[time-nuts] Neat toys on eBay for PN measurement

[Grant Hodgson]

John

Don't know if you got an answer, but this is my understanding :-

The 11729B needs to have an external 640MHz reference input for 
measurements above 1280MHz.  The 11729C has a SAW-based 640MHz 
oscillator built-in, but it can take an external 640MHz reference signal 
if one is available, such as from the 8662A.

The spec. for the 11729C internal 640MHz osc is not as good as that of 
the 8662A at offsets of up to about 20kHz or so, but at offsets above 
50kHz the 11729C SAW is better than the 8662A's xtal osc.

AFAIK, there is no physical difference between an 8662A with and without 
option 03.  All this option does is mean that somebody has done an extra 
set of measurements on the 640MHz reference output to confirm that it 
did indeed meet the spec.  There was also an option H40 which had an 
even lower noise 10MHz reference which reduced the phase noise at 
offsets up to about 100Hz by 8dB for both the main output and the 640MHz 
reference output.

Also, the 640MHz reference input to the 11729 does not have to be phase 
locked - as long as it does not drift at a rate outside the 11729's PLL 
hold-in range then it will work.  The 11729's PLL will pull the signal 
generator to compensate for drift in either of the two reference inputs 
(640MHz and/or external signal generator).

I've got an 11729B, but no 8662A, so I've started working on the design 
of a very-low phase noise 640MHz source based on an 80MHz crystal osc. 
But until that's ready, I've still been able to make measurements below 
1280MHz without the external reference input.  The Racal 9087 is a 
suitable reference signal generator up to 1.3GHz, which can be phase 
locked, and at offsets greater than about 1kHz it actually has lower 
phase noise than the 8662A.  Other low-noise generators are also 
avaialble...

And for those that can't resist tinkering with the fine offerings from 
the HP labs, I've been wondering if it would be possible or practical to 
replace the internal Step-Recovery Diode comb generator in the 11729B/C 
with one of the Non-Linear Transmission Line comb generators from 
Picosecond Pulse Labs.  PsPL are quoting a 15-20dB reduction in phase 
noise for their NTL comb generator compared to an SRD.  Seems almost too 
good to be true; don't know how much they cost for one-offs but it would 
be interesting to see if it could be used to reduce the residual phase 
noise of the 11729.

regards

Grant

> John Miles wrote :-
> 
> That's something I meant to ask you about, Rick, as a follow-up to an old
> Usenet post of yours from 1995.  From looking over the block diagram in the
> 11729B-1 app note, it appears that there is no reason why you couldn't feed
> any sufficiently-clean 640 MHz signal in.  Obviously, you want to drive it
> with the cleanest source you can find, but I don't see any other
> constraints.
> 
> But some of the 8662As did not have the optional (003) "specified SSB phase
> noise for rear-panel output" feature, including mine.  They all seem to 
> have
> provided a 640 MHz output at an unused internal SMB jack, though.  Is there
> something special about the reference multiplier section in an option-3
> 8662A that actually improves the noise level available at this jack?
> 
> I'd already added a BNC jack to the rear panel to bring the 640 MHz clock
> out in anticipation of buying or building a downconverter, and I expect it
> will work OK with this 11729C, but I am not sure whether I should expect 
> the
> fully-characterized option-3 noise performance, or something worse.  Any
> thoughts on that?
> 
>  >From your Usenet post I understand that there is supposed to be a feature
> that lets you turn the built-in 640 MHz SAW filter into an oscillator in
> case an 8662A isn't available, but I also understand that this is really
> only a utility/test function.  If there's no way to phase-lock the 
> resulting
> SAW oscillator, I can see why.
> 
> -- john, KE5FX
> 
>