2017 Drive Train

There’s obviously a lot of debate surrounding which drive trains can serve what purposes. This thread is created to centralize this discussion to provide easier access to all teams.

One of the more uncommon drive systems which still proves useful is the h-drive. This system has omnis on all corners, with a central omni to provide lateral movement. Our team has thrown around the idea of using h-drive as a less complicated mechanum system. The movement is near-identical to mechanum and debatably easier to build. This of course being in contrast to using tank drive (reliable and simple). Can we get some insight?

I’m not sure how a h-drive is simpler to build. With mecanums, you simply just mount wheels on each corner. The stock AM14U3 chassis even has support for it with Toughbox Micros. Meanwhile, with a h-drive you have to add a cross member and have a whole different drive wheel and gearbox in the middle. With H-drive you can have three gearboxes (you can tie the two side wheels together), but you need at least 5 motors. Mecanum has four gearboxes but four motors needed. Both offer challenges, but a mecanum drive is a lot simpler to mount to some versa chassis quickly (or c-channel as my team uses) and get started.

This year might be more of a swerve drive game than previous years. More control of the drivetrain can help place gears, and there’s a wide open field for swerve to dance around defense in. It combines the advantages of omni or mecanums (and more) without giving up resistance towards defensive robots.

Too bad I’m not confident in my team’s ability to do swerve properly.:frowning:

Mechanically, H-drive is harder to make because you need the center wheel, but programming-wise, assuming you’re writing your code from the ground up, H-drive is a breeze compared to Mecanum.

Mecanum vs. Tank?

Mecanum has an obvious mobility advantage, but how much does it really give up in traction/strength? Is it completely outclassed in strength to a tank drive system?

If there was ever a game to not overthink your drivetrain and just build a really simple tank drive, this would probably be it.

Agreed. Make it simple strong and compact

In a pushing match, yes.

Everyone should read this before building anything but a tank drive:

http://www.simbotics.org/resources/mobility/drivetrain-selection

Point to consider:
6-WD tank (preferably with a COTS shifter) is the fastest, easiest drivetrain to implement and program. Is a mecanum, H-drive, or swerve so much easier to line up that it outweighs the extra practice time you should have with the tank drive? Or the extra practice time with your practice bot because the tank drive is so much cheaper?

Also, a defensive tank bot is going to shove a mecanum or h-drive all over the place. Good luck scoring a gear while that’s happening.

Aw man, if only I were still with Texplosion, that last part is a pure gem.

Yeah, don’t overthink the drivetrain. You shouldn’t need a vertical adjustment for anything. And, if you are off horizontally, build it into a mechanism rather than moving the whole robot. If you are placing a gear, have a way to rotate it or move it side-to-side. If you are a shooter, have a way to rotate your shooter to keep it on target. Put your code into things like that, not into complex drive.

In 3946’s experience, H-drive is harder than mecanum.

In 2013 Aerial Assault we decided to do mecanum (foolishly expecting a less defensive game), and it worked as well as expected, pretty much the first time around (OK, after fixing some wheel order issues), but we had it running two weeks after we made the decision on our prototype “Woody” (which meant ordering wheels and extra gearboxes before we could finish the build) and running like a champ on the competition robot “Buzz” before bag day. We used the WPI mecanumDrive class straight up, without encoders.

In 2015 Recycle Rush, we decided to do H-drive, and never got it working correctly all the way through CMP. It was only after the 2016 season that we got it working. If you do do H-drive, do not depend on a fixed-height strafe wheel - there is no height which provides the correct amount of load. The central wheel(s) must somehow have a controlled force against the carpet, whether passively through springs or actively through pneumatics or torque actuation, or perhaps some other ways you can come up with.

Another caveat which may or may not apply to you: H-drive requires a wheel near the CoG of your chassis (or one forward and one behind). This will interfere with the “chassis opening and combine” style gatherer that is likely to be commonly used by fuelbots this year.

we are actually gonna use 610s DT from Ultimate Ascent but wooden mobility for us was not a big thing for us this year we want to just be a role player and focus on the gears but still win a pushing a match

For those of us who aren’t familiar with it, mind providing some details about what this is?

Link to picture: https://goo.gl/images/pvbwyc

6 CIM, 6 Wheel Drive, single speed. I believe it was designed around acceleration, not a very high full speed.

Pretty much this if you go mecanum or h-drive you run the risk of getting pushed around a lot as you have to make lots of trips to get gears or fuel

There is a reason that no robot on Einstein has ever had mecanum drive. (2015 is excluded)

So just so some newer teams don’t read this thread and think that mecanum wheels are completely useless at holding their own in a game like this, check out this video from 2014. https://www.thebluealliance.com/match/2014arc_qm29
The bright red and yellow robot (us) were running the new vexpro mecanums (before the metal casters oh how we wish they had the metal insert like they do now) and we easily could hold our own in a pushing contest. The key with mecanums is that the center of gravity needs to be as low and as perfectly centered as possible.
So, do a lot of mecanum drives get pushed around, yes. BUT, if they are optimized, can they hold their own against the most pushy of robots, absolutely!
Just something to think about before teams blindly throw it out because they are under the assumption they can be pushed around with minimal effort.

Frankly, in that video I see your robot pushing around other mechanum/omni robots, and using its momentum well to get the occasional hits on the tank drive 'bots. Your driver is doing a great job, within the physical limitations of the system… those limitations are minimized by the low CG & good system engineering, but still definitely there.

One was a swerve, one was a mecanum, and the third was a 6 wheel drive. But often times its more about how you drive the robot and how limitations are overcome such as using momentum. I wish I had video of it, but one of the matches during our first regional (which weirdly only a few matches were recorded) we ended up pushing an 8 wheel colson drive across the field from almost a stand still. Can mecanum drives push very often, not really, but should they be eliminated from decision discussion simply because there is a little defense, in my opinion I definitely do not think so because a good drive team with a slick drive train can do surprising things. So ultimately the point I guess I’m trying to get at is just because a drive train doesn’t have as much physical grip as another doesn’t mean it can’t at times match or outperform a technically gripper drive train.