On Apr 15, 5:36 pm, unruh wrote: > For ms accuracy you would just need usb for power and serial port, but > work by Lichvar suggests that you cannot just use the serial port if you > want usec accuracy.
It appears your bias against using DCD for PPS clouded your comprehension of what Miroslav said. He said he measured a 20-40us offset between a timestamp taken in the serial driver interrupt routine and one taken in userland by gpsd. Incidentally, I see about 40us offset on a PII 400MHz box between serialpps.sys timestamps and ntpd user-mode timestamps. Of course, if it's consistent it can be calibrated away. I also see about twice the jitter (~4us vs ~2us) from the userland timestamps, taken by ntpd running the realtime priority class, as with interrupt timestamps. The extra jitter can't be calibrated away. The more loaded the machine, the more the userland route will suffer compared to interrupt-time. Cheers, Dave Hart _______________________________________________ questions mailing list [email protected] http://lists.ntp.org/listinfo/questions
