GNSS SNR testing

RE: GNSS SNR testing

Arthur M — Mon, 04 Aug 2025 15:58:16 GMT

Hello Øyvind,

Thanks for this feedback.

Thank you for clarifying the "OpenSky" test. we where wondering how it could be done, makes more sense that the signal are simulated.

Thanks,
Best Regards,

Arthur M

RE: GNSS SNR testing

Øyvind — Mon, 04 Aug 2025 10:40:08 GMT

Hello Arthur, apologies for the late reply.

Here is the answer we got from our R&D team:

We define the GNSS sensitivity like this. Sensivity is the lowest signal level that enables the receiver to output GNSS fix in 2 minutes over 50% of the time. A GNSS simulator is used in the data sheet sensitivity test because it enables setting all the satellite signals to same power level. Yes, this is artificial situation, but that is the industry standard way of doing it because trying to derive any dB value from live signals (with varying power levels) would be nearly impossible.

There is no point in trying to compare results from AT%XRFTEST to our stated GNSS sensitivities due to completely different methodology. But as to why the results differ so wildly, the main reason is probably that the actual GPS signals are spread spectrum signals and for example GPS L1C/A provides processing gain of 30dB. XRFTEST expects a continuous wave input signal.

Let me know how this works for you.

Kind regards,
Øyvind

RE: GNSS SNR testing

ketiljo — Thu, 03 Jul 2025 15:57:27 GMT

Hi Arthur

Øyvind is away for the summer. I've tried to find the answer to your questions but many in the LTE team is away at the moment. I'll get back you when I hear something.

Regards,

Ketil

RE: GNSS SNR testing

Arthur M — Fri, 27 Jun 2025 15:24:50 GMT

Hello Øyvind,

Thank you for the response.

In the end, we performed tests with a cold fix in various conditions.
We managed to get a reproducible result, but we get sensitivity values ~ -120 dBm from our calculation.

Can we ask how the -146.5 dBm cold fix with Open Sky (in the nRF9151 datasheet) was tested in your setup?
Would you be able to share any insight on why our sensitivity would hit a wall around -120? Could this be because of the LNA reference that we use, or because of the fact we use a simulator instead of Open Sky?

For the SNR, we did include a test at production with the default gain, with the idea to have a repeatable and expected result.

Thanks,
Arthur M

RE: GNSS SNR testing

Øyvind — Tue, 24 Jun 2025 12:09:23 GMT

Hello Arthur, did this provide an answer to your questions?

Kind regards,
Øyvind

RE: GNSS SNR testing

Øyvind — Thu, 05 Jun 2025 07:21:35 GMT

Hello Arthur, my apologies for the late reply. Here is the answer I got from our team:

Same procedure should apply to 9151, yes.

<param0> is used to set the receiver gain. During normal operation the gain is automatically adjusted (automatic gain control) but such thing is not feasible in SNR testing. Theoretically increasing gain should increase RSSI but have no impact of SNR since both signal and noise get amplified. I believe this answers also the following question: “I am not sure why the command output "antenna_power" level changes depending on the "expected power level", shouldn't it be a raw value?“

Different input signal levels can certainly be used for testing. This will obviously results in lower SNR measurement. Note that <param0> should be adjusted accordingly.

It is recommended to have AFC ON to eliminate possible test equipment frequency error as well as any uncalibrated TCXO offset in nRF9151. During normal GNSS operation TCXO offset will anyway be resolved very accurately so there is no reason not the compensate it away during RF testing.

Kind regards,
Øyvind

RE: GNSS SNR testing

Øyvind — Mon, 26 May 2025 13:21:17 GMT

Hello Arthur,

I have forwarded your inquiry internally to get an answer to your question. I hope to have an answer by end of tomorrow.

Kind regards,
Øyvind