Investigate "brownout" recovery issues

[whymarrh]

Currently the video feeds seem to recover when the vehicle's 12V bus browns out, but the control software does not. A few things need to be looked into:

• What causes the Pis to not recover?
• What allows the video feeds to continue working?
• Where are the logs?
  • service eer status/systemctl status eer.service show partial output but journalctl -u eer.service shows literally nothing

[cal-pratt]

My theory is that the switch is browning out and that the event publisher is less robust than netcat

[whymarrh]

What do you mean by "the event publisher is less robust than netcat"?

[cal-pratt]

The video feed recovers after a disconnect but the control software does not. You can see live video after a brown out, but joystick communication is lost

[whymarrh]

Ok, but I don't know what you are getting at with "the event publisher is less robust than netcat"....

[whymarrh]

Our event publisher and our usage of nc use two different protocols—protocols that aren't comparable in their "robustness"

[cal-pratt]

Robust meaning that it can handle more adverse conditions such as gaps in network activity

[cal-pratt]

raspivid->netcat->mpv connection can handle the switch power cycling, vs the eventpublisher->eventpublisher connection which cannot

[whymarrh]

• Where are the logs?
  • service eer status/systemctl status eer.service show partial output but journalctl -u eer.service shows literally nothing

Running journalctl -u eer.service as root seems to show the correct output

[cal-pratt]

@FifoIronton Have you started looking into new thethers? Hopefully with a second conductor this wont be as critical an issue.

[FifoIronton]

Sorry about huge delay, still getting used to keeping up with Github, I need to adjust my email settings as not to get bombarded by miscellaneous messages.
Clar has contacted L-E about a new tether, but it was as part of a large order, so we have a while to wait to see if they'll approve that order.

To do Saturday:

• Characterize our voltage drop per over the tether for different loading (can't just use resistance because of heating), see if backup plan of chopping our tether to a shorter length is feasible (MATE says they'll never go back to the NBL so we'll never go that deep again)

• Give Michaela rundown of what we're ordering, get her to contact L-E, as she's contacted them in high school

[cal-pratt]

see if backup plan of chopping our tether to a shorter length is feasible

This seems like a big hack. Is there any way to create a current limiter on the just thrusters? I'd hate to see the tether get shorter, but it's better than brownouts I guess.

As a side note; can you update the wiki to include the tether breakdown?

[FifoIronton]

I made a breakdown of the current tether options and put it on the wiki. As we don't have a finalised tether, it's more of a discussion than anything.

[cal-pratt]

With the new design this shouldn't be a problem moving forward for this year. Going to close this ticket and re-open it at a later date if necessary