To start with, I underestimated the actual draw, and went in with a single 10A supply that needed brightness down to about 50% to push white to any appreciable number of pixels (of the 896 WS2812Bs). The event can get a bit, well, soaked, so I enclosed the PB in a sealed container about 5x3x4in. Around three days in the display started freezing a few hours after the PB was power cycled, and while its AP was still up and accepting connections, attempting to get into the UI spun with some interesting decorations:
This continued after pulling the top off its container and giving it a hard reboot. I’m curious how folks doing static-but-movable installations in harsh environments protect their PBs, and whether anyone else has seen similar issues with heat, sustained brownout, or the combination of the two. All my experience has been with tiny indoor installs and wearables, so this is definitely new ground for me.
Hey, a couple of quick questions before diving deeper:
Given how it’s wired or installed, can you remove the LEDs from it easily? If so, can you power it off a clean 5V source with no load, like the Micro USB connector?
Can you reach http://>/recovery.html?
Are you fairly confident the sealed container was getting extremely hot? I want to think about heat damage vs power issues.
As for how to protect electronics from rain but not have them overheat, I’ve been using a lot of “IP43” style, IE “rain-proof” solutions. This means they’re open to the air somewhere but the shape of the enclosure prevents them from getting wet from rain. They’d get soaked if somehow submerged. One example that’s been going for 3 years outdoors is a Pixelblaze in a small project enclosure box with holes drilled in it on the bottom, sitting in an inverted ziplock bag that’s very lightly cinched around the cords coming out, kind of like a balloon.
I used to worry about dust, for example in the Black Rock desert, but honestly, my experience matches up with a lot of other people: If you blow electronics off with compressed air after it gets really dusty, it has a surprising lifespan and corrosion resistance (especially if you apply some conformal coating).
Thanks for the quick response! In retrospect the wizard himself was at the event and I maybe should have sought him out, but our paths never crossed while I was debugging. I ended up swapping in a new PB, but it meant dropping the sensor board.
Removing the load and seeing how it fared was my first stab at debugging, and I saw no change.
I didn’t know to try, I’ll check it out once I find it during unpacking.
I’m not confident that heat was the issue, and don’t know how hot things were getting.
I would have been happy to help! Didn’t know there were issues. The sign was AWESOME!!!
It was probably not a heat issue with PB. They are rated to work up to a toasty 85C. Might have gotten too hot if the power supply was also in there but I would suspect the PS to give out first and most have thermal protection.
It could have been dirty power causing Wi-Fi issues. Or wire/metal doing stuff with the signal.
The page thing in your screenshot can happen when there’s a lot of packet loss and the connection drops mid way through transferring the page. It can also happen if there’s flash corruption but that’s much less common, but would persist no matter what you do with power. Recovery tool can force an update and would reinstall the web app.
If you were close to the power supply limits, my guess would be a weak power issue since some PS can get dirty under load, but troubleshooting without the LEDs would have improved it.
It still might have been a dirty power issue. Sometimes caps, especially electrolytic, give out over time and quickly under stress and heat.
Did the PB swap help? Does the old one work now that it’s outside of the system?
I’ve dug through the bins that came back from SOAK and have yet to find the faulty controller, I’ll do some experimenting whenever it finally turns up
