19458 Equilibrium.exe Progress Updates

[miriamsr]

We picked up our SystemCore and MotionCore on Friday and powered it on in the hotel. After some difficulty flashing the OS (see #243 which is now resolved), we got it configured and connected to the DS. The first task we plan to do is run one of the mecanum drivetrains we use at outreach events from the SystemCore.

[miriamsr]

I wired a SystemCore and Expansion Hub to a mecanum drivetrain and followed these instructions to create a sample program. However, after following instructions to deploy the code and VSCode said "BUILD SUCCESSFUL", my DS still says "No Robot Code" and I don't know how to resolve that.

[zachwaffle4]

Maybe they changed it on a more recent version of Android Studio than when I was a freshman in high school trying to speed up my development workflow.

Perhaps it's a Gradle thing? Or Android just being weird.

[miriamsr]

Yeah, looking at the docs, it's a little bit unclear since it says "build and deploy" and then lists all the ways you can build, and then it sticks a massive image in the way, then mentions at the bottom that you can deploy by selecting "Deploy Robot Code" instead, which definitely is asking for someone to stop reading partway through.

I think I might PR moving that sentence about deploying up so that it's above the image so that you read all the way through before hitting the image. Do you think that would've avoided this/is an improvement over current docs?

Yes, I think that will help a lot. Given that I expect others with minimal FRC experience to have a similar oversight to mine, anything you can do to make it extra clear that deployment is a separate step after you follow the build instructions would be great. I think putting a note in an FTC-centric place about this and other things that trip people up because the differences are subtle.

[miriamsr]

Perhaps it's a Gradle thing? Or Android just being weird.

I just did a build, connected a control hub, and then used the run button and the run button does still trigger a full build. It doesn't actually matter the cause, what matters is that this isn't something FTC teams can do currently but will be able to do on the new system.

[ThadHouse]

The run button is going to be something different, because generally that is going to run Simulation. We definitely need to document that deploy is what you need to do to deploy code. Also note that deploy does build as well, so you don't need to do it as a 2 step process.

[sciencewhiz]

I'd appreciate review on this PR for docs that attempts to make all of this clear. wpilibsuite/wpilib-docs#3309

[miriamsr]

We successfully got a robot moving with a SystemCore and Expansion Hub. I did notice that the motors didn't stop right away or even at the same time when I released the joystick. The attached video is all forward and backward motions from the gamepad, but you can clearly see that it seems to be trying to correct itself in an undesired way. I'm wondering if something for Teleop OpModes that makes MecanumDrive stop when there's no gamepad input would help mitigate this issue, although that probably isn't doing anything for the root cause of the problem. I tried setting floatOn0 to both true and false and the problem occurred both ways. You can find our code here.

IMG_1442.mov

[miriamsr]

Applying a deadband to each joystick worked. Is there a good place to document things that FTC teams never had to think about on the old system? I would like to suggest having a single place for all alpha teams to add to so nothing gets lost when writing documentation.

[ThadHouse]

We should probably create a new discussion for that. Evenutally we will have docs.

Note this specific thing is being fixed directly. We're going to add a deadband by default. We've generally not needed one because our joysticks were so low resolution.

[zachwaffle4]

I've created #266 for FTC transition suggestions :)

[auscompgeek]

We're going to add a deadband by default.

Will the deadband be user adjustable? My team has varying sticks of questionable quality after a few years.

[PeterJohnson]

Yes, it will be user adjustable on a per-axis basis.

[chaosleader]

I am trying to get autonomous motion working by setting the position setpoint of the motor but it seems like the motors are not receiving any power even after I set the setpoint. I have set a P value in the position PID. The code is here.

[Gold856]

You're setting PID constants in the robot constructor, which is run at startup, which means that it likely doesn't have a DS connected, so isAutonomous will likely return false, and the PID constants won't be changed.

[chaosleader]

Thank you, after making that change it worked.

[chaosleader]

I started working on a slightly more complicated auto and when I went to test it, it said there was no code running. I went to check it for the other opmodes which had been working previously and this TeleOp had the same issue which while I had changed it slightly since I last ran it I didn't see any changes that would cause it to break. This is the other opmode that isn't working.

[Gold856]

It doesn't look like you initialized MecanumDrive.

[ThadHouse]

We’re working on getting better crash reporting into the DS. But yeah as gold said you didn’t initialize mecanum drive.

[chaosleader]

I was able to get the square auto working without much trouble. In working on it there were a few things I noticed that were annoying/unintuitive. The first is that there is no way to get the set-point from an ExpansionHubMotor object so i had to store the target in a separate variable which is fine but it might have been a little nicer had there been a method to get the set-point from the motor. The second thing is that it seems like periodic() runs for the first time before start() runs. I'm not sure if this is intentional but it definitely makes some things harder. I'm not sure if there is a way to change the order in which the methods run but either way it was unintuitive that start is not the first thing that runs. The final thing is I couldn't find where runtime exceptions are logged which is probably on me but it would be nice if there was an accessible place where exceptions were logged.

[PeterJohnson]

It's not intentional that periodic runs before start; the fact that periodic always runs, not just during enabled periods, is a bug that is being fixed in wpilibsuite/allwpilib#8888.

[chaosleader]

I was able to get I2C sensors working pretty easily. I started with getting a Color Senor V3 working which was very simple and straightforward. I then wanted to get a GoBilda Pinpoint working. I was able to have ChatGPT port the driver to WPILib fairly easily and got the pinpoint working. I then started to implement simple PidToPoint pathing. While I don't quite have everything working yet. I did have a problem in tuning though. My tuning OpMode would crash whenever I would change the target values with Smart Dashboard. I'm not certain what I am doing wrong and I was unable to find anywhere that displayed runtime errors.

[ThadHouse]

You can get the runtime errors in the webui in the logs for the robot service. I’m not near a systemcore but hopefully someone can post it.

Eventually we will add the crash display into the WPILib, then crashes will show up on the ds. Hopefully the next release.

[gearsincorg]

If you are lucky, the error is also displayed on the Driver Station if you select the "LOG" mode of the small viewscreen. (Button on the far right of the DS pane).

[chaosleader]

We recently received the A301s and I had little bit of trouble updating the firmware. I would click the update button after uploading the firmware file and nothing would happen. This was resolved after a power cycle.

[chaosleader]

We were able to build an A301 drivetrain. However we had a pretty major issue with one of the motor ports. I have to do more throughout testing to confirm exactly what the issue is but I are confident that the motor itself and the wire is fine. The issue we were having was that when we set the throttle of the motor the motor would not spin. However we were still getting position data from the motor. Additionally, though this may be due to a code issue or the install not working it seemed like an adjacent port was getting power when I sent power to the seemingly broken port. I don't have time to do a more through analysis but I will probably be able to next weekend.

[gearsincorg]

There is a bunch of magic going on inside the A301 and I'm not sure how trottle mode is implemented. But I have seen that velocity mode does reach the max stated velocity of the gearbox, so if nothing else, you know exactly how fast it "should" be spinning. If it doesn't reach it, there must be some other factor (torque or current limit) kicking in. I wouldn't expect either of these on a lightweight Mecanum bot.

[gearsincorg]

Your breaker voltage observation is interesting. Typically it's never a good practice to use a device that is not rated for the intended voltage. I suspect that 18V isn't really beyond the breaker's actual max voltage, but that's just a guess.

In terms of operation, only the rated current is a factor. Since these are typically based on accunumlated thermal energy, and they are designed to not trip prematurely, you really have to exceed their max current for a while to trip them. Here is the chart for a simmilar unit used by FRC. You can see that you need to overload the current by 300% (80A total load) for at least 2 seconds to get one to trip (and it may take up to 7 seconds). With "only" a continuous overload of 100%, it may never trip :(

[chaosleader]

I did look at the diagnostic screen but I didn't see anything wrong on any of the ports. There was a fault that was labeled as sticky in RHC2 on all of the A301s we used but I wasn't really sure what it meant and it didn't seem to be effecting anything.

[revspin]

@chaosleader
Thanks for the update on the drivetrain build. Can I ask a few questions to help diagnose this?

Troubleshooting Questions
HWC2 Faults: Does the affected port or motor show any active faults or warnings in HWC2?
LED Status: What is the specific LED color and blink pattern on that port?
Command Response: When you send a forward/reverse command, does the LED respond accordingly (e.g., changing green/red), or does it stay the same?

Common Startup Issues
For context, if the motor isn't spinning, it is often due to one of these common states:
No command / loss of heartbeat
No gearbox detected: Indicated by a fast orange blink.
Over-temperature: Indicated by an orange/green blink.

For most drive trains I would recommend using the velocity control unless its not achieving your requirements.

[chaosleader]

I'm not going to be in my team's workspace again until the weekend but I'll definitely let you know the answer to those questions once I'm back in my workspace.

[chaosleader]

Me and the rest of my team were wondering if it would be possible to get the data sheet for the A301s. If we can't get the full data sheet would it at least be possible to get a motor curve for them? Thanks