[time-nuts] Phase jumps of HP3325A synthesizer ???

[SAIDJACK at aol.com]

In a message dated 7/7/2007 15:29:22 Pacific Daylight Time,  
magnus at rubidium.dyndns.org writes:

> One  comment: as Tom explains the FireFox has a native DDS mode with 32 
bits   
> resolution, this is probably comparable to the HP3325.

>Not  by far. Check your facts. The HP3325A/B has the equalent of 48 bits,  
but
>its synthesis is of strict BCD form. If you want to acheive integer  
multiples
>of some power of 10, then you need such strict BCD forms.  


Magnus,
 
I guess 1 microsecond resolution at 21MHz bandwidth does mean more  than 32 
bits. I was just taking a guess since I don't plan on looking into,  or using 
an HP3325.
 
Nonetheless the FireFox Signal Generator in Fractional-N mode has almost 49  
bits of resolution across a bandwidth of 10 microhertz to 1640MHz. BTW: it 
also  doesn't need adjustment procedures like the HP3325B, it is GPS  Disciplined 
and integrated solid-state. Our next generation Signal Generator  will have a 
bunch more resolution, maybe 64 bits or so. That's pretty easy to  implement 
in our architecture.
 
There is always a tradeoff, using binary mode may give you a small phase  
creep, but it does give you a lot more resolution than BCD mode at the same  
bit-width.
 
Phase creep at 64 bit resolution on a 400MHz bandwidth gives a frequency  
resolution of about 5E-011Hz, I think that should be sufficient for it gives a  
resolution error of 5E-018 at 10MHz. It would take some time to measure that  
phase creep.
 
bye,
Said



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