GPS time sync

[p98d]

I am using a GoPro Hero13 and I am trying to synchronize the video that I am getting out of it with a precise measurement system which is calibrated with UTC time.

I have activated GPS time sync on my GoPro through GoPro labs, but I am seeing an issue. I have plotted the difference between GPS time and the time stamp of the measurement over time stamps (as recorded by the GoPro internal clock).

The result I get is plotted in the image, one would expect a horizontal line for a perfectly calibrated clock, and a sloped line for a clock running slow or fast. What I did not expect was the sawtooth pattern that I am seeing here. It almost seems like every 100s or so the clock is being reset.
Do you have more information on how the time sync function works? what am I seeing here?

[dnewman-gpsw]

All the SYNC function does it use the GPS time to set the camera's RTC, which is used to set the timecode for a file (only set once per clip.) Timestamps are not affected, and timestamps are not using the RTC. So this could be aliasing between the crystal driven video timestamp values, and GPS samples. Yet just did a 10-minute capture, plotted the difference between timestamps and GPS9 time are got this:

So over 10m40s the internal clock to GPS clock drifted ~0.025s. But I'm not able to reproduce your sawtooth.

Here is the CSV
GX020004.MP4.GPS9.csv

[p98d]

That is interesting. I cannot reproduce it with your data, but I do see it with mine.
Here is a clearer image, on the right you can see it zoomed around one of the jumps.

I tried to simulate the issue and I managed to get similar results by assuming that the camera clock is ever so slightly slow, and querying the GPS sensor at 9.99Hz, while the GPS sensor is posting a new UTC time at 10Hz. That produced similar results.

Here is a .json with all relevant data
GX013601.json