Andy mark mechanum problem

So my team is having some issues with the Andy Mark 6" mechanums. They are in the correct orientation (X from the top) and they spin in the correct direction when they are on jacks, but they refuse to strafe side to side. Instead they follow a diagonal curving path. (Forward and to the left for a left strafe and backward and to the right for a right strafe).

Things we have tries:
Calibrating the Jags
Testing the output voltage on jags
attaching extra batteries to try different weight distributions (but i’m not sure the weight distribution is not the problem)
the rollers seem to roll fine
one of the mechanums wobbles when it runs and another is at a slight angle in relation to the frame (<10 degrees off)

Has anyone experienced this problem before?

Thanks

-baronep

When the robot sits on a level surface, are all the wheels equally touching the ground? Could the frame be twisted?

If the frame isn’t warped then maybe its a problem with the gearboxes not being broken in properly… I can’t think of anything else since you say that the wheels spin the correct directions. The last and final thing you might want to try is switching joysticks. However, the most likely cause is an imbalance of force being transmitted to the surface since the mecanums operate on the principal of vectors

if you are running on a flat frame, or a out of square frame–I expect you are experiencing a torquing issue with the frame when power is applied causing frame warpage not visible to the eye but causing all wheels to not contact the the floor evenly–have seen this.

Also, make sure that your rollers have some same level of movement. Some may be tighter than others. This might cause some of the problems you describe.

cims gain a bias in one direction after a year or two of use, if your cims are 2 years old i have found 12-15% difference between forward and backward, if you cannot get new cims, you might have to account for them not being the same forward and backward.

We have found the Jags are biased in one direction. Reverse the polarity of the motor leads not the power leads on the Jag and account for it in the code.

This isn’t true - and if it is, calibrating them would have fixed that.

the cims are bias, the jags are perfect, the rotor in the cims magnetizes the more you use them, so if you have old cims, the don’t spin with as much power in one direction.

Sounds like you know what the problem is. It probably is one of these.

Another option, when i drove mecanum last year, I first thought that I was pushing straight sideways on the joystick, but was not. This could be it aswell.

We used mecanum in 2008. It helped to have each wheel in a spring loaded suspension so they always had the same pressure on the floor regardless of the squarness of the frame (or the levelness of the floor).

http://www.chiefdelphi.com/media/photos/30031
http://www.teamtitanium.org/media/2008/buildseason/target26.html
http://www.teamtitanium.org/media/2008/buildseason/target28.html

PS. It was the fastest robot we’ve ever built.

Are team is using it this year.

Some keys that we have found:

-Evenly distribute weight
-Make sure the code supports straffing
-Make sure everthing is fastened properly (the wobble could be binding when trying to strafe.)

This and a number of minor tweaks that can be pretty self-intuitive should get you back on track.