Alligning swerve modules (and other problems)

[ekapalka]

I've never seen this discussed in a thread, so I assume it must be somewhat easy to figure out. One of the fairly common swerve issues I've read about is having the modules out of sync. I figured that some teams would use limit switches or hall effect sensors to re-calibrate the rotary position sensors of the modules during a match (similar to what most 3D printers do to identify their boundaries), but as of yet I haven't heard of a team resorting to doing that. I was wondering what teams have done to mitigate this issue as well as other crippling issues during matches.

[Tyler2517]

Swerve realignment is interesting my team had some problems early on in the season because we were new to the drive train.
We used thermal press fits to connect the coaxial shaft(outside shaft) to the top of the pivot yoke. We had one that was welded and the other 3 were using the press fits. The ones using the press fit slipped a lot during our first district.
In our second district we only had a slip ever 3 or 4 matches. In our last district we had no slips but we did loss a module after being catapulted by another robot.

The slipping was a major problem for us and in the future we plan on using a limit switch to automatically calibrate in a match.

When only one module was lost and it would just spin in place powered or not or just sit still our other 3 would just over power any movement it would say we would see a noticeable drop in top speed and acceleration.

Once the mechanical slip stopped we did not have any problems.

[cjl2625]

This was never a problem for us throughout the season. Never noticed any slipping.
Here's my attempt at a mechanical explanation of how we did it: The wheel is attached a triangle holding box thing which is attached to a stick which is attached to a gear which is attached to another gear or something which is attached to a stick which is attached to an absolute encoder.
Sorry, I'm a programmer.

The most crippling swerve issue for us was gyro alignment, as we used a field centric system. Sometimes the gyro would just go nuts, creating a phenomenon we like to call the "drunken apple"'

After putting the gyros in a protective shock-absorbing case, we didn't have any problems with complete gyro craziness, although gyro drift was still present and annoying. We have to zero the gyro many times during a match.
This is probably the biggest problem we've had with swerve, and I'm not quite sure how to fix it.

[asid61]

Hall effect sensors would be good if you are okay with wires. Otherwise, you can use a rubber tube stretched between the modules to straighten them out relative to each other before "zeroing" their encoders.The holes for the rubber tube need to be above the axle on the wheels.
I think 1640 had something on it.

[Gdeaver]

We use continuous rotation 360 degree absolute encoders on our coaxial modules. All modules are identical. We calibrate our modules on or off the robot. We use the convention of looking at the module from the front, the wheel sprocket is to the right. We get the wheel pointing straight forward by eye, twisting it back and forth making sure it is centered. (more lash than I like). We connect Ground and 5 volts to the sensor. Signal and ground to a volt ohm meter. We loosen up the set screw on the beam coupler and twist till we get the 0.25 - 4.75 volt crossover position. Tighten the set screw. One of our mentors made a break out board to make this easier. This also can be done by the crio and dash board. We did have problems in the past with plastic beam couplers from automation direct loosening and loosing calibration. This year we used the First choice voucher and use the Al beam couplers. No slipping. We have had arguments over calibrating to the cross over point or 2.5 volts. I don't think it matters. When things get out of calibration, the code is bad or something else is wrong, our team calls it the happy dance. The drunken apple? I think I know were that came from. I call it a grand mal seizure.