[time-nuts] Testing frequency using NTP Bruce GPS ps
Mike Monett
XDE-L2G3 at myamail.com
Sat Oct 4 00:13:46 EDT 2008
"Tom Van Baak" <tvb at LeapSecond.com> wrote:
>Hi Mike,
Hi Tom,
I want to thank you for the very nice reply you gave to my email a
while ago. Unfortunately, I am still learning about precision
frequency references, and I don't have enough knowledge yet to give
you an intelligent reply, or even ask a sensible question. But I
will as soon as possible.
In the meantime, your site is invaluable for newcomers like me.
Thank you very much for taking the time to organize it so well.
> Rick's CNSC02-O1 implementation has been discussed here a number
> of times over the years (google the archives). It uses a
> programmable digital delay line to compensate for the receiver
> reported quantization error on each pending 1 pps.
The single-shot is supposed to eliminate the 1PPS jitter? So it must
be triggered on the 1PPS, and the variable delay gives an average of
half the clock period?
Thanks for the tip. I'll research the archives and see if I can some
up with some better understanding of how it works.
> Here's a quick plot of an M12+ receiver without (pink) and with
> (yellow) sawtooth correction:
>http://www.leapsecond.com/pages/gps4/cns-1pps.gif
Thanks. I had not found that page yet. Just judging by eyeball, the
pink and yellow traces don't seem to track very well. Any reason?
And what's the blue trace for? I can't seem to find the parent page,
so I don't know if you have already explained it.
> No, the correction doesn't handle for the case of prolonged "zero
> beat" errors.
> As you continue your reading about sawtooth errors see also:
>http://www.leapsecond.com/pages/m12/sawtooth.htm
Yes, I had found that page. The top graphs make me believe my method
will work fine.
The bottom graphs say it won't:)
>http://www.leapsecond.com/pages/vp/heater.htm
Yes, that page is also very encouraging. All I have to do is figure
out how to close the loop on a long delay in the feedback. I believe
the PRS10 can average up to 8 hours, so it should be possible.
>http://www.leapsecond.com/pages/vp/sawtooth.htm
Yes, Yes, Yes. I had found that page. It is very important. The
uniform distribution in the top graph says my method will work.
The bottom graph looks very similar to ones I got from measuring
jitter on hard disk drives back in the 80's. The graph contains a
great deal of information on what is happening in the read/write
channel, head/media magnetics, servo system, and external noise
sources. As soon as I saw that, I wondered if a similar analysis
method might work in this case.
>http://www.leapsecond.com/pages/58503-cns2/
Thank you very much. That page has a lot of very good information. I
like the way you take data and present it. The data is exactly what
is needed to illustrate an issue, and the presentation is so very
clear that anyone should understand it.
Overall, I was a little disappointed to find the sawtooth correction
only gives about a factor of 3 or so improvement. The results I have
obtained with my method have given over two orders of magnitude
improvement in noise reduction. That's with no optimization.
But this is a new area, and there is a lot to learn about what is
going on. So I am very anxious to get my hands on some equipment and
start developing the necessary circuits. They are really very
simple. Anyone should be able to do it with no problem.
I just got from eBay an HP 8566A, an HP 53310A including the 53305A
Phase Analysis software, an HP 3456A, two TEK 2467's, and some
National Instuments AT-GPIB / TNT cards with cables and drivers. The
Win98 drivers for the NI were courtesy of a poster on the TekScopes
forum, and I uploaded them to Didier's site for anyone else who
might be interested.
My big problem right now is none of my working computers have any
ISA slots for the GPIB. I have some old motherboards with ISA slots,
but no software drivers to get the motherboard working.
And, of course, I have to get a TBolt and a rubidium oscillator.
Hopefully that's the next step.
>/tvb
Thanks,
Mike Monett
More information about the time-nuts
mailing list