[time-nuts] HP 5372A vs. 5370A

Ed Palmer ed_palmer at sasktel.net
Wed Feb 9 21:49:32 UTC 2011


SAIDJACK at aol.com wrote:
> Hi Ed,
>  
> excellent email. You raise the SNR of this forum!
>   
Thanks Said.  It's rare that I have a chance to contribute.  Many of 
these discussions are way over my head, but I'm learning.
>  
> I have tried to measure the CW-12 output myself with phase noise and Allan  
> Deviation equipment, and the constant cycle jumps prevent these from giving 
>  sensible results. Both PN and ADEV plots look awful, many orders of 
> magnitude  noisier than even an average oscillator. Glad to know that  your 5372A 
> could get the job done.
>  
> btw: jumping cycles 200 times a second is not disciplining an oscillator,  
> it's numerically controlling a frequency. Thus the CW-12 cannot  be 
> considered a GPSDO due to the massive 8ns phase jumps. 
Absolutely correct.  Navsync doesn't call the CW-12 a GPSDO.  They 
explicitly say that the 10 MHz is a numerically controlled oscillator.  
They even have an app note on cleaning up the output.  I was curious to 
see if I could use it as a 'new, improved' version of the Jupiter GPS 
receiver with it's 10 KHz output.  Haven't gotten around to it, but I 
believe others have.

Ed

> But for many  apps such as 
> driving a microprocessor clocks this may be good enough, even  though one has 
> to be careful that their internal PLL's can lock to such a noisy  (jumpy) 
> source. Of course the number of corrections per second should have  a very 
> strong correlation to temperature as the TCXO used on the CW-12 drifts  over 
> temperature.
>  
> bye,
> Said
>  
>  
> In a message dated 2/9/2011 13:14:16 Pacific Standard Time,  
> ed_palmer at sasktel.net writes:
>
> The  CW-12 also has a 10 MHz output.  I used the Histogram Time Interval  
> function to measure the periods of 100 Million cycles with a resolution  
> of 200ps.  It took less than 30 seconds to measure, process, and  display 
> the results.  The results showed that there was a normal  distribution 
> around 100 ns and a second normal distribution around 92  ns.  The 
> difference is approximately the period of the internal clock  (120 MHz).  
> This told me that the 10 MHz is kept on frequency by  occasionally 
> shortening the period by one cycle of the internal  clock.  For my unit, 
> this happens about 200 times a second.  I  know this by the ratio of the 
> number of long periods to short  periods.  This behaviour explains why 
> Navsync warns that this output  needs to be cleaned up before using it as 
> a frequency  reference.
>
> _______________________________________________
> 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_lists.febo.com mailing list