[time-nuts] GPS clock to RT-Linux for pulsar machine
nolan at naic.edu
Fri Jul 1 18:24:04 EDT 2005
On Fri, Jul 01, 2005 at 02:55:42PM -0700, David Forbes wrote:
> Thanks for the info. I plan to store timestamps with all the data
> samples, so any ntp drift can be spotted and accounted for by
> comparing it to the stored 1PPS samples.
> I don't anticipate CPU interrupt overload trouble since the quantity
> of data is so low, only ~20K words per second. At 80 GHz you don't
> need to de-disperse the pulse, so one data value per polarization
> channel is sufficient.
True. It just needs paying attention to. Our non-cpu-intensive
datataking has no problem arming from ntp. Just make sure you leave
plenty of buffer before (and after!) the 1 PPS trigger.
> I may design the interface hardware to synchronize the ADC clock to
> 1PPS as you suggest, since it would increase the phase accuracy. It's
> easy to do.
I think you will need to do that.
Most of our pulsar users start on a 1 PPS driven by our maser, then we
try to calibrate the maser offset wrt GPS as a function of time. We
keep the maser to within 250 ns, and monitor the offset to 25 ns or
better. If you're doing long-term timing, your users will insist on this
sort of accuracy. And if you give it to them, they will want better...
Mike Nolan +1 787 878 2612 Fax: +1 787 878 1861 nolan at naic.edu
Arecibo Observatory, HC3 Box 53995, Arecibo, Puerto Rico 00612
More information about the time-nuts