Navspark timer drift, #2

Second run, this time comparing the Navspark's internal CPU clock against the GPS time. Samples of both clocks are taken once per second, and the amount of time that has passed is calculated for both. The ideal CPU clock would be a perfectly horizontal line at 0.000 seconds. You can see from this graph that the CPU clock is running slightly slower, losing roughly 5 milliseconds per 40,000 seconds. This run was 101,313 seconds (1 day 4 hours).

The green line is a best fit y = a*x + b. The slope of the best fit line is: a=0.000000124 (or 124 parts per billion). This is very close to my last result (126 parts per billion), and the difference can easily be explained by the greater precision of this test.

The code used on the Navspark: Internal Timestamp Trigger