[time-nuts] Adjusting HP 5065A frequency

WarrenS warrensjmail-one at yahoo.com
Mon Oct 22 22:45:03 UTC 2012


Tom said:

>In general, when you discipline a OCXO you get that characteristic ADEV
>"hump".
>At some point there is a cross-over and you know/assume that at that point
>each must be contributing 1/sqrt(2) of the noise.

Although putting the Hump at the cross over point is typically a good
compromise, there are other choices which give different trade-offs.

Attached is a excell plot showing the effect of different PID tunings
on the Hump's size and location.
This spread sheet (originally is from Wenzel's site) was for plotting the
effect of the RC H/W values for a PLL when used as a clean up Osc,
but can be used for setting the PID values of a disciplined GPS
such as a Tbolt with a few changes and using different data.

 ws

***********************
----- Original Message ----- 
From: "Tom Van Baak" <tvb at LeapSecond.com>
To: "Discussion of precise time and frequency measurement"
<time-nuts at febo.com>
Sent: Monday, October 22, 2012 4:32 AM
Subject: Re: [time-nuts] Adjusting HP 5065A frequency


Hi Bert,

Not sure about the range/resolution. That would depend on how the standard
is used and what its frequency drift rate is. The stability doesn't have to
be too much better than the standard itself.
The Austron 2055 resolution is 1e-14, IIRC. The Symmetricom AOG is 1e-19
(overkill).

What I've found in some GPSDO and passive atomic standards (e.g., Rb or Cs)
is that as soon as you turn on the DAC and enable the loop, you get more
short-term noise, say in the range of 1 to 100 seconds. That's why for best
stability you always switch off the loop during a sensitive measurement.
Many older Cs had a "Cs off" switch for this. Not only did it conserve
cesium but it also means you're running straight off the high-quality OCXO.
This is also true for GPSDO, like the TBolt which allows you to turn off
disciplining with a s/w command.

In general, when you discipline a OCXO you get that characteristic ADEV
"hump". This is expected, a natural byproduct of combining two unknowns, one
that's assumed to be better at short tau and worse at long tau (e.g., OCXO)
and one that's assumed to be better at long tau and worse at short tau
(e.g., Rb cell, or Cs beam, or GPS receiver). At some point there is a
cross-over and you know/assume that at that point each must be contributing
1/sqrt(2) of the noise.

To see the humps in living color, refer to:
http://www.leapsecond.com/pages/gpsdo/

To answer your question about short/medium/long, I guess in this case short
is tau left of the hump; medium is the hump, and long is tau to the right of
the hump.

/tvb

-------------- next part --------------
A non-text attachment was scrubbed...
Name: PID effect on ADEV2.gif
Type: image/gif
Size: 55191 bytes
Desc: not available
URL: <http://febo.com/pipermail/time-nuts_lists.febo.com/attachments/20121022/76c920bb/attachment.gif>


More information about the Time-nuts_lists.febo.com mailing list