We are currently checking out ideas for our robot’s drive base and the idea of mecanum wheels came up as a main candidate.
We’ve been looking at videos and saw team 2848’s design (http://www.youtube.com/user/SteveJesuit2848#p/a/u/1/UQQ67SB8Acc).based on the wheel’s specs in Andy mark, The wheel’s friction coefficient is 0.7, which means that a robot using these wheels will not be able to accelerate while on the bump, since the coefficient is smaller than the necessary coefficient to accelerate on a slope of 45 degrees ((tan(45) = 1) > 0.7)) . Although our calculations indicate otherwise, It seems that 2848’s drive base was able to climb the bump easily, And so, we wanted to ask whether it’s us that misunderstood the problem and solved it wrong or that 2848 might have changed the wheel/carpet/bump/chassis somehow to make it easier to pass.
There is a chance you may be right. We haven’t tried it with 150 lbs yet, however, in the good old boy system of “reckoning”, “I reckon it will work” Still if not we have a work around for that. Just to put that question to rest, we’ll try to weigh it down and see if that makes a difference.
But Steve, weight shouldn’t have any effect on your effective pushing ability to get you up the bump. What with the whole friction increasing in direct proportion to weight thing. Weight distribution may cause you problems if your CoM is too high or too far forward or rear.
I think the problem with the math is assuming 2848’s robot is operating on a 45 degree incline. It’s clearly never entirely on the 45 degree incline, so it never has to exert the theoretical .707 * mg force to keep it from sliding down the ramp. It always has a pair of wheels on the ground or the top of the platform, or straddling the bump. So it’s never actually in a position that matches the physics land scenario of all four wheels on a 45.
Well, after putting some more thought into it, Kevin’s answer for the coefficient issue seems to be the right one.
It seems that our team will use mecanum wheels as a drive base this year, and so we wanted to ask 2848 some more questions.
First, in the video (http://www.youtube.com/user/SteveJes…/1/UQQ67SB8Acc) your robot uses different wheel brackets then the one given in the KOP. Is there any special reason for that?
Second, you didn’t use the Thoughtbox from the KOP. Why did you change the gear?
hello all
have a question about mecanum drive system…
on a prototype we have, we are able to drive fwd,rev and srafe (sp?) sideways. however when we try to do a 360, or even 180 the robot jumps/jitters and barely rotates(<90deg.). im assuming this is because of a high frictional coef. between the carpet and the rubber wheels, but it just seems like a weird/cheapo wheel design(the 3 small rubber rollers as wheels)?
we are using the andy mark 8" mecanum. all rollers are pretty loose, and even. No random tight roller to bog us down.
any thoughts on this are greatly appreciated.
oh and this is on a level (actual) field carpet ~20’x20’ area.
This can happen if you have the wheels swapped. The wheels will try to push toward or away from the center of the robot instead of spinning the robot. If you can turn it by hand without any power applied, you definitely have the wheels swapped.
We’re using wheels I used for an experimental system put together several years ago. For that system I used 12:1 Banebot transmissions adapted to a T-Slot frame. This allowed for easy chain tensioning. The chain path mandated the wheel sprocket be mounted with spacers. That means they wouldn’t fit in the standard kit of parts wheel brackets. During testing I decided it had too much speed so I re-geared the transmissions to 16:1
We still haven’t had the opportunity to try the bump with full weight on the chassis. We have been moving our building site and that had put our tests on hold. I THINK it will still work but if not, there are workarounds yet to explore.
Figure 4 is definitely wrong. I think that robot would have real difficulty going straight.
Figure 2 might be wrong, or it might just be a bottom view. Since it’s showing the force vectors, it actually makes sense to be showing the spot where the wheels contact the floor.
OK, agreed, if it’s bottom view it’s OK. If it’s top view it’s not going to spin in place well. The side vectors will be trying to spin in one direction while the forward/reverse vectors will be trying to spin in the other.
Having done the calculations, and real world testing, Team 1716 has found that the 8" mecanum wheels will successfuly go over the hump. Unfortunatly we did have to modify the KOP wheel brackets. Once that had been accomplished, we had no problems!
Just curious… How are you planning on configuring your operator interface for controlling the mecanum wheels?
e.g.
Joystick1 y-axis controls fwd/rev for all 4 wheels and
Joystick1 x-axis controls right/left slide and
Joystick2 x-axis control right/left spin
OR
Joystick1 y-axis controls fwd/rev for left wheels and
Joystick2 y-axis controls fwd/rev for right wheels and
Joystick1 x-axis controls right/left slide
OK FINALLY got the chassis together and loaded the bot to 165 lbs.
These are standard 8" Andy Mark mechanum wheels. 16:1 transmissions, the bump is built correctly and the carpet is reasonably close to the competition type. Here’s the result:
Kevin, I think you may have hit on a very important point. At any given time two wheels seem to be on a level plane and not the 45 degree angle of the bump.
Thanks for that insight