[ntp:questions] Re: hardpps difficulty on an ARM-7

Ulrich Windl Ulrich.Windl at RZ.Uni-Regensburg.DE
Mon Feb 2 15:15:07 UTC 2004


"Eric Hammerle" <ehammerle at spectracomcorp.com> writes:

> I'm currently porting the ppskit version 2.1.1 to an ARM-7 processor using the 2.4.20 kernel.  We are using a GPS antenna to receive a pulse, so I believe the pulse to be highly accurate.  Unfortunately once I have binded the kernel consumer to the gps 1pps, I keep getting the messages similar to:

Maybe test your system-time first: Using the "jitter.c" in the NTP
sources. If the clock seems stable, see if you can get stable
system-time during interrupts as well (using ppsctl.c, see usage()).

> 
> hardpps: bad pulse, delta = -39984010 ns
> 
> and every 24th bad pulse will output:
> 
> hardpps(0.481486004): early pulse, delta = 960006929 ns

You are 0.4 seconds off.

> 
> instead of the former message.
> 
> Any insight to this problem would be greatly appreciated.
> 
> - E. Hammerle
> 
> PS: Has anyone else out there tried porting this to an ARM7?

No, but I'd like to give you a few hints.

Regards,
Ulrich



More information about the questions mailing list