My team may end up building and refining a drive train over the summer. The goal is to have a bullet proof and reliable design that we can learn from to make a better drive train suited for next year’s game. The decision on what we could build seems to have come down to either a swerve drive, or a wheeled tank drive. Which of these designs would you favor for education?
For some context: We are what I would consider a high resource team (full shop with lathe, mill, band saw, etc.) However, despite our resources we don’t have a whole lot of collective experience (many more freshman and sophomores than seniors and juniors.)
I’ll do a full write up later. But in short:
In favor of education? Swerve. You’ll learn an incredible amount, both on design and machining principals, to controls and programming purposes.
That being said, if you’re looking for a no fail option for next year and years to come, I urge you to simply look into a WCD style or 8 wheel drop like 2056 runs year in and year out.
If my team had the money and resources to develop whatever drivetrain we wanted I would be pushing for Swerve. It’s the best of both worlds when it comes to traction and maneuverability plus it’s just so much fun to watch. (16’s bot looks like a hovercraft while on the field.) It will be very educational as I hear it is very complex on both the mechanical and programming ends of things. Also as 16 has shown with Stronghold, you can use Swerve even with crazy terrain obstacles.
Now of course there is nothing wrong with a wheeled tank design as it is very robust but I don’t feel you will learn as much unless you really get into testing different designs and/or building a multi-speed gearbox.
While building a good, traditional WCD is great for future years, I will always be one to recommend swerve to anyone who wants a design or programming challenge. A good lightweight swerve is definitely a challenge, although with work, I think it is very rewarding. On the design side, it teaches some good design principles, and on the programming side, it teaches vector math and how to use PID. So in short, for your situation and with your resources, I would recommend working on a good swerve drive.
What is the history of drivetrains that your team has used?
Define SMART goals. Compile a list of lessons learned, research and build on your team’s baseline.
Personally, I would avoid building a swerve.
If your goal is to have a bullet proof and reliable design within one off season, and your goal is to learn many practical things that can be applied to real build seasons, build a quality tank drive. You’ll be reaping the benefits of that prototype for many, many seasons.
Swerve is, quite frankly, a trap. A flashy, cool trap that is extremely hard to execute, and even when executed well is rarely a big advantage over a tank drive. Only the very top teams in FIRST can make great swerve drives both mechanically robust and reliable in software.
Agreed. The reward for a well put together swerve drive is minimally higher than that of a robust tank drive, and swerve is much, much harder to pull off. A tank drive is more than enough for most FRC teams and its simplicity will allow your team to focus on more complicated design challenges (scoring).
I’d recommend developing a robust tank drive first for use in future competitions and then exploring swerve after you’ve done that. A tank drive like a WCD will perform perfectly fine for nearly all FRC games.
I too agree, Five seasons with a wheeled tank drive or modified version. Always bulletproof, never failed. Always stock drive train gearboxes. Last year modified center wheels and added an h-pair in center of robot. This year was by far the most abusive game we’ve played and we ran 8" aluminum performance wheels from AndyMark and aside from some slightly bent wheels and axle bolts we had 0 failures. Our drivers literally beat the living crap out of the bot. While our design may be considered simple, we have never had a drivetrain failure and most teams in our district can’t believe our drivetrain is all stock parts every year.
We live by two rules:
Build a bot that doesn’t break
If it breaks we’ll fix it before the next match.
Unless you have enough money to iterate your swerve at least twice ($2000+), I would not recommend it initially despite being a big proponent of swerve. If you want a guaranteed and strong drivetrain for next year, just make a quick WCD using Vex parts. You don’t need to overthink it, just use the sliding bearing blocks and cams with some big holes and you’ll be fine. If you want to take it a step further, use the WCP sliding bearing blocks instead, or go direct center to center with the 17 tooth sprockets from Team 221.
Once you’ve got a physical drivetrain you’re sure will work for next year, then you might be able to focus on swerve. You will want to overthink your swerve design, to a point- going with a COTS option like the Revolution modules from Team 221 would be my recommendation so the programmers can figure out what they’re doing. That also lets you iterate cheaply because the modules are easy to integrate into different designs.
Unless you have access to waterjet or CNC mill, I would not work on a fully custom swerve. Just my $0.02
I have always figured that if I didn’t need to support the outside, I won’t. WCD lets me use lightweight 2x1 aluminum tube is basically a box shape, whereas if I had to add a plate on the outside for support I would have to spend space and extra pounds while mounting it via spacers and bolts- not to mention the added machining time. Plus with WCD I can run chain-in-tube and save even more space. There’s nothing wrong with a dead axle or outside support drivetrain, but I feel like WCD just offers more benefits than risks. I’ve never had issues with either drive when made properly.
As a driver I love swerve. Once you can wrap your mind around driving it it feels so natural and simply fun to drive. Swerve also comes with the benefit of the fact that it doesn’t matter what way you’re oriented for where you’re driving. This helps on offense and defense both. Offense wise, you can get to where you want to go in one fluid motion, even with defense being played on you. Swerve also allows you to shake defense rather easily as you aren’t limited to forward and backwards because you also have side to side movement at your disposal. Defense wise, it is outstanding. Swerve’s mobility allows it to be an amazing defense drive train as you can very easily stay in the way of the attacker (reference 16 on 2012 Einstein). Swerve has many benefits if you can build and program it right.
That being sad swerve is not for everyone. It has its limitations as does everything else. Tank drive beats it in certain areas and it kind of comes down how you want to play the game in my opinion. It’s fairly simple to mechanically build but programming is a beast. It also takes a while to execute a swerve (We’ve been doing swerve since the early 2000’s and are still improving our design). If you plan on doing swerve drive I would suggest getting advice from swerve teams as well as being very devoted to the project and continuing to improve on each design.
We ran a 10WD WCD and had no issues with it all season, landing on colson wheels over the defenses. WCD is very robust and easier to make now than ever before with all of the options available COTS. Machining isn’t even that needed depending on how you make it.
I have to very strongly recommend against swerve if you aren’t confident in your ability to build a tank drive. There are going to be years (like this year) where swerve just isn’t a practical option for everybody that isn’t team 16.* There are also going to be other years where swerve would technically work, but doesn’t fit your strategy well and so you would be better off putting the resources elsewhere. But as long as the kitbot is a tank drive, FIRST is never going to design a game where you need an omni-directional drivetrain to be competitive.
I think there were a couple of other robots that ran swerve. I’m not sure if any of them saw much success with it, though.
The offseason is a great time to test all of those seemingly crazy ideas you had during build season. I believe that the best thing you can make during the offseason is a tank drive for multiple reasons. One being that it is likely going to be your next drivetrain going on a competition field. Having a spare drivetrain allows you practice weeks 1-2 of build season, while having a spare for drivers practice. Want to build something else? You now have a somewhat realistic robot for mounting things. In the past few days, we have repaired our spare drivetrain so we can practice driving with our favorite style of drivetrain. This will lead us to making a manipulator, it being an arm, elevator, or whatever we want that fits. Use this project to catapult yourselves into a better position.
What if the OP chooses to purchase COTS swerve modules from AM or 221 Robotics? Given the very long runway until the next season, do the opinions change?
My opinion of “it’d be better, competitively, to really nail a 6wd and drive the heck outta it” doesn’t.
That being said, some folks aren’t about maximizing competitiveness and are about doing something cool.
So, it’s priorities - there’s very few times in which a swerve is going to offset the cost (weight, money, complexity) from a competitive perspective but there’s plenty of times in which it brings other benefits to your program as a whole.
Some WCD additions that accomplish both:
2 speed (auto shift)
Current limiting
Path planning with encoders and gyro (Math and PID fun)
Vision assisted aiming
PTO
Octocanum
If you want to take it a step further, use the WCP sliding bearing blocks instead, or go direct center to center with the 17 tooth sprockets from Team 221.