search

WHOIGit / PhytO-ARM

PhytO-ARM: Phytoplankton Observing for Automated Real-time Management
https://hablab.whoi.edu/phyto-arm/
GNU General Public License v3.0
0 stars 0 forks source link

feat: Implements JVL command for zeroing position encoder, calls on winch startup #22

Closed figuernd closed 3 months ago

rgov commented 1 year ago

There's a typo in the service definition, "registry".

If this is the correct way to do it, then it seems like a good change.

figuernd commented 1 year ago

There's a typo in the service definition, "registry".

If this is the correct way to do it, then it seems like a good change.

Good catch, fixed

figuernd commented 1 year ago

@rgov can you think of any side effects of using this? i.e. position drift from being reset too many times? (Seems unlikely since position is calibrated by CTD depth). Thinking about adding a config option to disable this command as a backup.

rgov commented 1 year ago

I have no experience outside this project. I don't think we need to add an option; either it works and we continue using it or it doesn't and we deploy a hot fix to revert back.

It would give me more confidence to know that this is indeed exactly how MacTalk resets the encoder position. We can do that by using MacTalk with WireShark, for example.

figuernd commented 1 year ago

I have no experience outside this project. I don't think we need to add an option; either it works and we continue using it or it doesn't and we deploy a hot fix to revert back.

It would give me more confidence to know that this is indeed exactly how MacTalk resets the encoder position. We can do that by using MacTalk with WireShark, for example.

Haven't merged yet because I still hope to do this; haven't had the proper setup yet.

rgov commented 3 months ago

I am not 100% sure about zeroing at startup since in theory the zero point would best be the median depth? If the total encoder range is more than enough to cover the depth range that's probably fine though.

Merge when you're happy with it!

figuernd commented 3 months ago

I'm going to take a closer look at MacTalk values in the lab next week - but we've been using this in the field long enough that I'm going to merge for now so that I can finally stop rebasing on top of it for every deployment. Will open a new issue if I find anything wrong with the procedure in lab.