My station stopped working


My station isn’t working anymore. I fear the worst because the web server isn’t starting up and it’s not responding to ssh anymore. I’ll have to disassemble it and see if the HDMI and keyboard will work.

I’ve had it for about a year, and I wanted to share some thoughts about it.

I used a Raspberry Pi 3 B+ for my RS4D, and I added enough SD disk space to handle a couple of years worth of data. I even increased the maximum number of days to keep to 360, and I think the filesystem only got up to about 53% usage. The most unreliable part of the system seems to have been recovering from a network failure. It would keep recording the data according to /heli/, but it would stop uploading the data. It would never recover from a network failure automatically. That would either require a toggle to off-line mode or a reboot. Even then it was unreliable on whether that would resolve the connection issue. It would be nice if the software was smart enough to reconnect to the Shake network when there is an issue, like if you unplug and replug the ethernet cable. If the web server is fine, then uploading the data should be fine too. After the last reboot, it never came back. :cry: Previously, if I just let the computer cool down for an hour by powering it off, it would start up OK.

I stored the device on the cement slab floor in the garage. Sometimes it would get hot. I think the hottest that I noticed it get to was about 160°F. Lately it was down to around 128°F. I suspect that it may have been hot for too long and on for too continuously. I stored it in one of those outdoor boxes to keep the bugs and dust out. Though I did modify the housing so that the cables would pass the cables through instead of using the connectors.

Are there any suggestions on lengthening the lifespan and improving reliability of a Raspberry Shake? I suspect that a passive heat sink that could expel the heat out of the plastic container could have helped, but I don’t know where I could get that or how that would fit. Computers don’t like to overheat.

If there aren’t hardware solutions, I guess I could review the software and make suggestions on how to reduce the CPU workload and thus generate less heat. That would just take time, since I’ll likely have to rebuild the system.

What are your thoughts?

hello blackdiamond,

sorry to hear your unit is “out of commission”. reading your post, there are at least two obvious issues:

  1. the internal environment is running too hot for your Pi to handle over the long term, this is an environmental condition resulting in a hardware issue
  2. the Shake does not automatically reconnect after a loss of connection to the internet, e.g., when the router is rebooted or the ethernet cable is temporarily disconnected, this is a software issue unrelated to the HW issue described above

taking these in reverse order:
we are aware that automatic recovery from a lost connection to the internet is not handled at the moment. this is not a condition that is difficult to recognize and handle appropriately; it will be addressed in an upcoming release.

regarding the overly hot Pi: as you know, running a computer at too-high of a temperature over a long period of time usually ends badly. while it is desirable to insulate the unit from unwanted intrusions (dust, critters, etc.), it is less desirable to do so in such a way that the possibility for heat to escape is removed; heat successfully dissipating is a necessary requirement when the IoT runs 24/7/365.

my suggestion would be to revisit your solution to insulate your unit from outside intruders, guaranteeing that heat is also allowed to escape. (maybe our all-weather enclosure would be of interest?) and, as far as i’m aware, this is the first case we’ve heard of where a unit has ceased functioning due to running at too high a temperature for too long.

regarding the CPU workload: under normal operating conditions, the system runs, on average, at or below 5% utilization of CPU, i.e., it is already rather efficient for the “real” amount of work it does.

that said, there is one thing you might be able to turn off:
if you don’t utilize the local helicorder plot, accessible through the front-end interface, or access the data locally using SWARM, you can stop the service rsh-data-consumer:

> sudo systemctl stop rsh-data-consumer
> sudo systemctl disable rsh-data-consumer  # when you also want it not started at boot-up

but i have doubts that this will provide much of a reduction in how hot the unit will get when the heat’s not allowed to escape anyway.

conversely, we have had several reports over the years of users finding their units had become either an insect’s home (i found a scorpion once in mine), or exceedingly dusty. the solution in these instances was to open it up, relocate the visitors, clean out the dust and cobwebs, and restart the unit, returning it to its normal functioning state. perhaps some periodic house-cleaning would be overall easier?

let us know how you get along or if you have any other questions, we are here to help.


1 Like

Thanks! I’m actually using the Raspberry Shake all-weather enclosure. It doesn’t have a heat pipe or other way to vent heat to outside the enclosure.

Thanks for the suggestions. I do use the Helicorder plot in order to see what’s missing from the online results.

I guess I’ll try to get a new Raspberry Pi 3 B+ with a new 32 GB SD card. Perhaps it’s time to think of a new enclosure. The all weather enclosure is unusually tall on the screws and a little too narrow for the cables.