to be COTS I think we would all laugh at them. Craig if you don’t want to use it don’t. More importantly, who said even if you DID use it that you had to use it for a drive train? Seems to me there are some parts in there that could be used for a turret or a ball shooter.

What is the estimated weight? Sensor options? Motor options? (Can I use a FP through a Planetary if I want to?) Will individual parts be available? 6" wheels are too big, when (if ever) will you get around to offering us an option to use smaller wheels? :wink: It really does look great Anthony, keep up the good work.

These Swerve Modules are pretty awesome… If they would’ve been availible in years past I think I would’ve definitely considered using them.

But this brings up a good question that we as a community may need to address and some of us have already addressed in this thread. At what point in time do we draw the line between what we buy and use and what we don’t?

[quote=
]

FIRST moved past this a long long time ago. Yes there was a time when you could not buy metal sprockets or gears and if you wanted to use them you had to make your own. I think a couple came in the kit for the drivetrain and that was it. Small Parts Inc was the only catalog you could purchase parts from and there was an absurdly low limit like $400 total or something close.

While I’m not a big fan of ready made bolt on FIRST parts, like drivetrains, shifters etc. I most certainly do not want to go back to the days of making sprockets again either. And no, you don’t need a cnc or a laser jet to do that, a bridgeport and prototrac will do nicely.

I think the discussion here is the difference between standard off the shelf parts that you can buy that weren’t designed for a FIRST robot vs. ones that were designed specifically for a FIRST robot.

but…just wanted to point out the fact, COTS items have not always been allowed.[/quote]