NRF9160: GPS Tracking satellites, but can't get fix.

[mlang3440]

Hi, we have a custom board using the nrf9160, running modem firmware 1.1, and using the nrf9160_gps driver.

Sometimes we see messages like this:

Tracking: 8 Using: 0 Unhealthy: 0

Where the GPS is tracking satellites but is not using any, and never gets a fix. Eventually we time-out and stop the GPS. But this behavior continues until the next device is restarted.

Is this a bug with the modem firmware? Is there any fix for this.

Thanks

[phipi]

We are still trying to reproduce this issue on a device running the modem trace however, we have found some pattern on one of the devices that has been running for a longer period. The pattern is as follows:

1. The device runs as expected with great (30~60s) TTFF
2. After approximately 6~7 days we enter this failure mode where the device is tracking many SVs but using 0 and times out (20 minutes)
3. On the next GPS fix attempt (6 hours later), once again it tracks many SVs and uses 0.
4. At this point we do a software reset
5. On the next GPS fix attempt (6 hours after software reset), the fix is successful (again with a very good TTFF)
6. After this the next few GPS attempts all time out
7. Steps 4-6 repeat maybe 3-4 times before we finally come out of this mode and things work again for another 6~7 days

Based on this it almost seems like an almanac update might be somehow responsible.

• Can Nordic confirm how often an almanac download is attempted?
• What happens if an almanac download isn't able to complete in the 20 minutes? Is it thrown out until the next fix attempt or is partial download possible?
• What happens on reset? Does the firmware maybe revert to the last known almanac and that's why our first fix is successful before it attempts to download the new almanac?
• Is there anyway to get information about the status of the almanac download status in the device firmware?

[Øyvind]

Hello phipi, 

I will move this discussion to the private ticket you have registered, as that is where I will ask questions regarding your code.

To answer your last questions here, I have forwarded your questions to our nRF91 GPS expert, so far the issue leans towards software. 

"The device never needs the almanac to get a fix, it should be able to do a cold start without any prior knowledge of anything whatsoever. The only thing the almanac would do in this case is to further speed up the TTFF by aiding which satellites to search for.

In this case, the device seems to be always able to find the satellites, but using 0 would be an indication that it cannot get the ephemerides and/or TOW downloaded from the satellites.

A log would indeed be helpful here. We can try to recreate the situation, but that will take 6-7 days for things to start happening if I understand correctly.

One way or another this is an SW issue somewhere. Since the device is able to get fixes, the HW seems to be fine. Since it is trying to get a fix every 6 hours, it is always cold starting, and as said, a cold start does not need any prior knowledge of anything.

But somehow the process of synchronization to the satellites seems to get corrupted and never finalized. It is useless to try to speculate here what could go wrong. If we don't get a log from the customer, I will have to forward this for recreation in our own lab. There is a good chance we would find nothing wrong => customer's logs needed anyway."

mlang3440 said:

Tracking: 8 Using: 0 Unhealthy: 0

Have you tested our GPS sample alone? Do you get the same results? Have you tried your project on an nRF9160 DK?