I love the work that was put into this elevator and I think it will benefit many teams, but at what point is too far when it comes to buying somewhat of a sub-system. In no way am I saying this elevator is bad for FRC but more interested to see what you guys think about buying whole sub-systems like climbers, drive train, elevators, etc.
If it helps “inspire young people to be science and technology leaders and innovators,” then I’m all for it. And I think the ability to buy a while subsystem does that.
8 Likes
That ship sailed several years ago… the GreyT elevator was released in 2019 (or was v1 in 2018?), there’s also the GreyT shooter and GreyT turret products released in late 2019. Just like COTS swerve, COTS products like this tend to raise the floor, as it enables teams without machining capabilities to build better robots. An elevator like this is still not a complete subsystem–the team needs to figure out how to best use the pieces and parts and interface it to other things they are building. “Advanced” COTS availability enables teams to spend $ instead of mentor/machining resources to get to a similar level of capability, and the latter is often more difficult for teams to get. And just seeing the COTS products designs can be a source of inspiration–I’m sure there are teams that didn’t buy the GreyT turret, but rather used it to figure out how to build better turrets of their own.
12 Likes
last I checked, they just hand these out for free now, so…
13 Likes
Well I’d like FIRST to be less COTS oriented, and to see more innovation and creativity. I’d also like a pony for Christmas and that ain’t happening either.
FIRST has changed. So has the industry it mirrors. I can’t let my origins as a junkbot builder hinder the success of the team. Besides, Ryan and the Thriftybot crew are great to work with. The degree to which they are willing to help teams out is in itself inspirational.
11 Likes
If mods could split this off into another thread that would be great.
To quickly summarize my view:
COTS mechanisms raise the middle tier into upper tier. They do not impact the floor except via second-order mechanisms (seeing more successful elevators). Limelights, COTS mechanisms, COTS swerves, all raise the amount of money a team needs to spend to remain competitive relative to their peers. Most teams do not know that COTS mechanisms even exist, except for swerves and possibly Limelight.
I have a longer post somewhere that I might dig up.
9 Likes
(respectfully)
The KOP chassis sure as heck isn’t free. You pay as part of your registration fee, and if you opt-out, there’s a lovely $450 AndyMark voucher.
4 Likes
Having the ability to have a crucial complex mechanism given to me in a way I can use out of the box or modify how I see fit is amazing. This also allows teams to focus more on other subsystems, such as a collector and work on perfecting that rather than worrying on how I’m gonna get my elevator or telescoping arms working
2 Likes
I would argue 100%, getting a CAD of another team’s robot, and using it to build a duplicate version…is one of the most productive, thoughtful, and engaging learning experiences students (and mentors) could have with respect to programming, design, construction, and testing (driving).
Speaking from experience…
If we never had said design, the learning curve of getting to the same point after going through that 3 month experience would have either been too high or we would have given up. We decided to use Java this year for the 1st time ever, and built a robot that didnt use pneumatics at all, something we never did before.
The misconception with a lot of people is, there is no learning involved in doing such a thing. I’d equate this topic to our experiences as being very positive. We hope to do it every off-season moving forward.
19 Likes
With all these COTS things, the game has changed (and I really like it).
Now we get to decide what we put effort into, and what we don’t, but will still have a highly functional robot. For example, we can buy our SDS modules since it’s not worth our time or money to design our own. It’s less time, and less cost (with all the r&d we would have to go through).
The best teams will still customize many things to keep themselves the best. Maybe a 3 state elevator is better? Maybe continuous is better? If they have the resources, they get that advantage, but anyone can now buy the ttb one and still have an elevator.
Is there a line? Not sure, but probably not with the rules as they are . Even if someone wanted to sell full robot kits, the logistical challenges are going to make it worse than a clever team. A company would have no design until after kickoff, then have to be able to manufacture more than 1 rapidly, then figure out shipping. A good team that only has to build 1 or 2 robots is going to have the better final product.
6 Likes
In many ways, it makes FIRST resemble the “real world” of engineering, where “make vs. buy” is a very real consideration.
Given a budget, internal capabilities (people, machines, etc.), and the required schedule/available time, which “things” should we design and make ourselves, which should we design and have others make, and which should we buy off-the-shelf? If we buy a solution, what is required to integrate it into the rest of the system and make it perform well? Can we support/maintain it as well as if we’d built it ourselves? Would we be better off to start out by buying a component and later transition to a more optimized approach once we learn how it performs?
Notice that all those questions could apply equally to an FRC robot, or to a professional design project.
Of course we are here for students to learn - but let’s not think so narrowly as to limit that to learn mechanical design. Consider a team that has strong software students but is struggling on the mechanical side. Having COTS mechanisms and sensors might well allow that team to field a robot that is at least mechanically adequate, which their software folks can perhaps make into a very competitive machine through clever code. And in the meantime their mechanical students are learning how to assemble and integrate and maintain those mechanisms, which will make them much better for next year. Everyone on the team is going to have a better, more inspirational season.
Respectfully, I disagree. I’ve seen too many teams, especially rookies, awfully discouraged with barely-working homemade mechanisms. Something as simple as a COTS shooter or climber-in-a-box takes them from a “floor” of having little gameplay capability, to being able to contribute meaningfully to their alliance and come back with smiles on their faces. That’s a win, whether they won the match or not. And there’s not necessarily a high cost involved.
I don’t believe this is true. COTS mechanisms are available from major FRC vendors and in widespread use; even a casual Google search finds these things immediately. “FRC elevator” turns up at least 2 COTS solutions in the first page of results. I saw plenty of climber-in-a-box-es at events in 2022.
To a point. Swerve is probably the poster child here, but then we come back to the perennial not every team needs to have a swerve conversation, which I won’t rehash here. Putting swerve aside, some of the other commercial mechanism options are going to cost a comparable amount (even just in parts and materials, not factoring in time spent) to a similar in-house design.
Regardless, without the COTS solutions there is a vast gulf between the “peers” with advanced design, machining, and sponsor capabilities and large teams of students to carry out the work, and “everyone else” with much less. Having a bunch of more capable teams thanks to their COTS mechanisms makes for more interesting gameplay and events, and again a better experience.
In the end teams are each going to make different choices, based on a lot of factors. I’m glad that available COTS solutions increase the options.
7 Likes
How many rookie teams have you seen that make use of COTS mechanisms? I can think of extremely few low resource teams that know about them. The perennially low-competitiveness teams scoff at COTS mechanisms or simply don’t know they exist.
Swerve is a huge competitive advantage. If your team doesn’t have swerve, they’re being left at a significant disadvantage. The time investment to returns ratio is very good. This is not the same as it was ten years ago when making a swerve required tons of machining and design knowledge, all for something that probably didn’t even work as well as an MK1. This is raising the price of robots. Yes, it obviously makes robots better, but in a competitive setting that means the teams that don’t have $3k to drop at a disadvantage.
Before 2017, teams spent an average of < $80 trying to get vision processing working, and few teams had it working well. Now, teams spend $300+/yr on Limelights and maybe half the teams make it work trivially without even understanding vision processing. It’s pure competition.
My point is, COTS mechanisms can’t be beat. 98% of teams that choose not to use them are giving themselves a competitive disadvantage. The logical step beyond that is that spending money on COTS mechanisms is the most efficient way to become competitive. Which means that the price of a competitive robot goes up. The GDC could simply design games that are slightly easier to complete and ban COTS mechanisms, but as long as we have climbers in a box, tasks like the 2022 monkeybars are a lot more viable in game design.
3 Likes
I work as a software developer. I can program using Java. I don’t understand how the underlying structure of Java works or how it translates to a lower level of code. Is that a problem? I can do my job effectively and I can create what is asked of me. If I wanted to, I could go out and learn that but for my job I just want to know what is necessary to get the job done.
If a team feels it’s important to learn the underlying structure of vision, that’s their prerogative. Same with building a swerve, designing a climber, etc. I don’t want to force my students to learn the intricacies of building a WCD if they don’t want to. Not every team has the same needs as yours or how you want to teach your students.
If they want to, great! But I’m so happy we don’t have to spend the first competition troubleshooting the issues with having manufactured our drivetrain since we didn’t find the KOP drive to fit our needs, leading to immense frustration in the team.
3 Likes
Having all the COTS options makes FRC more accessible to more teams and people on the teams. Buy vs. build is a valuable thing to learn to contemplate. A team with few members that can only meet 8 hours a week might find it best for them to use the KOP chassis, use a COTS climber-in-a-box, and use an COTS elevator kit (with some customizations for fit)… and then the ONE fully custom-designed and iteratively refined part of the robot being the intake/ejector mechanism connected to the elevator. That could be the best way to use their time and resources, and if they tried to do more customized pieces than that & failed due to running out of time, then they could end up being little more than a bare chassis. I think the program wants these teams, and the students on them benefit from the experience. (i.e. I agree with Anthony)
1 Like
My point is that this used to be a competitive advantage. Now, wasting time learning about machine vision processing is a competitive disadvantage. You have to sacrifice competitiveness to learn, and I think that’s tragic.
You didn’t have to teach your students about vision processing back before Limelight either. You just also wouldn’t have vision processing - just like almost everyone else.
This is a great example. While this lets lower resource teams play at a high level (assuming they have money and know about all of these products), what happens when those COTS mechanisms are so good that only top percentile of teams doesn’t benefit from using them? What student is designing and building an elevator that’s finished faster and is more reliable than a Greyt V3? Who’s making the custom elevator that beats out COTS options? Who’s making a custom swerve and reaping any competitive benefit? Innovation is discouraged when COTS mechanisms are so good that they can’t be beat by anyone with less than 4 years of robot experience. Teams should be allowed to design their own mechanisms without being almost guaranteed to shoot themselves in the foot as a result.
2 Likes
Possibly we were the exception, but I’m not sure we would have known to search for “FRC elevator” when we were rookies. After the first season if we saw one at a competition, then maybe.
2 Likes
I’m not so sure about the “almost guaranteed” part. I think plenty of teams will still design their own elevators if we have a lifting/placing game. Swerve is an unusually complicated thing to do yourself, but having it is so fun and valuable, that I’m sure most people are loving the serve COTS. Your work with alternative co-processors for vision and similar work by others proves there’s an opportunity to do something custom that outperforms COTS (i.e. Limelight). As has been the case, the teams that design the best game piece handling mechanisms and software solutions will reign. That part’s not going to change I don’t think.
How many custom tube elevators do you think outperformed COTS ones this year? I saw less than 10 all year, and except for the ones on Einstein level robots, none of them were better than COTS. My favorite example of the “ultimate COTS robot” is 5499 2019. An absolutely stunning robot that was based around a Greyt elevator kit that ran circles around everyone except 254 at Silicon Valley Regional. Most teams just have too much pride to use COTS solutions, like 1072 in previous years. That doesn’t change their value proposition.
I would bet the Limelight 3 outperforms it, actually. This is only possible at all because we don’t need those annoying green lights anymore, and the community has put a huge open source effort into software. This will never be true for anything mechanical. And ultimately, teams are still going to spend $100-300 on a custom solution that they would never write themselves.
That’s the top 2% of teams, and I’m not too concerned about them. Let them pursue their custom solutions and benefit. I’m talking about the middle and lower tiers when I say that COTS can’t be beat.
1 Like
I agree with the premise more or less, but not necessarily with the conclusion. But that’s based on how important the competitive aspect is, which doesn’t need to be re-argued again. If it is all about the competition, then I would agree that there’s very little incentive for most teams (but not all) to go the custom route, except when money is an issue. But I’ve never seen FRC as being all about the competition, and I think it’s good to have COTS options for those times/teams when they would like to shorten development time.
I don’t think innovation or customization is discouraged, but maybe some competitiveness might be sacrificed, especially on the lower end. You can choose to go custom any time you like. I don’t think FRC has been “solved” either and there will be more times than not that COTS mechanisms are unable to win games on their own.
1 Like
Before COTS mechanisms, teams could choose to learn and benefit competitively as a result - mostly. Now, using a COTS mechanism is a choice between competitiveness and learning/having fun. It’s not “all about the competition”, but designing something that’s COTS should not come at the expense of competitiveness. It shouldn’t be futile for most teams to make their own elevator, a basic mechanism.
There’s limits; I don’t think teams should be casting their own aluminum or milling their own gears. But there’s a world of difference between that and buying an entire mechanism off the shelf. COTS aluminum and gears are available everywhere. FRC-specific ones just make them cheaper.
4 Likes