Re: #DRAWS oddity, GPS and ALSA headaches #draws


Basil Gunn
 

Hi Jack,
Taking your responses out of order.

it ain't worth fretting about. Solution found.
Yep, totally agree. Get another RPi and continue having some fun. A
friend of mine had exactly the same symptom on a Pi 3B. I have 6 RPi's
on my bench running 24/7 and haven't seen the problem with any yet. At
least two of the RPi's are RMS Gateways that have been running for a
couple of years.

I am going to add a cron job to all my RPis that check's being throttle
& over temperature as in the case you had where vcgencmd get_throttled
returned 0xd0005.

vcgencmd get_throttled
https://www.raspberrypi.org/documentation/raspbian/applications/vcgencmd.md

// Is it OK to run a Pi4 off the DRAWS buck regulator? - I vaguely remember
discussion concluding it was.
Yep. I have 2 RPi 4's (2GB, 4GB) running powered by the DRAWS 12 V buck
regulator with no problem. Actually all my RPi's run from 12 V. I never
have problems with under voltage & it is a lot easier to manage a bunch
of power pole connectors than a bunch of wall warts plugged into a power
strip.

// it is on all 40 pins correctly - that is one of my personal QC checks -
burnt up my first pi making that error
You are not an RPi expert unless you burn at least one up.
I killed two due to mDin6 control cable laying on antenna coax. Didn't
learn my lesson the first time.

Glad to hear you are back up & running.
/Basil n7nix

Jack Spitznagel <@flyingfrawg> writes:

Hi Basil,

Good news/bad news story:
DRAWS hat appears to be OK, the Pi3B+ is a sick puppy. Silver lining:
discovered just how fast everything DRAWS seems to respond with a Pi4. That
may be the replacement board I pick. The 1GB versions have gotten fairly
cheap.

See inline after your suggestions below:

* Keep a continuous eye on the temperature In a console window cd cd
n7nix/debug watch ./pitemp.sh
// never bothered with pitemp.sh - just used vcgencmd measure_temp to start
with & GPU temp
// kept rising, so removed the DRAWS board, let the Pi cool, booted vanilla
Rasbian Buster
// and the GPU temp headed toward 65C very quickly. Both GPU and CPU sinks
were very hot to touch so
// shut it down so not to waste time to create pitemp.sh on the sick Pi in
vanilla buster.
// AT this point I decided the Pi3B+ board was ailing and did not push it.
* Using pitemp.sh note the temperature at boot & after 1/2 hour
// Skipped this as noted above
* Removed the DRAWS hat, boot, note the temperature & after a 1/2 hour
// Skipped, see above
* When you re-install DRAWS hat check for being off by one on the 40pin
connector.
// it is on all 40 pins correctly - that is one of my personal QC checks -
burnt up my first pi making that error

* Power down RPi with DRAWS hat, let RPi cool down At cool power up is the
DRAWS hat working?
* just run alsa-show.sh and look for something reasonable.
// I did this and the symptoms were the same, nothing in alsa-show.sh
output and nothing
// from gpsmon other than it "could not find the device at localhost port
2XXX" whatever.

* Do a close visual inspection of the DRAWS hat & RPi.
* Anything look charred?
// Nope! Just two very hot to touch processors - both have heat sinks on
them too! That may have
// been what prevented a total meltdown.
* Feel the USB cable powering your RPi does it feel warm?
// Neither power cable configuration got warm
More things to try as you suggested.

The Pi did not seem to be acting sick by itself.
Could it be the power source or should I not trust the PI at all?
// wired the buck regulator feed to the display power line and saw the same
symptoms with the Pi3B+
// THEN I subbed in a Pi4 (4GB) that I have been playing with and it fired
right up. DRAWS came up and
// behaved normally for alsa-show and gpsmon including no lightning bolt
while running off the buck
// regulator on the DRAWS board.
QUESTION:
// Is it OK to run a Pi4 off the DRAWS buck regulator? - I vaguely remember
discussion concluding it was.

It is suspicious to me that the RPi gives the low voltage symptom (lightning
bolt). Is it the power supply or is the DRAWS card loading the RPi voltage
source?

The following is not a bad idea but I would confirm that there is not
something seriously fried on the DRAWS hat first.

If you think I can trust the Pi, I could always take a different power
supply tactic and use the buck supply line in the DRAWs card for the
Pi and feed the display from a separate line.
SO: The Pi4 is good with the DRAWS but the Pi3 is the second I have
consigned to the junk bin. Something probably munged up the GPIO lines
needed by the sound and GPS subsystems - have no idea what and @ <$30 for a
replacement, it ain't worth fretting about. Solution found.

Join udrc@nw-digital-radio.groups.io to automatically receive all group messages.