pi@raspberrypi:~ $ uname -m
aarch64
The only thing ive not done
is plugged in the airspy mini yet
pi@64bit:~ $ uname -m
aarch64
SITE 98275 up and running since about 18:45 yesterday.
I definitely canāt run the CPU as hard without dropping samples.
Previously samples started being dropped at greater than 88-90%, now 67-68% is as far as I can push it.
Itās strange, a 0.1% of -e increases usage from 67% to 90%+, was not like that with 32 bit Iām sure.
Please see this:
Added the new binary to be used for uname -m giving arm64 or aarch64.
So, I found that the CPU governor was throttling down the CPU because we donāt use much of it, and this results in the bus to also run slowly which in turn results in packet drops. I hope they will fix this behavior in the official Raspberry OS 64. Not much impressed by the beta so far.
Iāve had a similar problem previously with the 32 bit decoder - I have force_turbo=1
set in /boot/config.txt, which disables cpu throttling so it remains at full speed all the time. As long as you arenāt overclocking or using the over voltage flag, it wonāt trip the warranty bit. Fixing the cpu frequency solved that issue.
I have documented that for the Pi2/3 but I was hoping for better with the Pi4.
Iāve seen this before with single-threaded (or more generally, threads < cores) CPU-bound loads; it can happen on x86 too.
Just got an Airspy Mini and am setting it up on my RPi 3B+. I am having a problem where its running fine for a few minutes but is then stopping unexpectedly. The Airspy Mini firmware is the latest version. Iām running it in 12MSPS with the bias-tee on. The bias tee is running the RTL-SDR ADS-B LNA. I am powering the RPI using the POE hat. Here is what the log is saying:
Jun 26 10:49:38 piaware airspy_adsb[279]: Decoding stopped
Jun 26 10:49:38 piaware airspy_adsb[279]: Client disconnected 127.0.0.1:50166 (beast)
Jun 26 10:49:39 piaware airspy_adsb[279]: Acquired Airspy device with serial 260868C82A329B9B
Jun 26 10:49:39 piaware airspy_adsb[279]: airspy_start_rx() failed: AIRSPY_ERROR_LIBUSB (-1000)
Jun 26 10:49:40 piaware airspy_adsb[279]: airspy_open() failed: AIRSPY_ERROR_NOT_FOUND (-5)
It just repeats the not found error until I shutdown the RPi, remove power and start it again. It will then run a few minutes again and then throw the same errors.
Sounds like power issues.
The early POE hats didnāt provide proper USB power i believe.
Maybe you have one of those?
Not sure, could maybe also be the Airspy Mini getting hot?
Maybe get an external bias-tee and feed it 5V from the RPi?
That way itās not going through the Airspy and might be less of an issue.
Check if you have undervoltage warnings:
sudo dmesg --ctime | grep voltage
Make sure no other USB devices are connected. The shared USB bus (SD card, ethernet/wifi and USB) cannot handle another device being connected whilst the airspy is in use.
I ran the airspy with POE adapter and bias-t enabled on an RPI 3B+ without any issues. Make sure that you have a quality POE power supply. The 4B is much better. You will need to keep the RPI3 CPU cool as it often overloads the CPU.
I powered the LNA separately with an injector. This will leave a bit more juice for the Airspy from the limited USB port.
I definitely have one of the newer POE hats that shouldnāt suffer from that problem.
The POE injector Iām using is a TP-Link TL-POE150S V4 which should be able to provide enough power.
I have an external bias-tee coming that I can power from somewhere else to see if that fixes it.
Strangely enough Iām also seeing this when I restart the processes according to your instructions. Itāll start throwing that error again. So any time I change options I have to restart the Pi and then unplug and replug the Airspy a few times before its found.
Did they change how power is distributed on the Pi4B? Am I better go just get one of those if Iām going to use the Airspy Mini?
So actually iāve often seen the same issue with an Airspy R2 used with a laptop.
It usually only occurs after reboot and might be a separate issue.
@prog
Is there anything to diagnose this issue further? (the below error happening after the RPi decoding for a couple of minutes)
Jun 26 10:49:38 piaware airspy_adsb[279]: Decoding stopped
Jun 26 10:49:38 piaware airspy_adsb[279]: Client disconnected 127.0.0.1:50166 (beast)
Jun 26 10:49:39 piaware airspy_adsb[279]: Acquired Airspy device with serial 260868C82A329B9B
Jun 26 10:49:39 piaware airspy_adsb[279]: airspy_start_rx() failed: AIRSPY_ERROR_LIBUSB (-1000)
Jun 26 10:49:40 piaware airspy_adsb[279]: airspy_open() failed: AIRSPY_ERROR_NOT_FOUND (-5)
This could be a power issue. Check the PSU and the cable.
What LNA were you using?
I plugged my RPi directly into a power cable and that may have fixed the issue where it only ran for a few minutes. But I am still having the problem if I restart the Airspy_adsb service. I have to go and unplug and replug the dongle once or twice for it to show up. Hereās the error its giving again.
Jun 26 15:30:14 piaware systemd[1]: Stopped Airspy ADS-B receiver.
Jun 26 15:30:14 piaware systemd[1]: Started Airspy ADS-B receiver.
Jun 26 15:30:14 piaware airspy_adsb[24484]: airspy_adsb v1.85
Jun 26 15:30:14 piaware airspy_adsb[24484]: Listening for beast clients on port 29999
Jun 26 15:30:14 piaware airspy_adsb[24484]: Acquired Airspy device with serial 260868C82A329B9B
Jun 26 15:30:14 piaware airspy_adsb[24484]: airspy_start_rx() failed: AIRSPY_ERROR_LIBUSB (-1000)
Jun 26 15:30:15 piaware airspy_adsb[24484]: airspy_open() failed: AIRSPY_ERROR_NOT_FOUND (-5)
and then that last error repeats every second until the Mini is replugged once or twice.
Strike that. Plugged in to a normal power supply that I have no reason to doubt is still causing it to unexpectedly shut down and gives me this error:
Jun 26 15:35:08 piaware airspy_adsb[24484]: Decoding stopped