Page 1 of 1

ADSB performance on PC vs. Raspberry Pi

Posted: Sat Apr 02, 2016 11:12 pm
by gregw916
Has anyone noticed a significant difference in performance for ADSB signals from a Windows PC vs. running on a Rpi?

I'm using a Nooelec dongle with ADSB# on the PC, and for example sake, I can see about 20 aircraft this is with the basic antenna supplied with the SDR. Now I take the exact same SDR, and antenna, and hook it to the Rpi, and using either dump1090 or modesdeco2 I will only see maybe 1/2 to 1/3 of the same performance. The difference in performance measured is by how many aircraft I can see. I have tried everything I can think of, and can't isolate why they would be different. I have 2 dongles, and have used one on the PC while simultaneously watching the performance of the other one on the Rpi. Yes, each SDR has the proper ppm correction loaded, because they aren't the same.

I'm not using any kind of signal filtering or amplifying between the antenna and the SDR. I've played around gain settings, and using auto gain control, and none of them make a difference. I'm also doing ppm correction on both PC and Rpi. I'm using the max gain setting of 49.6 on the Pi as well as on the PC with ADSB#.

Does the Pi require filtering, or an amplifier to get the same performance that the PC is getting?

Re: ADSB performance on PC vs. Raspberry Pi

Posted: Fri Apr 15, 2016 10:18 am
by osramrtlsdr
Check the source code to see if there is anything different for PC or pi compilation.
If it is the same, it may be the power issue of the USB port on pi.
It can also be the problem of low CPU power on pi. You got discontinued samples.

Re: ADSB performance on PC vs. Raspberry Pi

Posted: Sun Apr 17, 2016 1:49 pm
by gregw916
Thanks for the advice. I'm not a software developer, so I wouldn't know what to look for digging around the code.

The idea of the pi processor not being as powerful as even an older Intel based processor I would think is the most likely scenario. Even though, I've got a Pi 2, and the processor load is running around 20%. There's going to be some very subtle differences between the Osmocom and Zadig drivers, as well as ADSB# or Modesdeco2 or Dump1090.

Since I have a few USB devices running off the pi, I added a powered USB external hub. That has seemed to help too.

I also moved the pi, as well as the antenna to a different location where I'm getting better results.

Re: ADSB performance on PC vs. Raspberry Pi

Posted: Tue Apr 19, 2016 12:14 am
by rtlsdrblog
Performance on the PI should be similar, as it should have enough processing power.

Have you tested using the same program on both the PC and Pi? Why don't you test with dump1090 on Windows too?

Other things: Could the Pi be more noisy than your PC? Run a rtl_tcp server on the Pi and log on to it with SDR# and see if there is any difference in the noise floor at 1090 MHz.

Re: ADSB performance on PC vs. Raspberry Pi

Posted: Sun Apr 24, 2016 7:37 pm
by gregw916
Thanks for the advice. I'm now getting consistently better results since the antenna move a few weeks ago.

I did try looking at the noise floor on both the pi and the pc. I used the same gain value of 49.6 dB. I used SDRSharp to see the output.

Pi: -45 @ 1090MHz
PC: -35 @ 1090 MHz

So the PC is a little noisier if I'm seeing things right.

I didn't check dump1090 in windows since the results are similar enough. As usual with radios..antenna and location are big factors in performance. I'm sure if I could get one of those dedicated ADSB antennas from Flightaware, my performance would improve even more. I also seem them offering a LNA(line amplifier?). I'd spend my money on the antenna first, and see what happens.