servo.setAngle(angle) code crash

[cpapplefamily]

I apologize I don't have the exact log output. This evening we took the System Core off of a test chassis we use for a lot of... test i know, right, and put it on our competition robot to get it ready for this weekend and the gantlet of fall off-seasons we have the next 6 weeks. This was my first time hands on with the hardware up until now, I had our younger software mentor leading the tasks. I have a few comments on the new work flows and hang ups but I'll wait till after the first event to compile all my thoughts.

We have everything ported over except one final feature of the robot. We are using a PWM 270deg servo. Current 2027 WPILib documentation l mentions no new methods to use the SERVO When the code calls the set Servo Angle the code crashes with a HAL warning.

One of our other mentors, though he seen something about one of the Alpha testers working through a simmial issue but we are not finding any records of this.

Getting this working will be vital for us to run this system this weekend.

[ThadHouse]

I can work up a workaround, but the bigger issue is how are you powering the servo? The SystemCore outputs do not output enough voltage for servos directly. So you either need something like a servo power module or a rev servo hub.

Eventually the servo class will likely be completely removed due to this limitation. Servos directly off of systemcore just won’t work.

[cpapplefamily]

"So you either need something like a servo power module or a rev servo hub."

Well that's going to be a problem

Lets see what we can get for a work around. Ill look for a servo Power Module.

[ThadHouse]

Yeah. Servos directly off of systemcore are not supported. There’s only 3.3V on the output, and nowhere near enough current to run a servo even if you tried.

[cpapplefamily]

Either way, I suspect this is something that should be highlighted in the 2027 WPILib Docs for Servos.

[ThadHouse]

Like I said, the more likely option is we just remove the servo class from 2027 wpilib. We just haven’t gotten around to doing that yet, and there’s a lot of docs gaps in the alpha.

This is the reason why none of the systemcore specs have mentioned anything about servos.

[cpapplefamily]

Do you think I could hijack the 5vdc from the CTRE Candle and tie the 0v candle output and 0v PWM pin together to bond the reference? Will the presumable 3.3v PWM output have enough pixies to drive the servo input?

[cpapplefamily]

Will the servo class still be useful if a team chooses to add a power hub? Or will the plan be to direct teams to use a CAN Servo Hub and let the 3rd parth libraries handle it?

It will take a few years until the community forgets that you used to be able to plug a servo into the controller. The kids will forget but us midlife mentors will never forget.

[ThadHouse]

That is part of the reason to remove the class. To make it clearer that direct won’t work. The power module has some weird problems, the can servo hub is a much better product.

[cpapplefamily]

Will some try to use the edu.wpi.first.wpilibj.PWM? I was just starting down this path as a "workaround" or are these not ready for system core yet too?

Is there documentation I have not found to use the "Smart IO" ports in PWM out? if I wanted to pain myself into using the power hub?

[ThadHouse]

Using the raw PWM class would be the workaround. That class does work, as raw PWM output does work (Motor controllers do work).

But at least manually using PWM requires thought, and we can add a doc there that trying to use it with a servo directly won't work. Additionally, the rules won't allow hooking servos directly up to systemcore. And it could be possible that the rules don't allow the servo power module either, and servos have to be connected with a servo hub.

[cpapplefamily]

@ThadHouse do you have any ideas how I can hard-code a workaround for this weekend? We have a CAN PWM controller on Overnight shipping so maybe we will be ok.
I have the PWM Power Module currently working. I changed the SERVO to a PWMTalonSRX to get a sign of life. I can control the Servo but when the robot disables it returns to neutral and this can NOT happen on our mechanizam.

[ThadHouse]

SystemCore currently always returns the pwm to center when disabled. There is no workaround for that. It’s what’s currently in the io chip firmware.

[cpapplefamily]

Ok I'll stop trying. Will the CAN device react correctly? Or remain at the last postion.

IMO, that a bad practice. In industry (My world) when an estop is pressed, there should be no movement.

I can confirm this action now

Test 1

1. enable robot
2. command servo to move off center
3. unplug servo *signal pin
4. disable robot
5. Plug servo back in !!! SERVO MOVES to Center !!! PWM port appears to be continuously outputting a signal when the robot is disabled

Test 2

1. enable robot
2. command servo to move off center
3. unplug servo
4. Power Cycle Robot
5. Plug into a different unconfigured Smart IO Port
6. No Movemnet

Test 3

1. enable robot
2. command servo to move off-center
3. unplug servo
4. Power Cycle Robot
5. Plug in servo
6. Power on Robot !!! SERVO MOVES to Center !!! PWM port appears to be continuously outputting a signal when the robot is disabled

This means ANY PWM controled output device running on Systemcore will move if its not in its nutral position when the system is power on.

Seems like there is an issue here somewhere that needs to be look into.

[ThadHouse]

It’s a known bug. But it doesn’t really affect motor controllers, since to them receiving 0 and receiving no signal are basically the same thing. So it only affects servos, which due to not being supported are not really tested.