team5818 / shermancrab-2019 Goto Github PK

Arm moves at a constant slow speed when robot is turned on, regardless of gravity. Input from joysticks modifies this speed when pressed, but returns to the slow speed after joysticks are centered. Still uses competition code.

It does this occasionally, despite joysticks not being touched. The current deadband is too large to be a deadband problem.

SparkMax motors can't follow Talons. They also don't have motion magic, so I suspect some custom PID looping will be needed.

We should add a gyroscope and accompying code to the kitbot for rotation testing. This means a new field for the DriveTrain, and a new command like Rotate(double degrees) for testing.

These motors will flip into their idle mode between manual orders. A workaround is to use coast and set to brake when disabled.

Slow down the drive train before shifting and resume

The arm can hit the hatch mechanism when it is extended, so we need to prevent that somehow. Perhaps with a flag of some sort?

Comprehensive log of any time the robot does anything w/ timestamps

It would be nice to make a joystick map. Perhaps we can make an auto-generated one via Shuffleboard, but if not it'd be nice to just make it in Google Drawings or similar.

Assisting collection and deployment with alignment will help a lot.

Not sure what's up, but in teleop, pushing forward on the joystick goes in a direction, which we consider "forwards". Motion Magic likes to go the other way when given a positive distance.

Avoids gears over-torquing and stripping

Make something that will run diagnostics on the robot, potentially out to a UI

Hardware wants to swap the default state of the clamp to be "closed", and true will open it. We need to flip all the software clamp solenoids first.

Hatch push is currently using very short timeouts. We want something like the following:

• Push top/bottom pistons out
• Release the panel
• Wait until button release (this requires another command, or us to use end)
• Retract top/bottom pistons
• Keep the clamp in "closed" state -- this is connected to the next issue as well

The encoders read negative values when going forward. Potentially the drive train inversion needs to be swapped. The code does still work with some absolute value trickery.

Something that would be nice is to calculate the hood's angle relative to the ground / robot, rather than to the arm. This would mean injecting the ArmController into the HoodController, and adding a new method returning degrees (I think this will be easiest, since we can calculate it for both). The hood degree calculations can be done similar to the arm, and we can just measure the positions later. This will help with calculating a gravity offset for the hood (something I thought of, after watching the hood fall while they were driving), and for maintaining a constant position of the hood (we can do a PID to the new degree value from the method).