A small issue with our Mecanums

[AGWSPilot]

Durring testing before crating up(sadly our only day we were able to get the bot on the ground), we noticed an issue with our drive, when moving sideways, our front wheesls sweep out, esentially making us pivot around the rears, also turning is very poor. I apologize, I have no video of the tests to demonstrate, but I can provide any other information that might be needed. Thank you all in advance for any help you can provide my team.

our wheels are layed out as such (lines reprresent the rollers):

/...............\
/...............\
/...............\

\.............../
\.............../
\.............../

[Bill_Hancoc]

Well i think you have your wheels backward...you need to swap the right and left i think...Talk to M. Krass. She knows a lot and is very helpful.

[GUI]

Is that diagram from the top or the bottom? From the top of the robot, the rollers should form an X, or looking up from the bottom they should form a diamond (<>).

[DB_UPS]

Yes, your wheels are backwards. It should be like this
\.............../
\.............../
\.............../

/...............\
/...............\
/...............\

[AGWSPilot]

Originally Posted by Bill_Hancoc

Well i think you have your wheels backward...you need to swap the right and left i think...Talk to M. Krass. She knows a lot and is very helpful.

Thank you, I can see how that would definitely assist in turning, but how would it change horizontal motion?

I will try to get a hold of M. Krass.

[AGWSPilot]

Thank you all.

[meatmanek]

The rollers touching the ground should be able to roll toward and away from the center of the robot. If you can spin your robot in place without any of the wheels moving, you've got your wheels backwards.

[devaaki]

It helps if you think of the wheels applying forces perpendicular to the axis of the rollers. when the rollers are arranged as you had them, the diagonals were applying forces in the same directions, which led to problems. You can either arrange the roller axis as an X across the bot, or an 'O' around the bot... not quite sure which one is better, as flipping the bot upside down will result in the other configuration.

Good luck

[meatmanek]

Originally Posted by devaaki

It helps if you think of the wheels applying forces perpendicular to the axis of the rollers. when the rollers are arranged as you had them, the diagonals were applying forces in the same directions, which led to problems. You can either arrange the roller axis as an X across the bot, or an 'O' around the bot... not quite sure which one is better, as flipping the bot upside down will result in the other configuration.

Good luck

See my previous post. The axis of the rollers must be radial to the robot, not tangent to it. If the rollers are tangent to the robot, 90 degrees to the rollers will always be pointing out from or in to the robot's center, hence, you cannot turn your robot, or prevent others from turning it.

[AGWSPilot]

Originally Posted by devaaki

It helps if you think of the wheels applying forces perpendicular to the axis of the rollers. when the rollers are arranged as you had them, the diagonals were applying forces in the same directions, which led to problems. You can either arrange the roller axis as an X across the bot, or an 'O' around the bot... not quite sure which one is better, as flipping the bot upside down will result in the other configuration.

Good luck

I get what you are saying, this won't be the first year we have had to swap wheels at the regional, at least it isn't for a new set this time!

[AGWSPilot]

Originally Posted by meatmanek

See my previous post. The axis of the rollers must be radial to the robot, not tangent to it. If the rollers are tangent to the robot, 90 degrees to the rollers will always be pointing out from or in to the robot's center, hence, you cannot turn your robot, or prevent others from turning it.

Thank you. We should be moving much better after we get that working at regionals. Luckily this is our only issue this year.

[devaaki]

Originally Posted by meatmanek

See my previous post. The axis of the rollers must be radial to the robot, not tangent to it. If the rollers are tangent to the robot, 90 degrees to the rollers will always be pointing out from or in to the robot's center, hence, you cannot turn your robot, or prevent others from turning it.

I'm pretty sure that's what I meant, but no worries. Thanks for clarifying.

[AdamHeard]

Also it sounds like one side is outputing more power than the other.

Are you just setting PWMs or is there some feedback to control the speed?

[AGWSPilot]

Originally Posted by AdamHeard

Also it sounds like one side is outputing more power than the other.

Are you just setting PWMs or is there some feedback to control the speed?

Currently, we are just using PWM's. I tried in the code to reduce power to several different permutations of wheels, and it made little or no change to the behavior of the system(it still pivoted on the rear wheels), I don't think the problem lies in wheel speed. But thank you for the suggestion.

[dtengineering]

If you can get some feedback happening you are going to be a lot happier with your handling, I think. We use shaft encoders to ensure the wheels are all turning at the correct speed to give us the motion we want, while I know of other teams who have reported success using the gyro to stabilize their heading.

If you've got the banebots encoders and divider cards on board, we've developed some code that might be able to help, which we are willing to share. Drop me a PM. If not, it is probably easier to just tack on a gyro... perhaps there is someone willing to share their techniques for gyro stabilization of a mecanum drive.

To see some clips of how we move try http://www.youtube.com/watch?v=gyn9e0cT3-A

Jason

P.S. Some posts surrounding mecanum drives insist that you need a suspension to have a working mecanum drive system. While that may be true on a rough surface, we are quite satisfied with just having a bit of flex in the frame.