pic: Traction test



We ran some traction testing tonight. 2 Vex robots, same frame, same wheels, the robot on the left is four motor drive, one motor per set of tandems wheels. The robot on the right is four motor drive, two motors into a 1:3 sprocket to wheel ratio. The wheels on the robot on the right turn 300 rpm, the left robot 100 rpm. We ran the test on a piece of FRP. The weight of the robots was the same, except for the extra tandem wheels.

The robot on the right won every time. Why?

I do not know much about vex motors, but I am willing to bet that due to the less available torque for the one on the right due to the inverted gear ratio that it’s acceleration was below the acceleration needed to break free.
Breaking free means reduced friction coefficient …

This is a very interesting test. Thank you for sharing.

Just so you know, surface area in no way at all effects friction. It can effect traction though where traction is the sum of all forces you use to propel your robot. One non friction force that is included in traction is “Gripping” force. Carpet to rubber has “gripping” force which is caused by the carpet meshing like gears with the rubber. “Gripping” force is effected by surface area, more grip points more resistance. On this years surfaces, however, it is only friction.

The reason friction does not go up with surface area is that as you add contact point the weight per point (pressure) goes down, netting a even friction force. The formula for friction Ff=µ(Fn). Where Ff is Force of friction, µ is coefficient of friction and Fn is normal force (is this case weight of the robot). Te left robot most of one because of discrepancies in the wheels that made it have slightly more friction with the floor

tl;dr: Surface area doesn’t effect friction, winning robot was blind luck.

darn

and i thought we were clever.

I would say keep both at the same gear ratios and number of wheels.

what happens when the only variable is the gear ratio (same # of wheels on both)?

I agree with Copioli here, and the results are corroborated by our own experiments. We noticed that a single CIM driving a single rover wheel with all the weight on it (120lbs total or so) running on a Toughbox had much better acceleration than the same robot with two CIMS into two Toughboxes running one wheel each (with the same 120lbs).

We then measured the actual maximum tractive force provided by each setup, and discovered that there was no noticeable difference in the total force available. The reason for the difference, as we determined, is that since all the weight was on one wheel, that wheel itself could source twice as much force (and thus double the torque), but a single CIM could only input half the torque of two CIMs. Therefore, the single wheel accelerates better because it essentially has built in traction control, as it takes a lot more to put enough torque through the single wheel to slip it.

I assume you’re going full power with your two robots, in which case the robot with the lower torque will be slipping the wheels less (or maybe not at all), and will have more traction than the robot that has more torque and is slipping all the wheels.

Both robots were running at full power. All wheels were slipping right from the start of the test. We also, to add another variable, added approx a 1/2 lb weight to each robot, the tandem wheeled robot, then the single wheeled robot. In each case, the single wheeled robot won the pushing match. We did notice about a 1 degree increase in surface temp of the wheels during the pushing mathces, which quickly dissipated upon shut down.

Tonight is the next phase, dual wheels and 1:3 on both robots.

This sounds about right to me. Equate it to getting your car moving on a slick surface (ice). If you start in 1st gear you will most likely spin your wheels because the engine doesn’t have to overcome much resistance to turn the wheels. Start in second gear and now your engine has to overcome the gearing before your wheels will spin. (This works great on a standard but not so easy to do with an automatic unless you have second gear start, my old '96 Grand Prix had it & it got me out of a few tricky spots.)

A few thoughts about this -
Given a perfectly level surface and a rigid chassis - how many contact points would there be on the 4 wheels vs 8 wheel? 3 points determine a plane.

So, given that the chassis and wheel assembly might very well flex a bit, and the surface may not be perfectly flat - the # of contact points and the amount of force could vary between the 4 wheel and the 8 wheel.

As stated earlier, the normal force will be divided among the number of contact points - contact points are more relavent in this case than surface area.

I believe the difference that made the 4 wheel win, is due to the 4 wheel having more contact pressure per wheel in contact, and that the gearing reduction caused that machine to slip less than the 8 wheel. I know you indicated that both machine slipped right away, but I don’t think that both slipped at the same rate. My guess is that the 4 wheel slipped slightly less.

Just my 2 cents worth,
Looks like you’re having fun

Mike Aubry

Static vs dynamic friction
Static inline 0.06
Dynamic inline 0.05

4 motors allows the wheel to reach top speed faster, then the two. The Cof, increases until it peaks at static, then approaches 0( or some value, that is less then static friction), as the wheel speed increases. Since the 4 motors reach top speed faster the can’t exert their larger force, the 2 two can since they accelerate slower due to the lower torque.

Also, the force of fiction is CofP (pressure Fiction) by having two wheels per side driven you decrease the pressure P (P=F/A, force/area), the force is the weight of robot, a large area will decrease your ability to move, by increasing the force of fiction. This also hurts the 4 motor design.

Thats two answers to your questions, heres the solution:
TCS, Traction Control System.

Are you using this year’s acetal tread Rover wheels or some other variation?

It would be interesting to see the test results with this year’s acetal tread Rover wheels on the FRP…

Update:

We rebuilt the tandem wheeled robot to include a 1:3 gear ratio. So now both robots run at 300 rpm wheel speed. I didn’t have my camera with me last night so I do not have any photos or video, but the single wheeled robot won the shoving match every time.

Conclusion: 4 wheel drive is the way we are going, maximizing stability and power to the ground.

Thanks for the update. We have a running discussing going here about whether having two wheels at each corner for a total of eight wheels is better than just a single wheel at each corner (total of 4). I say it doesn’t help and that having a wheel at the center would do more good because it would take some of the weigth off of the corner wheels, which would help turn.

There is a video posted on Youtube by the Nutrons, (Team 125 at Northeastern Univ) of them driving around their 2007 frame outfitted with the Rover wheels on a piece of FRP. Don’t worry about unloading the corner wheels to facilitate turning!

I couldn’t help to notice that none of the videos on Youtube that have rover wheels appeared not to be weighted down to simulate the typical 150 pound robot … the actual weight factor may affect how these rover wheels perform on that surface…

Also, I wonder if the conditions of that field service will get better as one and one half days of robots drivinig on them during the competition…

Hopefully we will have our practice field ready for testing tonight…