So last night I was thinking of a way to improve my team’s drive train, when I had an idea. We usually use a 6 wheeled tank drive, with 2, 6" plaction wheels in the center lowered 1/8", with 4 standard FIRST wheels on the outside, with each wheel connected by chain.
My idea was this: A similar drive train, using 6" PLaction wheels in the center, lowered 1/8", but with Andymark trick wheels on the outsides instead of the standard FIRST wheels.
Has anybody tried this? Anything to change? Should all wheels be connected via chain, or only the center wheels powered?
We’re going to try a WCD next year, so will that affect anything?
What are you trying to improve? What is the problem you are trying to solve?
From the AndyMark web site:
Trick Wheels are the original AndyMark omni-wheel. The other, newer omni-wheels (Aluminum omni wheels and Plastic omni wheels) are improved designs, made from lessons learned after applying the Trick Wheel to various applications. These wheels carry a high price because inventory on them is kept low.
If you implement a 6-wheel skid properly, you shouldn’t need more maneuverability. Take a look at 254’s drivetrain if you want a WCD; they are very maneuverable without sacrificing the vital side-to-side static friction that allows you to resist pushing. In other words, you should probably go with all traction wheels.
As for the wheels, basically pick any omni wheel that isn’t an andymark trick wheel. Personally, I like the plastic ones. They aren’t too expensive, they’re robust, and I haven’t heard of any common failures.
Be careful. There is a point at which the robot becomes much too easy to turn, and you will start having issues with the simple things like driving straight or driving while being defended.
Also, make sure you experiment ahead of time. All 6" wheels are not exactly 6", and all 4" wheels are not exactly 4": they sit on the ground differently. We found this when we didn’t pay attention to the 4" plaction and the 4" omnis from Andy Mark.
we did exactly this, and i would recommend against it. 6 wheel drive with the outer wheels as omnis turns way to fast and is difficult to control without a lot of driver practice.
AM trick wheels are not what i would recommend, the plastic omnis are excellent in the 6 inch size, and they are also not quiet 6 inches, they are actually 5.94 if i remember right, which means you don’t need to make your center wheel lower.
if you want to make a tunable robot, i would take a look ate 1986 team titanium’s robot from this year, they used an 8 wheel tank drive, with omni wheels on the ends, and were amazing, because it still has 4 traction wheels, it doesnt turn too easy, but turns better than a standard tank drive.
you might try reducing your gear ratio some if you want to turn better, even just 5 percent will make all the difference in the world.
if you want to use omni wheels on the ends, use 8 wheels total, 6 turns way to easy and its easy to get spun, and the plastic omni wheels are the best option as far as cheap, versitile, durrible, and lightweight.
Edit: here is a video showing 1986’s drive train, if you want an drive with omni corners to learn from, theirs is very good.
Contact me and let me know when and where your team is meeting. I should be able to help with this. Your mentors should have all of my contact info. If not, just PM me here and we will set something up.
If you do opt for the omnis, you’ll want to remove the drop from the center wheel.
The reason why: if you leave the drop in there, you shortened your wheelbase (the point of the drop), while simultaneously making it easier to turn the resulting 4WD (the point of the omnis is to make the 6WD easier to turn). It’ll most likely try to act like a 2WD, which is really easy to turn, but might not be the best with traction.
What I might suggest instead would be to:
Increase the center drop slightly. This will pick the “up” set of wheels up just a little bit more, resulting in less scrub from them.
2) Change the center wheels to pneumatic wheels. This will allow for slight variations in the drop without permanent modifications, and, depending on the wheels you use, it may increase your center traction. (It may also be fairly simple to test out during the offseason–make your modifications and drop them in.)
If you’re going for maneuverability, you can’t ignore the other 2 fundamental concepts that help your maneuverability (i.e. helping you to turn). The Poofs (et al) understand these fundamentals so well that they keep implementing similar designs year after year that are tailored for them. Quite honestly, I don’t know if there will ever be a skid-steer configuration superior to WCD since WCD plays so well into these concepts while maintaining a very simple, efficient approach to the design. Note that the KOP is also setup for both concepts, yet no kit drive train can force you to put your center of mass in the right spot ;).
Concept 1: Wheel base is wider than it is longer.*
This is achieved with the 6WD drop center with since only 4 wheels ever touch the ground in the right configuration for our robot dimensions.
In a 6WD corner-omnis configuration, this is somewhat achieved by the reduced lateral frictions of omni wheels.
Concept 2: The center of mass should not be close to your center of area.
For a 6WD drop-center there are 2 centers of area depending on which 4 wheels are on the ground – it’s either in the middle of one or the other set of wheels. For a 6WD with omni-corners, it is dead in the middle of the robot. To find the center of area overall, draw lines from wheel to wheel – where they meet in the middle is the center of area.
By placing the center of mass in the dead center of the robot with a 6WD drop center
, the center of mass is never the same as the center of area. The center of mass is also far away from the sides of the robot, which minimizes tipping issues. - With 6 wheels on the ground, special care has to be take to move the center of mass away from the center of the robot – which then may cause tipping issues if moved too far. Moving the mass towards the corner omnis also puts more weight on them, meaning more lateral forces that play into Concept 1.
You can get by with corner-omni placements and ignoring the 2 concepts. However, you will be much better off by at least understanding the concepts and somewhat adhering to them. Robots that turn easier are not only more agile at mid-high speeds, but also draw less current overall during a match when compared to non-optimal configurations (all else equal).
Omnis on all four corners make things difficult–more difficult to control, and lack of ability to resist being pushed or turned.
Perhaps you could put omni/trick wheels only on either the front pair or back pair of wheels and remove the 1/8" center drop. This maintains all 6 powered wheels on the ground at all times, removing the minor rocking due to the drop center, but it increases the manueverability without loosing contol. Teams 48 and 3193 did this with great success this year, using custom 4" treaded wheels (modeled after the AM peformance wheel) with 4" omnis up front. The robot was very maneuverable, very stable, and had no issues being pushed/turned.
Two notes: we made the custom treaded wheels 1.5" wide instead of the standard 1", and we located our main mast/arm at the center of the 4 traction wheels.
Tip of the day: Do not waste your time trying to improve drivetrain performance when the kitbot is so good!
My team ran a 6WD Kitbot with 4" Plaction Wheels and Supershifters from Andymark. It was the most reliable and effective drivetrain I have ever been involved with. If you preped the parts prior to build season, it literally takes a couple hours to put together and build (lead times for parts were the only reason we were just driving in week 2). I have seen lighter drivetrains, but the weight we had on it was only 2" off the ground, dropping CG of the robot.
Focusing on the mechanisms that get you points will make you more competitive than building what is essentially the kitbot, but lighter. Engineering is all about picking the most efficient areas to make gains.