[time-nuts] Correcting measured PPS values with receiver sawtooth correction values

[Mark Sims]

The F9 has a totally new internal architecture / processor, so it's best to make no assumptions about how it works compared to older models.

For the LEA-6T the correction strategy is clearly "add  previous sawtooth value".  That does not work for the F9P.  

The two candidates for the F9P are "add current sawtooth" or "subtract previous sawtooth".  I had been favoring "subtract previous" but some more detailed analysis seems to favor "add current".  The difference between the results of the  two strategies is rather subtle and it is very hard to determine which is the correct one.

----------------

> I don't think that uBlox changed the defaults over their previous
generation, so whatever worked for the M8 series should be applicable
here.