Mecanum forward rotation error

[mattachew12]

We're having auto-disable troubles with our mecanum drive system. Using RobotDrive's cartesian system, we can move in the XY plane fine, and our left rotation can receive full power, but whenever we push right rotation (positive side) to 38% power, all four jaguars disable simultanously and require a disable/re-enable to work again. The ports don't matter and neither does the axis on the controller. The Jaguars are fine because they all go 100% during every other movement, just not positive rotation. Currently we have a functional robot by dividing the rotation inputs by 3 so it never hits above 33%, but I'd prefer to cure the cause rather than the symptom. I think it is a problem in RobotDrive, any ideas?

[shindigo]

I have witnessed jaguars cut out when they are under excessive load so look for something mechanical that may be causing problem with rotation in that direction. Btw, we have also found that victors do not have this problem/feature.

[Joe Ross]

I suspect that your digital sidecar isn't properly powered. It gets some power through the DB-37 cable, but not enough to send a full speed PWM to 4 motors. Assuming that your right motors are inverted in code, a left turn is minimum PWM to all motors and a right turn is full PWM to all motors.

Disconnect the DB-37 cable from the digital sidecar and make sure that all 3 power LEDs are brightly lit.

[mmaunu]

It sounds like your wheels aren't spinning in the correct directions. We had similar problems a couple of years ago. Have you inverted the motors on the left side (or right side) like Joe mentioned? This ended up being the cause of our woes; we hadn't inverted any of our motors.

First, are your wheels mounted correctly? When you look down on the robot from above, do the front rollers want to roll in towards each other when the robot moves forward? Do the back rollers want to roll away from each other? Here is a top-down view:

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

Put your robot up on blocks and look at the way the wheels spin (again, look from above).
1) When you direct the robot to go forward, all wheels should spin forward.
2) When you direct the robot to go backwards, all wheels should spin backwards.
3) When you direct the robot to strafe right: the right front wheel should spin towards the back of the robot, the right back wheel should spin towards the front, the left front wheel should spin forwards (towards the front), and the left back wheel should spin backwards (towards the front).
4) Test again with strafe left: left front should spin towards the rear, left rear towards the front, right front towards the front, and right rear towards the back. (The wheels diagonally across from each other should turn in the same relative direction.)
5) When you rotate clockwise, the right wheels should rotate towards the back of the robot and the left wheels should rotate towards the back of the robot. (The wheels on the same side of the robot should turn in the same direction...and each side of the robot should turn in opposite directions.)
6) When you rotate counter-clockwise, the right wheels should rotate towards the front and the left wheels should rotate towards the back.

I am not sitting in front of a robot so take those directions with a tiny grain of salt. Somebody feel free to correct me.

Again, it sounds like a symptom that we had to track down before. Try inverting the left motors but not the right (or vice-versa). This can be done in code (Java and C++ have method calls on the speed controllers and LabView allows you to set this property as well).

I hope that this helps. Mecanum is fun to drive ~(_8^()

[mattachew12]

This problem is occurring even when the robot is on blocks and the wheels aren't bearing any weight, so I'm fairly sure it isn't mechanical. The wheels are in the correct configuration with the from top X from bottom O, and the signals to the wheels are all correct in regard to strafing, rotating, etc. I haven't checked the sidecar recently but we have been having troubles with those lights. We already switched out one sidecar because of it. Any ideas what can be done to keep that part of it from breaking? I noticed on tank drive that this also happens in one specific scenario, when all four motors are putting out negative power and more than two go above the 38%, which even though we have positive axis rotation on mecanum the jags are also putting out four negative signals.

[mattachew12]

Update: We've isolated the problem to the lights on the sidecar. Something is breaking the 6V light. Every time this disable problem happens the sidecar's 6V light is off. When we switched a new sidecar in, the three lights turned on and we got our full rotation functionality back, but today we just broke the light again by rotating too much apparently and somehow overloaded the sidecar. It auto-disabled again and now the 6V light won't come on again. What I don't get is jaguars shouldn't be affected by the 6V since their PWMs use 5V.
What breaks a sidecar's 6V capability?

[Joe Ross]

Update: We've isolated the problem to the lights on the sidecar. Something is breaking the 6V light. Every time this disable problem happens the sidecar's 6V light is off. When we switched a new sidecar in, the three lights turned on and we got our full rotation functionality back, but today we just broke the light again by rotating too much apparently and somehow overloaded the sidecar. It auto-disabled again and now the 6V light won't come on again. What I don't get is jaguars shouldn't be affected by the 6V since their PWMs use 5V.
What breaks a sidecar's 6V capability?

The only thing that uses 6v is servos (when the jumper is installed next to the PWM). I suspect that you are not getting power to the digital sidecar. Did you remove the DB-37 cable and look at the LEDs like was suggested earlier?

[mattachew12]

The only thing that uses 6v is servos (when the jumper is installed next to the PWM). I suspect that you are not getting power to the digital sidecar. Did you remove the DB-37 cable and look at the LEDs like was suggested earlier?

When we take off the cable the 5V and RSL lights turn off. What should happen?

[Ether]

When we take off the cable the 5V and RSL lights turn off. What should happen?

How are you powering the sidecar?

[mattachew12]

How are you powering the sidecar?

We have the DB-37 cable attached as well as a WAGO connecting to the PD board with a 20A fuse.

[Joe Ross]

When we take off the cable the 5V and RSL lights turn off. What should happen?

They should remain on. Did you wire the digital sidecar per the wiring diagram? http://www.usfirst.org/sites/default/files/uploadedImages/Robotics_Programs/FRC/Game_and_Season__Info/2014/ElectricalLayout.pdf

[Joe Ross]

We have the DB-37 cable attached as well as a WAGO connecting to the PD board with a 20A fuse.

If the lights went out, and you are wired to the PD with a 20A fuse, there is a problem with that connection. Maybe one of the wagos is crimped on insulation rather then bare wire.

[Ether]

We have the DB-37 cable attached as well as a WAGO connecting to the PD board with a 20A fuse.

Like this (http://www.chiefdelphi.com/forums/attachment.php?attachmentid=15847&stc=1&d=1389920269)?

[mattachew12]

They should remain on. Did you wire the digital sidecar per the wiring diagram? http://www.usfirst.org/sites/default/files/uploadedImages/Robotics_Programs/FRC/Game_and_Season__Info/2014/ElectricalLayout.pdf

Yes we have, the square white connector is plugged in to the DSC under the (-) and (+) symbols, and the red is with + and black with -. They are plugged into the PD board under a 30A slot with a 20A fuse.

[Ether]

We have the DB-37 cable attached as well as a WAGO connecting to the PD board with a 20A fuse.

Have you checked the fuse?

[Ether]

Yes we have, the square white connector is plugged in to the DSC under the (-) and (+) symbols, and the red is with + and black with -. They are plugged into the PD board under a 30A slot with a 20A fuse.

If the fuse is good, unplug the white connector and use a voltmeter to check if there's voltage at the connector, with the correct polarity.

[mattachew12]

If the fuse is good, unplug the white connector and use a voltmeter to check if there's voltage at the connector, with the correct polarity.




How do we check if the fuse (it's really a circuit breaker, right? I mislabeled it?) is good? I'm also signing off now, I'll be back tomorrow evening.

[Ether]

How do we check if the fuse (it's really a circuit breaker, right? I mislabeled it?) is good? I'm also signing off now, I'll be back tomorrow evening.

Just swap it out for a known good one.

If that doesn't work, then disconnect the white connector at the sidecar and check to see if you've got voltage coming out the end of the cable.

[mattachew12]

So it isn't the wiring as far as we can tell. We replaced the wires and used a new 20A breaker, still no 6V light. We are getting 12V at the end of the cables connected to the PD board. The DB-37 cable keeps the 5V light on by itself, and the WAGO keeps the Bat light on. Both of them together will turn on the RSL. The 6V light would come on when we used a new sidecar, so since the wires were capable back then I believe they still are now. When the sidecar broke, the only change was that in the situation in which the disable would occur, we were able to run the motors in that situation for a few minutes while all three lights were on, and then the jaguars disabled themselves and we noticed the 6V had turned off. So I don't think it is the wiring since it was capable of running the robot and the sidecar before it broke. Something physical changed when everything stopped working, and my guess is we broke the sidecar, but what we don't know is how sending four -1 PWM signals killed it.

[Joe Ross]

My best guess would be a small metal shaving shorting the 6v supply. It could have worked its way into a bad place over time. Look in the area of the PWM connector/servo jumpers.

[mattachew12]

I didn't see anything but a bit of dust at the very bottom, and we've switched out two sidecars already in the past couple weeks.

[mattachew12]

It is solved! It turns out that though nearly all documentation requires a 20 amp circuit breaker for the sidecar, one obscure schematic requires a 10 amp breaker. So far it is still functional, so it looks like we were giving it too much current.