STM32F4 time experiments

Goal

1ms or better accuracy on a USB-connected GPS.

Hardware

stm32f4 discovery board using a 10.5mhz internal timer is measuring the PPS interval from a Fastrax UP501 GPS reciever (claims "+/- 50 ns (RMS) accuracy"). The stm32f4 is connected via full-speed usb to a pc

Software

See: usb gps

Changes from Run 6

Send timestamp data to NTP on a 32 second interval instead of 16 second

Graphs

Histogram

This takes all samples within a 32 second window and graphs the number of samples within X µs of the mode

Comparison between run 6 and run 7:

percentile Run 6 Run 7

50% 3 µs 5 µs

90.0% 22 µs 20 µs

99.0% 46 µs 43 µs

99.9% 56 µs 53 µs

Run 7 and run 6 were very close

percentile	Run 6	Run 7
50%	3 µs	5 µs
90.0%	22 µs	20 µs
99.0%	46 µs	43 µs
99.9%	56 µs	53 µs

PC clock

Loopstats graph - local clock error (PPM, green) and USB/GPS offset (µs, red)

Overall conclusion

A longer poll time did not help

Future enhancements

Use USB High Speed (avoid the PC's built-in High Speed/Full Speed hub) - I'd need to get a USB HS PHY for this
Use a OXCO or other higher-quality oscillator for the microcontroller
Add a display to the microcontroller