Is my feeder in trouble?

Discussion in 'Feeding' started by Dutch, Aug 9, 2020.

Tags:
  1. Dutch

    Dutch Member

    In October 2019 I set up my feeder and it worked fine. I irregularly check it and it kept working fine. When I checked last week, I saw it was down. I soft rebooted it (shutdown -r) and it came up again. But it went down again pretty fast. So I hard rebooted it (shutdown, then unplug, wait a minute, plugin in again) and now it seems to work more or less, but it's health is varying wildly between 95% but sometimes going very low:

    upload_2020-8-9_12-35-53.png

    upload_2020-8-9_12-39-6.png

    I never changed anything in the setup. I there something I can do to fix it? I don't want to be banned
     
  2. wiedehopf

    wiedehopf Administrator Staff Member

    Likely the power supply is dying / insufficient.

    Check if there is any output for this command:
    sudo dmesg --ctime | grep voltage

    No output is good, any output is bad ;)

    Alternatively you've not entered a precise location (doesn't look like that though, rather intermittent issues)
    If you installed using the feed scripts, it is probably easiest to just run the script setup again and this time put in the precise location:

    curl -o /tmp/feed.sh https://raw.githubusercontent.com/adsbxchange/adsb-exchange/master/install.sh
    sudo bash /tmp/feed.sh
     
  3. Dutch

    Dutch Member

    About half a bazillion times this:
    Under-voltage detected! (0x00050005)

    I think I will have a look at the power supply indeed. Thanks
     
  4. wiedehopf

    wiedehopf Administrator Staff Member

    I'd recommend depending on RPi4 or RPi3

    Raspberry Pi - official power supply for Raspberry Pi 4
    Raspberry Pi - official power supply for Raspberry Pi 3

     
  5. Dutch

    Dutch Member

    Thank you very much. You seemed to have nailed my problem. I have switched now to the most powerful power supply I have and this looks to be pretty steady now. Health fluctuates between 90 and 94% and the timeout penalty stays pretty solid at 0
    upload_2020-8-9_16-1-40.png

    Also, sudo dmesg --ctime | grep voltage now does not yield any results anymore.

    The annoying part is that I got 3 RP2's and the power supply I used was actually part of the kit the RP2 I used came in. But it was a USA power supply on a convertor plug. Maybe that wears down after a long time of uninterrupted use. \

    Thanks for the quick diagnosis. I will continue to monitor my feeder to see if it remains steady
     
  6. Dutch

    Dutch Member

    I am acutally using a RP2 ;). But you are right, I have had problems with this before. That is why I ordered the powerful power supply I use now. But the one I was using, as I said, came with the kit so I - wrongfully - supposed it was okay.
     
  7. wiedehopf

    wiedehopf Administrator Staff Member

    The one for the RPi3 will work fine for the earlier RPi models.

    RPi4 went to another connector, that's the difference.
     
  8. tedder

    tedder New Member

    fwiw, instead of dmesg and grep, you can use `vcgencmd get_throttled`. Both dmesg and vcgen require the bit math:

    "0x0" is good, anything else is some form of bad. Roughly, if it ends in 5 (eg 0x00005 or 0x5), you are currently voltage throttled. If there's a 5 near the left side (eg 0x50000) you have been throttled in the past. It's common to see 0x50005, which is both of those. And as of Rpi4 it's common to see an '8' in either of those places, which means thermal throttling.

    As wiedehopf said the "official" or canakit supplies are recommended; they voltage-compensate to the end of the cable, while a "phone charger" type USB supply is only to the outlet, and so you tend to drop voltage in the cable even if it is supposedly rated high enough.

    Here are the docs if you want the full set of possible bits.
     
    Bob Randolph and Mark Richardson like this.