Swerve/crab drive and tank turning
 

HI,
we tested are swerve drive recently and we were wondering how other teams who have created a swerving design made the robot pivot/ or tank turn/ turn on one spot. we right now have one motor turning all the wheels if u can see fromt he picture and have too much tractions. we are not really facing a problem with tank turn but want to improve it. Im wondering if other teams have faced this type of problem and how they have improved it. WE can get rid of the treads and put othe rstuff for them to skid when we piviot with tank drive. All help would be appreciated.
Thanks
Akshay

Re: Swerve/crab drive and tank turning
 

When working on the Wildstang bot last year, we tried to stay completely away from tank style turning as we had to overcome so much friction. While we were on the top of the HDPE we found we were able to perform the tank turning because of the slicker surface. The only other way we found to turn was with monster truck turning. This caused us to build a small drop-down skid which propped up one side of our robot and allowed us to do tank style turning on only 2 wheels and the skid.

With all the wheels linked, you will be limited to what you can do. Hopefully this helps.

Re: Swerve/crab drive and tank turning
 

Last year, one of our three control schemes was tank steering. We had all of the wheel modules pointed forwards, just run the left two wheels in unison and the right two wheels in unison.

Re: Swerve/crab drive and tank turning
 

<EDIT>Chris beat me to the punch!</EDIT>

Akshay

We (111) faced this problem last year. The traction on the wheels made it very difficult to turn in place. While it was possible to do, it caused a great deal of stress to be placed on the drive modules.

We ended up using a pneumatic foot to raise two of the wheels off of the ground to reduce the friction. The foot had a smooth surface that acted like a skid plate, allowing near zero-point turning.

I've attached a picture of the underside of last year's robot. The primary foot can be seen at the bottom of the base. There's a secondary foot on the left for when we went into what we called 90-degree mode. That mode made the stacker side the front of the robot to make the driver's life easier. The stacker was removed at the first regional last year, and I don't think the second foot was ever used after that.

Re: Swerve/crab drive and tank turning
 

Though it was still tested before each match....

I may have won, but you got the picture. I couldn't seem to find one I liked.

Re: Swerve/crab drive and tank turning
 

yea we have the same treads as u must have noticed. It creates a huge stress on the wheel. Also right now we are using the CIMS on diagonaly and Drills diagonaly too. will it matter if we use chips in the front and drill on the back or vice versa? SO basically you guys(111) lifted the back of your base and used the front to tank, am i right? do you think if we take off the tread and use a different form of belting like the black ones from mcmaster( not treads on them, just plain black plasticy not sure what material)

question for 258- did you guys use treads last year i cant seem to remember?

Re: Swerve/crab drive and tank turning
 

Actually we lifted the front of our base and used the back to turn, but yeah...
I'm not a mechanical guy so I can't comment on specific tread types, but from a generic point of view there's simply too much traction. You need to reduce that somehow. Last year we decided to temporarily reduce our traction by lifting two wheels off the ground using a slippery piece of Delrin. Putting something with less friction on your wheels will help, but then you're sacrificing traction. You could also use what we call BUPODs (others seem to call Omni Wheels) in place of two of your wheels, but they're a little complicated to construct. The advantage is they have good traction in one direction but also slip easily from side to side, allowing for easy tank turning. Our 2001 robot had these and it drove quite nicely.

Here's a picture of someone else's omni wheels: http://www.chiefdelphi.com/forums/sh...ght=omni+wheel

Re: Swerve/crab drive and tank turning
 

Akshay,
Our swerve had the drills and CIMs like yours. We used the same tread that you have in that picture you linked to for the Silicon Valley Regional, and we didn't have any trouble with normal tank style driving. We changed to a different read that was more resistant to the wire mesh for Cal Games, but only for that reason. Our modules were geared to go about 10 feet per second on 6" diameter wheels. Our setup also had our two drill modules in the front of the robot, and our two CIM modules in the rear (although we had a mode where we could rotate all the modules 90 degrees, and then change the front of the robot to whataver direction they rotated towards). Not that that mattered much last year.

Here are a couple links to some pictures of our modules...
http://www.seadawgs.com/Pictures/2003/SM1.htm
http://www.seadawgs.com/Pictures/2003/SM2.htm

-Bill

PS - I'm glad to see another CA team tackle swerve in such a challenging year to do so. I can't wait to see your robot in person :)

Re: Swerve/crab drive and tank turning
 

thanks Bill, Dave and other wildstang members.

Bill, yea i think you guys are being able to do tank turns because you had thin wheels. We have wide wheels and the creats sooo much stress we tried to turn. I think its just the fact of too much traction. We have couple of ideas and were gonna try them. thanks for all your help. if other teams tried to tackle the same problem please do post.
thanks
-Akshay

Re: Swerve/crab drive and tank turning
 

Our wheels are 1.625" wide, with the cross hatch tred. Combined with the fact that we are geared to go 14 feet a sec and actually go 12 ft/s ,meaning we have no torque. We just stall the modules when tanking.

What I am thinking to do is get a small pneumatic cylinder with a caster on the end of it to pop down, not necesarily clear the 2 wheels, but decrease their traction. This way, we can turn with only 2 wheels and just spin around on that caster.

Thanks all on your input.