[time-nuts] Jupiter modules: 1pps or 10kHz to lock an OCXO ?
didier at cox.net
Mon Dec 22 13:20:20 UTC 2008
If you are trying to have precise enough frequency for ham microwave
applications for instance (1E-9 would be 10 Hz at 10 GHz), the 10kHz output
of the Jupiter can be used to your advantage to phase lock a cheap VCXO with
a short time constant. A perfect example is the Jim Miller design which Tom
(and others) has tested and found to perform very well. A short time
constant is advantageous because it means the system will phase lock quickly
after power up, and many ham microwave stations are mobile and do not remain
powered all the time.
However, if you try to phase lock a relatively good OCXO like the HP 10811
or better (and achieve 1E-11 or better), any time constant faster than 20
minutes or so is likely to actually degrade the available performance of the
OCXO because the GPS signal has a lot of fluctuations in the short term due
to atmospheric perturbations and satellite position in the sky, refered to
as geometry, so a fast loop will force the otherwise stable OCXO to follow
the short term instability of the GPS signal, not what you want. With that
kind of time constant, it makes little difference if the phase detector
works at 10 kHz or 1Hz, aside from saving a few dividers. As it has been
pointed out, long time constants are better implemented in software, due to
the negative effect of large value resistors on noise.
Tom has a lot of very useful charts on his site www.leapsecond.com showing
the relative effects of GPS jitter and OCXO drift on overall stability.
1) if you have a Jupiter and your needs are modest, use the 10kHz and an
analog loop, it will make your life easier (fewest parts), but depending on
the reference oscillator you use and the performance you try to achieve, do
not speed up the loop unnecessarily.
2) if you only need 1E-9 stability, the Jim Miller design is excellent and
so simple, anything else would be a waste.
3) if you need 1E-11 or better, you will need much more hardware, including
a good quality GPS receiver, antenna and OCXO and a microprocessor to
implement the loop filter, and probably some holdover algorithms,
essentially what you will find in a Fury, Z3801, Thunderbolt or Brandywine
4) if your needs are in between, you have a lot of room for experimentation.
> -----Original Message-----
> From: time-nuts-bounces at febo.com
> [mailto:time-nuts-bounces at febo.com] On Behalf Of Bob Q
> Sent: Sunday, December 21, 2008 8:34 PM
> To: Discussion of precise time and frequency measurement
> Subject: Re: [time-nuts] Jupiter modules: 1pps or 10kHz to
> lock an OCXO ?
> I thought of something else. Time to lock can be really
> long. Say the OCXO can swing +/- 1 ppm, +/- 10 Hz at 10 MHz.
> The +/- 1 ppm also applies to 0.5 Hz. To swing 180 degrees
> at 0.5 Hz, assuming the low pass filtered control signal
> pushes the OCXO the full 1 ppm, will take 1e6 seconds unless
> you use a circuit similar to what I used to prealign 50 Hz
> signals in the 100 Hz Motorola Oncore design.
> Bob Q.
> ----- Original Message -----
> From: "Patrick Reynaert" <preynaert at yahoo.com>
> To: "time-nuts" <time-nuts at febo.com>
> Sent: Sunday, December 21, 2008 2:13 PM
> Subject: [time-nuts] Jupiter modules: 1pps or 10kHz to lock an OCXO ?
> > Hello,
> > Using the 10kHz output to lock an OCXO is common practice
> and seems to
> > give good results.
> > Has anyone tried to use the 1pps output? One? could first make a
> > simple divider to create a 0.5Hz reference with 50% duty cycle, and
> > then use this signal to lock an OCXO. This would also allow
> the use of
> > other OEM GPS modules that don't have the 10kHz but only
> the 1pps signal.
> > Of course, as the 0.5Hz signal is much lower in frequency
> than 10kHz,
> > the short-term stability is pushed more to the OCXO, making the
> > requirements of the OCXO even harder.
> > But maybe the 1pps signal has better stability than the
> 10kHz signal?
> > Or is this the same since these signals will only influence the
> > long-term stability of the oscillator and have the same accuracy?
> > Any thoughts are highly appreciated.
> > Thanks,
> > Patrick.
> > _______________________________________________
> > time-nuts mailing list -- time-nuts at febo.com To unsubscribe, go to
> > https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
> > and follow the instructions there.
> time-nuts mailing list -- time-nuts at febo.com To unsubscribe,
> go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
> and follow the instructions there.
More information about the time-nuts