Spurious "Tracking" messages appear even if no antenna and no signal present and gain set to 0

[RobertSwirsky]

I wasn't getting NAV messages but I was getting Tracking messages. I assumed either my signal wasn't strong enough or, maybe, my B200's internal oscillator wasn't stable enough.

After trying everything, (antenna position, external clock on B200 USRP, etc.) I set the gain to zero in the config file and disconnected the GPS antenna. I still get the "Tracking" messages!

The TCP/IP server of RTCM messages is up and running. Accepting connections ...
Current receiver time: 1 s
Current receiver time: 2 s
Current receiver time: 3 s
Tracking of GPS L1 C/A signal started on channel 2 for satellite GPS PRN 03 (Block IIF)
Current receiver time: 4 s
Current receiver time: 5 s
Current receiver time: 6 s
Current receiver time: 7 s
Current receiver time: 8 s
Loss of lock in channel 2!
Tracking of GPS L1 C/A signal started on channel 7 for satellite GPS PRN 07 (Block IIR-M)
Current receiver time: 9 s
Tracking of GPS L1 C/A signal started on channel 0 for satellite GPS PRN 09 (Block IIF)
Current receiver time: 10 s

I didn't expect this! Is it just matching noise, or is there something wrong?

And if it is normal to see this when no signal is present, I thought I'd put this in the issues, so someone else doesn't think that these "Tracking" messages mean they're "almost" there with respect to signal strength, etc.

[JohnNatHaysys]

We have a system which generates a simulated satellite signal. Even when switched off we can see tracking messages. I assumed this was seeing a real satellite giving a strong signal. It usually lost lock, and never resulted in a Nav message. However, more recent testing suggest the acquisition pfa value may result in a 'lock' on noise. Any views on a suitable pfa value?

[RobertSwirsky]

It's definately noise! I can turn the gain to zero and put a 50 ohm resistor on the input and I still get tracking messages. (I'll never see a NAV though.)

I'll have to look into the "acquisition pfa" setting I see that it defaults to zero, which I take to mean "no threshold" for a false decoding.

I'm having trouble using GNSS-SDR on real signals in general. I need a very strong signal to get it to work. I can get it to work from a simulator, but I don't have a solid enough signal from real satellites (in a place where a uBlox module will have no problem).

I tried from a USRP 200, 205, and an x310 and none of them was able to hear. Yes, I'm using a gain antenna, and I have a bias-T injector to supply power (and I verified that the antenna was getting power).

[JohnNatHaysys]

Well, I can confirm that a RTL-SDR V3 dongle connected to a good external antenna (Tallysman triband) works perfectly well. Replacing the antenna with a 50R and a DC block ( to prevent the bias T driving 50R) resulted in an occasional tracking message - perhaps one or two a minute on average. The loss of lock occurs 15s later - which is a function of some timeout counters, even if there is a brief DLL match.

[RobertSwirsky]

Thanks! I tried an RTS-SDR using the GNSS-SDRLIB software and didn't get a signal either but I was using one of those little square black cheap antennas. (I verified it was working by putting the simulator into a dummy load near the antenna).

Maybe all the antennas I have are junk! I have some old "marine" style antennas I thought were good (e.g., Trimble 57860 https://novotech.com/bullet-gps-antenna-210.html ) , but maybe not (though they work with uBlox). I'll try getting a the same "Tallysman triband" you have and put it on a pole on the roof.

[JohnNatHaysys]

Our gnss-sdr based system appears to detect non existent signals! Locking on noise I can understand but we also see GPS NAV messages with a constant pseudorange from channels that are not supplied with a signal. It seems to be a function of the configuration. Anyone else seen this?

[jmfriedt]

Tracking is meaningless, just telling you that a given SV was assigned to a given tracking channel and searching for the signal. It has no relation to the actual existence of a signal, it is the internal state of gnss-sdr.
Only when seeing
GPS L1 C/A tracking bit synchronization locked in channel ...
can you consider that at actual signal was seen, strong enough if you reach
New GPS NAV message received in channel
and hopefully if enough SV are visible, you achieve the ultimate dream of
Position at ... using ... observations is ...

[JohnNatHaysys]

Not sure I agree 100% with this. If there is NO signal then we never get a tracking message. However, I can now confirm that because PRN cross correlation rejection can be as low as 24dB then as the signal is increased, the software will generate a spurious correlation output. Since this also carries the Navigation message then it will get bit synchronization and output a pseudorange.

We note that later Nav Message formats support including the SV (PRN) ID in the message. This makes rejecting these 'ghost' signals much easier.