Drop-center drivetrains: Why?

[CENTURION]

So, this may sound extremely n00bish of me, but I can't seem to see the advantage of a drop-center 6 (or 8) wheel drivetrain over other types.

I understand the point of dropping the center wheel: to reduce friction by making the robot rock so that two wheels are off of the ground. But I can't see why you wouldn't you just have a 4-wheel drivetrain, only 4 wheels are going to actually be touching the ground anyway.

We went with a 6 wheel drive this year, but we lowered the friction by putting omniwheels on the corners of the robot.

Is there some secret purpose that you could enlighten me of?

[NickE]

With a fixed robot width and fixed wheels all facing forwards, the shorter your wheelbase is, the easier turning will be. This is because the wheels have to scrub less to make the turn.

With a 4WD robot, the wheelbase is quite long, causing difficult turning.
When a 6WD dropped-center robot, only the front four or back four wheels are on the ground at any given time, effectively halving the length of the wheelbase, resulting in easier turning.

[Garret]

With a dropped-center 6 or 8 wheel robot the wheels contacting the ground when turning are closer together than on a 4 wheeled robot. Because the wheels are closer together it makes it easier to turn. This is one of the reasons it is easier for wide base robots to turn.

[Chris is me]

All of the below assumes a roughly centered CG and even tractioned wheels (all roughtop or wedgetop or kit)

With four equal and high traction wheels on the robot in the long orientation, turning on a dime is basically impossible. It can be done with large wheels and a lot of "jumping", but it's close enough to impossible that the robot's basically worthless.

Adding another set of wheels with no drop fixes this a little bit, but generally not enough without some specialization. 25's drivetrain is an exception due to a lot of subtle reasons I don't really want to go into here.

Dropping the center wheel means that effectively you have shortened your wheelbase by half, since only 4 are on the ground. This makes your wheelbase wider than it is long so you can turn just fine. As a bonus, when you're pushed on one side it's very easy for the robot to "lean" on those wheels to make you harder to spin than, say, the same drivetrain with omni wheels, leading to great pushing resistance.

With a 6 wheel no drop omnis, you get comparable performance and even better turning at the cost of being more easily spun yourself. 4 wheels with 2 traction and 2 omnis exaggerate this effect further. 6 wheel drop is basically the very best of most worlds, for most games.

---

In the case of a wide body robot, Lunacy demonstrated you don't _need_ 6 wheels to turn well.

[CENTURION]

Ah, I see now, of course!
I guess I underestimated how much friction the 4-wheel drive created...

Thanks a ton for the insight!

[Chris is me]

No problem.

I felt really, really bad not including "the math" supporting the claim, so here's a link to an excellent paper on the topic of drivetrains that explains all of the phenomena. It never specifically addresses the case of the 6 wheel drop, or what happens if you lower the sideways traction of one of your wheel sets (i.e. if you put omnis on one part of your drivetrain) but it shouldn't be too hard to see how a 6 wheel performs better than a regular drivetrain as a result.

If we're engineers, I feel we shouldn't be content with the what without the why...

http://www.chiefdelphi.com/media/papers/1443

[Ether]

If we're engineers, I feel we shouldn't be content with the what without the why...

ditto that !

[AdamHeard]

Originally Posted by Chris is me

If we're engineers, I feel we shouldn't be content with the what without the why...

We're not engineers, small correction. Statement is true otherwise.

[Al Skierkiewicz]

Simply put, without dropping the center wheels, drive motors rapidly go to stall current in turns.

[IKE]

Do a little more digging. A common questiont that many 6x6 drop center guys will ask is "How much drop are you using?". While a fair question, the same amount of drop will have various effects on different chassis. CG placement and Chassis stiffness, and overall goals/objectives play a big part into what is considered effective amounts of drop.

[JamesCH95]

Originally Posted by IKE

Do a little more digging. A common questiont that many 6x6 drop center guys will ask is "How much drop are you using?". While a fair question, the same amount of drop will have various effects on different chassis. CG placement and Chassis stiffness, and overall goals/objectives play a big part into what is considered effective amounts of drop.

As an example, 95 has used dropped center designs when the robot did not have precise manipulators mounted on the chassis, where the robot rocking back-and-forth did not adversely affect controllability. On two robots that DID have bigger arms that needed to be controlled with precision we used omni wheels on the four corners so that the robot was more stable and the arm was easier to control.

EDIT: To avoid confusion, I am talking about a 6wd platform with two traction wheels in the middle, and omni wheels at each corner.

[JesseK]

Originally Posted by JamesCH95

... On two robots that DID have bigger arms that needed to be controlled with precision we used omni wheels on the four corners so that the robot was more stable and the arm was easier to control.

Anecdotal evidence would suggest that rocking drive trains still maintain the maximum manipulator control aspect given a proper c.g. As an example from 2005, 254's robot had a rocking 6WD with a very long 2-PID arm. When fully extended the arm would shift the c.g. slightly, yet when it was down the tetras were easy to pick up because the c.g. stayed towards one side or the other when the robot was not accelerating.

Unfortunately it is exceptionally hard to integrate c.g. from scratch ahead of time; rough estimates are the best I've ever seen derived at the early stages of design. Since drive train is one of the three systems that MUST work (drive train, electronics, programming) in order to do anything effective in a game, most teams prototype like crazy on a pre-season prototype drive train. Additionally, finishing the production drive train early in the season is atypical of actuality for most teams.

Ergo a wise suggestion to anyone worried about c.g. shift on a rocking drive train is to have a pre-season prototype ready to go such that manipulators may be directly mounted to it during build season prototyping. This will give earlier estimates on c.g. and should produce constraints on shifting c.g. that may dictate where to put the 'heavy' elements (compressor, battery, drive train gear boxes, etc) on the production robot.

[JamesCH95]

For all of the analysis that could go into a chassis like that it is simpler to use omni wheels on the corners and be done with it. That frees up all the time and energy saved in chassis analysis for use in more unique challenges.

[RyanN]

Originally Posted by Al Skierkiewicz

Simply put, without dropping the center wheels, drive motors rapidly go to stall current in turns.

And when drive motors rapidly go to stall current, that means...

1. Robot doesn't turn
2. Breakers pop
3. or Worse

We've been using 6WD with dropped center wheels for years (with the exception of last year).

We like using pneumatic center wheels. Makes adjusting the offset very easy.

[Chris is me]

Originally Posted by AdamHeard

We're not engineers, small correction. Statement is true otherwise.

aspiring to be engineers

Originally Posted by JamesCH95

For all of the analysis that could go into a chassis like that it is simpler to use omni wheels on the corners and be done with it. That frees up all the time and energy saved in chassis analysis for use in more unique challenges.

I really think "just use omnis" can be a pretty poor solution depending on the game challenge. If you have a long arm, a gentle tap moves your game piece several feet away and there's not much resistance to speak of. Really, it's not that hard to do "all of the analysis". I can tell you what the analysis will say right now: The more rigid your chassis, the better your 6WD can perform. Building a rigid chassis doesn't take tons of time and energy. Even if you do want to do careful and thorough analysis for such a drivetrain, it's not exactly a huge resource drain, and depending on the challenge it could result in a much, much better chassis for your needs.

While some games you probably could get away with a chassis with four omnis, in the past I think more often than not you couldn't. Being spun off target in Aim High, knocking you away from your spider leg in Rack 'n Roll, or taking a mid to long range shot in Breakaway are all situations when I'd rather have a bit more resistance to being spun.