[time-nuts] another Ebay mixup, 5370

SAIDJACK at aol.com SAIDJACK at aol.com
Sun Jun 10 07:49:03 UTC 2007


 
In a message dated 6/9/2007 23:54:50 Pacific Daylight Time, jmiles at pop.net  
writes:

>Measurements taken from 3325B at 1 MHz:

>5370B  reference:       INT          EXT  (same 5087A as 3325B uses)
>Square wave, 2v pk:    15-20  ps     11-12 ps

>-- john,  KE5FX





Hi John,
 
now I am jealous. I checked the switches, they all seem to work as  required. 
Wanna trade :) ?
 
Tom wrote:
 
>What definition of jitter are you using? Are you looking  at
>every zero-crossing at 10 MHz? Or every Nth one? What
>sample  size? Or would you rather know the ADEV at tau
>1 second?  etc.

>/tvb

Hi Tom,
 
the Wavecrest can be set up with arbitrary bin sizes. In this case it was  
1000 measurements per bin I think (default setup). Going higher doesen't  change 
the measurement much, and the Manual states the same.
 
The unit uses zero crossings.  I usually AC-couple square-waves into  it.
 
The unit calculates RMS jitter at the touch of a button, and I used all the  
standard settings.
 
Not sure what the hold-off is, but it's asynchronous. Basically the unit  
starts as soon as it can after a stop event. The histograms I sent are  averages 
of many measurements, there is some info in the picture.
 
On a good oscillator the histogram is a perfect Gaussian curve, with about  
2.7ps RMS jitter (basically the internal jitter noisefloor of my Wavecrest  
unit).
 
I once correlated the results of the Wavecrest numbers with an Agilent  
86100A 20GHz scope on my best OCXO, and the RMS jitter numbers were within  
200femtoseconds of what the Wavecrest gave!
 
Since the jitter is so large, any good scope should be able to show  the 
5370B jitter, such as a 54720D or similar (has about 60ps noise-floor if I  
remember correctly). The >200ps jitter on the 5370B should be visible on  that, but 
then again maybe it's a fault in my unit :(
 
I get about 100ps best case RMS jitter on my measurements with the internal  
OCXO, and about 30-50ps using a Fury as a reference.
 
Still think besides the noisy time-base that the unit just needs a very  
careful calibration...
 
I wonder who has the needed equipment to do a by-the-book calibration on  
these guys.
 
bye,
Said



************************************** See what's free at http://www.aol.com.



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