I saw a passing refrence to a very interesting (and obvious) design I wish I’d thought of this year. Basically the robot has skids on all four corners and two wheels in the center, allowing it to turn quickly and minimize the friction put on the skids. I am aware this has several downsides, but I’d like to see pics of robots that use this drives system. Anyone? If your team used it, how did it work out?
We went with this drive system our rookie year. It was… interesting.we had casters on all 4 corners of the robot, though. We had the casters lifted atleast half an inch above the drive wheels to keep us from high centering, even though the field was flat in '98. This lead to the robot constantly tipping back and forth when it drove, and I imagine was a pretty odd thing to see and drive. I think this would be a problem with most drives of this type, because you need to guard against high centering as much as possible, and this would make you rather tippy. So I’m slightly doubting the wisdom of this choice since FIRST seems to have a penchant for non-flat playing fields lately.
Isn’t is possible to put the casters on springs to always keep the wheels even?
A robot like this is necessarily an inverted pendulum. You can’t stabilize it and keep the CG directly above the wheels without a Segway like control system. You could, however, purposefully shift the CG to one side of the wheels and keep it low. This would keep your robot leaning to one side the majority of the time, except when you’re accelerating or something.
Yes, it is.
In practice, you’re better off with some delrin type skids or small wheels. The ones that 885 used this year were litterly made out of some plastic I found in the shop and done by eye in about an hour. They ended up workin surpringly well, so they stayed on instead of getting replaced by more professional examples.
Lets see about pictures…
This link shows one side of 885’s drive train for 04. Extreamly simple. The only real problem we had with it was one that was ultimatly tracked down to a 40 amp breaker that was opening far to soon. At 8fps the speed diffrence in the drill motors became annoying. It wasn’t unmanigable, but enough of a problem that our driver wanted a solution. A PID loop was installed to compensate, using encoders and some simple programming. After that, she drove wonderfully. You can also see the skids, on the inside edge of the frame. They appear as light brown blocks, much like the plywood frame. They are acutally made out of ‘stallpoly’ (the stuff used to make bathroom stall deviders). A bolt is acutally threaded through them and holding them on (I love this plastic!). One was ripped off late in the season and replaced with a Delrin copy. As far as I can tell, they were quite sasifactory for a quick hack job.
This shows the location of the wheels. The center of the wheels is where those gray dots (encoders) are. The balence came out surpringly well, favored to the front (ideal, for our needs). We didn’t really try for that, it just work it’s self out that way. As long as you can rock the bot back and forth with out much effort, your fine.
This shot shows 95’s bot from 2002. You can see the wheels near dead center of the bot, acutally a little rear of center. The skids on this bot are also very flashy (as was the bot period). Instead of just plain skids, delrin ‘wheels’ were made, and mounted on plain shafts. Their shape allowed them to ‘bite’ into the carpet when going forward and helped keep the robot tracking straight. When turning, they provided enough friction to keep the bot from over rotating. This is by far the easiest to handle 'bot I’ve ever driven.
Here’s 95’s 2000 bot. The drive wheels are hidden, but you can pretty much figure out whats going on. I can’t recall what the skids used were, but they don’t stand out much, so probably just small wheels or some delrin.
Also, 95 used 2 wheels this year. They havn’t uploaded any pictures from this year yet, but it is just a refined version of 02’s drive.
Obviously, there is a little bit of family in all these bots. 
I have loads more pictures of 95’s 2 wheelers and 885s latest. Just let me know if you want to see them.
-Andy A.
And I just can’t resist…
We are currently just finishing up the build on a robot similar to what you describe. It has 4 omni wheels, one on each corner that are unpowered (although they could be powered) Then in the middle, there are 2 NPC gear motors that have a dual output shaft (shaft goes al the way through the motor and goes out both sides) There is a wheel on each side of each motor making for 4 powered wheels in the dead center of the robot all on the same axis. Anyway, it drives pretty good and turns like a spinning top. It is amazing how maneuverable that thing is. Sorry I have no pic.
our (60) 2004 design was a little different, but I think it does the same thing. It had six wheels with the two center wheels being 3/16 of an inch lower than the other four. It was very maneuverable and reliable. If you want to check it out, look at http://www.team60.com (I don’t have a direct link)
I have many pictures of the 263 2 wheel drive robots from 2001 through 2003 if you are intereted you can IM me for some more detailed ones and things you want to make sure you don’t do.
2002:
http://firstrobotics.net/02gallery/pages/263-1_jpg.htm
Wow, Andy A, those are some of the most aesteticly pleasing bots I’ve ever seen! Looks like Ikea-esque white pine!
Very cool post, just what I was looking for
Thanks 