Was CANivore support for sysid added to 2023? I opened up sysid v2023.1.1 and don’t see it.
I know there was an issue opened for it last year. But if it was implemented I am not sure how to denote that a motor is on the CANivore or give the CANivore bus name.
IIRC, there’s a pull request for it that also adds phoenix pro support as well.
So… Not here but coming soon?
Found it.
That PR has has most of the features, I just haven’t gotten to adding CANcoder Pro and haven’t yet tested all the configurations with CANivore/Phoenix Pro.
I do know there’s a bug in the gains calculation that appears to produce gains that are off by a factor of 12, which I’m working on solving.
Anyone able to test it and provide feedback on what does/doesn’t work would be helpful, as this is something I’m doing outside of work and haven’t had a lot of time to verify everything works as expected.
Is there a plan for this to make one of the point releases this season?
Tytan asked for testing and I haven’t seen anyone post testing results either in this thread or the PR. Are you planning on testing it? I can’t imagine it being merged without testing.
We aren’t there ready for testing yet. I might be able to get some, but you may want to make a top level topic asking for canivore testing on sysid.
I’d really like this to get to a state where it can be merged in an official release, but I’m not comfortable doing this without more testing either from the community or from myself once I find time/do testing with my team.
At a minimum I’m aiming to maintain the fork against the main repo, so users can at least checkout my fork and use CANivore/Phoenix Pro while getting the latest features.
I may put up a top level topic once I’m a little further along. Like I said, I don’t have CANcoder Pro implemented and have the gain calculation off by a factor of 12 bug, which I haven’t had a lot of time to dig into. Testing my fork either on a real robot or reviewing it is really appreciated, since I’m only one guy and this is a relatively large change that touches a lot of components, many of which I’m not intimately familiar with.
So just to understand, if we were to use Phoenix Pro on a canivore, and then use SysId, the cancoder data can’t be trusted yet in SysId.
Is there a way to down grade just the cancoders in a Phoenix pro canivore loop, so that the numbers can be trusted in SYSID?
I actually just got CANcoder Pro implemented, so you’re able to use it in SysId now (untested however).
Still looking at why the gains are incorrect, but the data collection is correct as far as I can tell.
If CANcoder Pro is causing issues it’s perfectly acceptable to put Phoenix v5 firmware on them and collect data with it. You can mix and match v5 firmware and Pro firmware on a CANivore Pro bus, both for SysId and normal robot program operation.
Should probably make a standalone thread asking specifically for people to test it out. Just a suggestion.
Off by a factor of 12 sounds like an issue with voltage conversion somewhere (you didn’t say which way) I know pro allows voltage control without converting to percent output, if you are setting previous percent calculations as voltage that would be off by a factor of 12. Haven’t looked at the code so you can ignore my blind guessing if you want.
I haven’t tried it out yet, but it looks like support was added in release v2023.3.2.
I saw that, I have it installed as of 5 am this morning but haven’t had a chance to try it out yet, probably Thursday
Looks like it broke the ability to load old configs. I guess I’ll recreate my config. Not the end of the world.
This topic was automatically closed 365 days after the last reply. New replies are no longer allowed.
