[time-nuts] How do I know my GPS stabilizedoscillator is working?

Magnus Danielson cfmd at bredband.net
Mon Jul 31 14:42:08 EDT 2006


From: "Tom Van Baak" <tvb at leapsecond.com>
Subject: Re: [time-nuts] How do I know my GPS stabilizedoscillator is working?
Date: Sun, 30 Jul 2006 19:54:34 -0700
Message-ID: <000801c6b44c$a828d400$f10ff204 at computer>

Tom,

> Drift -- Another consideration is when you know if
> the DUT is drifting forward or backward in time. In
> this case arrange the start/stop channels so that,
> over time, you always diverge from zero rather than
> approach and cross zero. This argues for putting
> the device with higher frequency as the start channel
> (lower frequency as the stop channel). Or is it the
> other way around.

This is a problem for some counters. Others will not have that problem, in
particular those with zero dead-time. Then event counters and time counters
with interpolators will keep you on track regardless. It also depends on the
particular measurement selected. They may vary in particular details, which may
bite you. Choose the right one and your post-processing becomes a bliss of
simplicity.

> 1PPS -- Lastly, there is one convention I found handy
> with 1 PPS sources, specifically those GPS boards
> that are designed to suppress the 1PPS signal when
> they loose lock. In this scenario make the reliable
> 1PPS ref the start channel and the GPS 1PPS the
> stop channel. If lock is ever lost -- your TI readings
> will reveal the number of missing pulses. For example:

We in the telco buissness do this, it is really necessary and the only way to
convey a known error. Even when you have a side-channel with more info (like a
serial line), you need to be prepared for the lack of "propper signal" so this
is a good way, since relying on the side channel may lie to you if you loose
signal, and besides, no side channel looks the same and may even be non-
existent or otherwise unreasnoble to handle.

> (REF=start, GPS DUT=stop)
>     8.9755  us
>     8.9368  us
>     8.8986  us
>     8.9214  us
>     8.8829  us
>     34.0000088063 s
>     8.8310 us
>     8.7332 us
> 
> If you wire the TIC the other way around, then those
> missing 34 pulses won't even show up in your data.
> For example:
> 
> (GPS DUT=start, REF=stop)
>     8.9755  us
>     8.9368  us
>     8.8986  us
>     8.9214  us
>     8.8829  us  [followed by 34 seconds of silence]
>     8.8063  us
>     8.8310  us
>     8.7332  us
> 
> In summary, there's isn't a right way and a wrong way.
> There may be, in some cases, a convenience of one
> over the other.

Well, in your above measuring case with 34 missing pulses, you *DO* have a
problem if you are going to do Allan Deviation (ADEV), Modified Allan Deviation
(MDEV) or Time Deviation (TDEV). Replacing the start and stop as you did above
will destory that information. The gap will cause a measurement error.

For ADEV/MDEV/TDEV measurements you also want to compensate out driftrate
before doing the actual calculation. The frequency error is not a problem for
the calculation, but the frequency error can be a problem for some, but not
all, counters.

Cheers,
Magnus



More information about the time-nuts mailing list