Drivetrain First Thoughts?

[Chris is me]

Originally Posted by noa_n_f

What is everyone's opinions about a swerve drive for this year's challenge? You would have the maneuverability of a mechanum drive but also the traction of a tank drive. Is it a matter of difficulty when it comes to building and programming that prevents teams from using swerve drives or are there other reasons?

That and it's a lot of weight, time, and resources that could be better spent on other parts of the game (for a low resource team like mine).

[Michael Corsetto]

Originally Posted by Kevin Sevcik

If your argument it that scoring in 2007 was easy with a 6WD, you might want to look closer at how teams were doing it. It looks to me like the 6WD bots were approaching the scoring pegs from the side to line up and drop. Which worked really well with a widely spaced round grid.

This year's grid is flat, and more tightly spaced. I think approaching a peg from the side is going to be significantly more difficult than it was in 2007.

Good point. However, from a driver perspective (I drove 3 competitions in 07) we used the side approach often because it was simply a straight line from where we were to where we wanted to score. Also, space was limited between the rack and the side walls, so scoring sideways was almost required since turning and going straight on would require careful positioning. Remember how two bots couldn't even fit through the gap between the rack and wall sometimes?

Side note, holding the tube horizontally allowed us to come at the spider from almost any angle (straight on or otherwise) and score with equal effectiveness and reliability. This year, scoring straight on, or close to it, will definitely be more common, as there is essentially a lane that forces a right angle approach.

Back to the thread. My first thought, 6wd. Second thought? 6wd. You get the picture

[davidthefat]

I am 100% certain on my design. Its just a simple 4 wheel drive. I will post a picture of the robot later on when it starts coming into shape. It will blow your mind what can be done with something so simple as a 4 wheel drive.

[exprg:melonhead]

Here's what 2181 has come up with:

We decided on an omnidirectional drive, so that we can let the robot, in essence, drive itself. When we get to the guidelines, rather than us lining it up, have the robot set itself to angle 0 (via gyroscope) and follow the lines from there. Little room for human error and we'll always be facing the wall with omni drive.


I don't think we'll lose on speed but has anyone ever thought of acceleration being an issue with the omni drive? Just a thought since there isn't as much power in those wheels.

[MrForbes]

Originally Posted by davidthefat

It will blow your mind what can be done with something so simple as a 4 wheel drive.

We had 4wd our first year. We also had a reliable autonomous (so we could get the bonus and usually win matches and qualified highly), and great scouting (so we could get two great 6wd robots to help us) and we won a regional.

If you haven't driven a few different robots, including mecanum, 4wd, 6wd, etc then you might want to do a bit more research before deciding on a drive configuration.

[JVN]

Originally Posted by davidthefat

I am 100% certain on my design. Its just a simple 4 wheel drive. I will post a picture of the robot later on when it starts coming into shape. It will blow your mind what can be done with something so simple as a 4 wheel drive.

Hi,
What about a 4WD will blow my mind? I've got a pretty good idea of the physics involved, so I tend to understand the tradeoffs pretty well...

In a narrow configuration:
You'll have either higher turning scrub than a 6WD so you won't be able to turn as well, OR you'll need to reduce lateral traction on some of your wheels (and likely your primary traction as well) OR you'll need to reduce your wheel base which will make you less front/back stable OR you'll need to shift your CG drastically forward/back which will make you less front/back stable.

In a wide configuration:
You'll be MUCH less front/back stable and wider (which makes it harder to get through traffic).

So yes, you can make a 4WD work great for the game -- but there are tradeoffs involved. Typically those tradeoffs involve reducing traction (at least partially) or reducing wheel base or reducing stability.

A "rocking" 6WD allows you to keep full traction while keeping full wheelbase (because it essentially has two short wheelbases that it alternates between. The tradeoff there, is the added complexity of additional wheels.

Note, this post only applies to "skid steer" robots commonly known as "tank drives" -- when the wheels can turn it is a whole different animal.

Please try to remember that you're preaching (in part) to a group of veteran robot builders and experienced engineers on this forum. Conduct yourself accordingly, and you'll get additional opportunities you can't possibly imagine.

-John

[pfreivald]

A note for rookie teams:

It isn't uncommon for teams to get to competition for the first time ever, and then learn that their skid-steer robot cannot turn very well (or in some cases at all!)

It is CRITICAL that you test the turning ability of your robot ON CARPET comparable to the carpet used in FIRST games AS SOON AS IT IS BUILT, so that you can make modifications if necessary. (4WD with omni wheels on the rear are simple, lightweight, have good forward/reverse pushing power, zero turn radius... and unfortunately can be spun like a top when pushed laterally in the rear.)

[Chris is me]

Originally Posted by pfreivald

It is CRITICAL that you test the turning ability of your robot ON CARPET comparable to the carpet used in FIRST games AS SOON AS IT IS BUILT, so that you can make modifications if necessary. (4WD with omni wheels on the rear are simple, lightweight, have good forward/reverse pushing power, zero turn radius... and unfortunately can be spun like a top when pushed laterally in the rear.)

For what it's worth, 4wd with omnis does have a small front-biased turning radius. They are vulnerable to being pushed, but they are probably the simplest drivetrain a team can build. Low resource teams should look into this drive if they feel they don't want to take any chances with their drivetrain this year. And really, when do you want to take a chance on the most important part of your bot?

[JVN]

Originally Posted by pfreivald

A note for rookie teams:

It isn't uncommon for teams to get to competition for the first time ever, and then learn that their skid-steer robot cannot turn very well (or in some cases at all!)

It is CRITICAL that you test the turning ability of your robot ON CARPET comparable to the carpet used in FIRST games AS SOON AS IT IS BUILT, so that you can make modifications if necessary. (4WD with omni wheels on the rear are simple, lightweight, have good forward/reverse pushing power, zero turn radius... and unfortunately can be spun like a top when pushed laterally in the rear.)

Make sure you test your robot turning on carpet at full weight, as well (including bumpers, battery, and mini-bot).

-John

[Brandon Holley]

Originally Posted by JVN

Make sure you test your robot turning on carpet at full weight, as well (including bumpers, battery, and mini-bot).

-John

Ditto, especially with a potential of 160+lbs of weight on the robot this year.

-Brando

[Lil' Lavery]

FWIW, if you're testing early enough, you can play around with where that weight is to see how it impacts your drive as well.

-Sean (joining the cool kids club of signing my post outside of my signature).

[JesseK]

Originally Posted by Kevin Sevcik

If your argument it that scoring in 2007 was easy with a 6WD, you might want to look closer at how teams were doing it. It looks to me like the 6WD bots were approaching the scoring pegs from the side to line up and drop. Which worked really well with a widely spaced round grid.

This year's grid is flat, and more tightly spaced. I think approaching a peg from the side is going to be significantly more difficult than it was in 2007.

Your post really made me say 'hmm'. It's a very valid concern.

Yet after a bit of thinking, I think robots will not NEED to approach from the side this year due to the very large differences in the plates on the front of the pegs. The room for error this year is quite a bit bigger for the Circle, and still rather large for the Square.

[PVCMike]

I am surprised to see so many people referring to Mechanum drives as being slow. In reality, yes, they are inefficient but that doesn't make them slow, gearing is what ultimately determines speed. Like others, I think there is merit in exploring them as an option. Elegant driving can quickly get you away from even some of the best defensive bots.

Not saying we're going with with this system, just saying I don't think it's something i would write off.

[Ether]

Originally Posted by PVCMike

I am surprised to see so many people referring to Mechanum drives as being slow. In reality, yes, they are inefficient

What do you mean by the word "inefficient" in this context?

[PVCMike]

Originally Posted by Ether

What do you mean by the word "inefficient" in this context?

I mean that you lose energy through a vector other than forwards. Those mechanum rollers are rolling even when you're going straight, so you're losing energy and it could be taking you more wheel rotations to cover the same distance as a non mechanum wheel of the same diameter.