Swerve vs. Mech?

[yoshibrock]

I know there's a thread about effective drive systems, but I specifically want to discuss swerve drive vs. mech drive. I personally think that swerve drive would work better this year because it gives more traction to push and hold your ground, and it seems like it's going to be a physical game this year. My team voted to use mechs though, so that's too bad for me What does everyone else think? Mechs or swerve?

[Jeffy]

Today we voted for mechs, not because we thought they were "better" but because there is no way we could do swerve with our woodshop.
Swerve does everything mech can do and with better traction. The only tradeoff I can see is swerve is more complex. Too complex? You decide.

[yoshibrock]

Originally Posted by Jeffy

Today we voted for mechs, not because we thought they were "better" but because there is no way we could do swerve with our woodshop.
Swerve does everything mech can do and with better traction. The only tradeoff I can see is swerve is more complex. Too complex? You decide.

Yeah, that's basically the same thing that happened to us. Personally, I don't think it's that much more of a complex system. I also think it's more manueverable because you can do more, programming wise, with it. But it also takes more practice (arguably) to learn to drive it. I agree that it's a tradeoff.

[lexxasaurus]

Our team has been debating this too. The team voted mech but before we finalize that the drive team has decided to prototype both. At the moment I'm kinda torn in between the two because both have strong points. I'll let you guys know how prototyping goes.

[Chris is me]

There's a billion threads on this before but I"ll add my personal inexperienced cents.

All drive trains are tradeoffs, and neither of these are exceptions. Mecanum drives do not respond well to defense, and can have high power draw. Swerve eliminates the defense problems at the expense of a heavier, tougher to build, less durable drivetrain and tricky programming. Given an infinitely good shop / mentors / resources / machining / programming, I'd pick swerve over mecanum always (but a 6 wheel drive is sometimes better than both), but that's obviously not the case.

If you've never built a swerve before this year, don't start now. It's too late.

[kirtar]

Originally Posted by Chris is me

There's a billion threads on this before but I"ll add my personal inexperienced cents.

All drive trains are tradeoffs, and neither of these are exceptions. Mecanum drives do not respond well to defense, and can have high power draw. Swerve eliminates the defense problems at the expense of a heavier, tougher to build, less durable drivetrain and tricky programming. Given an infinitely good shop / mentors / resources / machining / programming, I'd pick swerve over mecanum always (but a 6 wheel drive is sometimes better than both), but that's obviously not the case.

If you've never built a swerve before this year, don't start now. It's too late.

I'll add on to that. Swerve drive also takes a lot of resources. Unless you do coaxial swerve (drive all of the wheels off of a central gearbox), you'll use probably three or four CIMs + whatever you're using to turn the wheels. In addition, it takes up a lot of space in comparison to mecanum drive. This year, where almost all interaction with the ball is below the bumper zone, you might not want to take that space under your bumper zone to put all the components of a swerve drive (then again, you also might decide that it's worth it).

In terms of complexity, as noted above, swerve is pretty ridiculous compared to mecanum. In addition, as a backup plan on mecanum, you can switch out your wheels for traction wheels.

[Alexis Howell]

Yeah. Plus mecanum wheels are really expensive. A set of four from the andy marks site is 700+ dollars.

This year our team is really considering a holonomic drive, we've got most of the resources collected from the past couple of years. Downside's lack of traction. Also worried about how difficult this would be to drive...

Anyone has experience with this drive base?

[Chris is me]

Originally Posted by Alexis Howell

Yeah. Plus mecanum wheels are really expensive. A set of four from the andy marks site is 700+ dollars.

This year our team is really considering a holonomic drive, we've got most of the resources collected from the past couple of years. Downside's lack of traction. Also worried about how difficult this would be to drive...

Anyone has experience with this drive base?

1714 did it in 2008 and it was the worst drive base we ever had. Though part of it was code issues, it was way too easy to push around and extremely vulnerable to defense. The robot would over spin almost every turn. I've heard mecanum is a bit better, but I doubt the team will ever do a holonomic drive ever again for anything.

[Alexis Howell]

Ooh yikes. Our programmers said that they already have a VI for holonomic drive... and we're trying to decide whether it's worth it to build it or not. We don't have the experience or resources for crab.

Worst ever? hmm.

[XaulZan11]

Originally Posted by Alexis Howell

Ooh yikes. Our programmers said that they already have a VI for holonomic drive... and we're trying to decide whether it's worth it to build it or not. We don't have the experience or resources for crab.

Even if you got it programmed 100% perfect, I don't think it is the answer for this year's game. A team in 2007 had a 3 wheeled holonomic drive that they seemed to be able to control very well. On the downside, it was unable to climb ramps and was very easy to shut down with defense (just sit in front of the rack and push them accross the field once they got close). If you don't face defense or don't plan to climb the bumps, then it may be a good choice.

[Alexis Howell]

Thanks for bringing that up. We were planning on just staying in one area, but defense is going to be a definite issue. Driving straight is a problem I hear too.

We want something really maneuverable, but crab and mecanum are pretty expensive... any ideas?

[Bob Steele]

Be very wary of doing anything with mecanum or any type of omni-wheels this year. Two big reasons....

The Ramp...

and Traffic ...

Omni wheels are easy for a team to push around...

As someone said before... all drive trains are trade-offs and they take lots of resources... this year the question will be does a team dare to use omniwheels even if they don't go on the ramps.?..

I think there is going to be huge amount of pushing around on the field this year... especially in the midfield...

if you are using Omnis in a 4 wheel drive put them on the back and then you might be able to go up the ramp... just be really careful when you are coming down the other side... its just so easy to slide sideways with omni wheels...

You will have to be aligned pretty square to the ramp to try it... and hope that no other robot comes along to push you sideways...

My suggestion is you need to power all the wheels... you are going to need the pushing ability... but of course the trade off of using sticky wheels all around is going to be loss of turning control...

so perhaps wheels with a medium amount of traction all around say a coefficient of friction of .75-.8 ... you could turn easier...

Get a variety of wheels and see what you can do...

You can use really sticky wheels if you do a swerve... mechanically they are not extremely difficult...the programming is the hard part...

Don't be set back about the design... we designed our Skunk Swerve last year AFTER the kickoff...

if you need some help .... we are not far away from you guys!!
We can help!!

Good luck

[JesseK]

Swerves are fairly heavy, and to lighten them without ever actually doing one before means a team risks quality. Mecanums are the 'easy' way to upgrade during the actual season, because if all else fails with the code you can easily just tank steer like usual. Additionally there's aren't any tricks to mounting a Mecanum wheel since it goes in the same spot as a normal wheel.

The BEST is definitely crab this year*. Yet just because that's a probable reality doesn't mean your team should do it. It's complicated in both hardware and software.

*AFAIK. YMMV. Only applies to certain strategies. See the other robots at your local competition arena for details.

[Jimmy Cao]

Swerve versus Mech/Holonomic is a discussion that seems to come up every year.

My team did a crab drive last year, and swerve drive the 2 years before that. I can say that they're great for somethings (maneuverability), but at a huge cost.

In 2007/2008, we made the full 8-motor 4-wheel independent swerve drive. All 4 CIMs, 2 FP, and 2 Globes went to this application. What did that leave us elsewhere? Not all that much. The 6 most powerful motors in the it all went to the drive. Worse yet, in terms of pushing power, many (most?) teams w/ a simple 4/6WD with 4 CIMs were stronger. Why? Using high traction wheels (IFI/Pneumatic) would mean that it would be necessary to gear down the steering motors further. If they steer too slow, there's a noticeable delay between going straight and translating. You get one or the other, power of maneuverability.

Another thing that is commonly overlooked is HOW you control the robot. If you have "full swerve", you can control it rotate/translate (the same way you play a FPS game, for example). Difference is, in a FPS (and most other games where you use this control system), the system responds instantly and repeatably. In a robot system, it does not. The end result? It becomes difficult for the driver to fully utilize this system. Now you have a heavy, resource expensive (as far as time, (maybe money), and motors) system that's difficult to program and more difficult to control.

Last year, we downgraded a bit to a left/right side linked crab drive system. With the slick surface, it was unnecessary to put 4 CIMs to drive, since 2 would (in our assessment) always be sufficient to break the wheels loose. This system was really no less complicated than the independent swerve system. When you link sides, you need some way of passing power around. The benefit, though, is that you now can use less motors (minimum of 4 for left/right linked, I suppose you could get away with 2 for all linked). I suppose another advantage is controlability/programming. The control scheme we used last year was simple, left side of drive points in the direction of your left joystick, and the right side of the drive follows the right joystick. In my opinion, it was a little more controlable, but still very difficult to fully utilize (most of the time it was enough to just drive like a tank-drive robot, forward w/ minor adjustments to get balls).

Now, I have never built a mech/holonomic drive system before. I am also quite afraid to go in such a direction. It seems, to me, as if a mechanum drive would, at least, have trouble going over the ramp. I don't think a holonomic drive can do it at all, but that's just my gut feeling. Now, consider the alternative to the ramp, the tunnel. First of all, its narrow. Second, a mechanum/holonomic system would, almost undoubtedly, be unable to (or have great difficulty to) push a conventional 4/6/8/10/12... wheel drive system. If you are really the biggest scoring threat on your alliance, just about any robot would be capable of largely eliminating your ability to maneuver across the field. How big of a deal is that? I can't say for sure. Maybe crossing zones won't be necessary at all, but even then, your inability to stand your ground would prove to be quite detrimental, I suspect.

Just my 2 cents.

EDIT: Oh, as far as the comments about swerve drives being less robust goes, I wholeheartedly disagree. The only year we had issues with the swerve drive was in 2007, mostly due to calibration and chains. If you build a swerve drive, avoid chains if possible, and make a system that, once calibrated, will not lose calibration. If I recall correctly, we recalibrated 0 times during the season and never lost functionality in either side of the drive during competitions. A properly engineered swerve drive is, in my opinion, as robust as a well engineered conventional drive.

REEDIT:

Originally Posted by JesseK

Mecanums are the 'easy' way to upgrade during the actual season, because if all else fails with the code you can easily just tank steer like usual.

While that is true, you can also do the same with a swerve/crab system. I would imagine that swerve/crab is better than mechanums in tank mode because the wheels would have more traction.

[JesseK]

Wait a second ... iirc 540 did Mecanum drive in 2007 as part of the Sparky Triumverate (and won VCU right?) -- thus it should be very easy for your team. This simplicity is probably why they chose it. It'll be practically just as good for your team as swerve since your team will probably have more time to practice with it than they would swerve.