search

osqzss / gps-sdr-sim

Software-Defined GPS Signal Simulator
MIT License
2.61k stars 760 forks source link

Possible to create consistent SNR/CN0 simulation? #382

Closed antmerlino closed 10 months ago

antmerlino commented 10 months ago

Hello,

I am trying to use this software to test a GPS receiver in manufacturing. To do this, I am trying to generate satellite signals with consistent and stable SNR values. I have tried fixing the gain to 100 as mentioned in other posts, as well as disabling ionospheric delays. When the simulation first starts, I get a very flat set of SNR values across all the satellites, but then as the simulation progresses, I start getting increasing disparities between the satellites.

I am wondering if you could point me in the right direction as to what is causing this behavior. I'd be happy to contribute any improvements back and also willing to make a donation for help supporting this.

Thanks, Anthony

osqzss commented 10 months ago

I added the "-p" option to disable path loss and maintain a constant power level. It simply fixes the gain at 128 for all visible satellites. I tested the generated I/Q samples with a u-blox receiver, and there were no disparities between the satellites. If the signal level is too strong, it can lead to cross-correlations among the satellites. Please consider inserting appropriate attenuation in the RF line.

antmerlino commented 10 months ago

Hi @osqzss.

If the signal level is too strong, it can lead to cross-correlations among the satellites.

This was ultimately my issue. I had 3x 20dB attenuators and a DC Block coming from the HackRF One, but that still wasn't enough. Adding a 4th attenuator fixed the cross-correlation issue and gives me consistent SNRs.

Thanks for the help!