[time-nuts] Spec An for phase noise measurements

Grant Hodgson grant at ghengineering.co.uk
Mon Jan 21 20:18:07 UTC 2008


Matt

It depends on what you want to measure.  If all you want to do is 
measure the relatively close-in phase noise of (say) a single-loop PLL, 
then an 8560 is fine - I know, because that's what I used for some time. 
  But if you want to measure the noise floor of a VCO, PLL or other 
source then even the best spec. an. just does not have enough dynamic 
range, so other techniques have to be employed, such as the 
down-conversion techniques that have been discussed here recently.

There's a circuit on the Wenzel website for a suitable PLL/LNA that will 
do most of the downconversion, just add a suitable mixer and off you go. 
I was seriously looking at building one until an 11729B popped up for a 
price I couldn't refuse.  The only problem with building one is the time 
- your choice as to whether you want to build/buy.

And it also depends on how close to the carrier you want to get - most 
8590 series analysers go down to 9kHz, if you down-converted and want to 
get closer than that then you will indeed need another analyser.  There 
is an option on the 8560s that take the lower frequency limit down to 
30Hz, which is close enough for most applications.

Beware the options on the 8560 - option 103 replaces the nice OCXO with 
a cheaper TCXO; the stability and accuracy with opt 103 should be fine 
for most uses, but due to the poorer stability of the TCXO, you lose the 
1Hz Resolution bandwidth which may or may not be a problem.  Just check 
exactly what options you are getting in advance.  However,  option 103 
was $$$$ cheaper, which tipped the balance for me; I can use an external 
10MHz reference if I want greater stability and accuracy, but I still 
can't get the 1Hz RBW !

Also check the frequency range if you want to go above 22GHz - the 8563E 
will definitely go to 26.5GHz; some, but not all, 8653As stop at 22GHz.

I love the 8560s, they are an order of magnitude better than the 8590s 
for serious RF work.

regards

Grant

Matt wrote :-

> Message: 1
> Date: Mon, 21 Jan 2008 11:20:03 -0800
> From: "Matt Ettus" <boyscout at gmail.com>
> Subject: [time-nuts] Spec An for phase noise measurements
> To: "Discussion of precise time and frequency measurement"
> 	<time-nuts at febo.com>
> Message-ID:
> 	<ad9f78230801211120w2b4ae144g3d117cd2d343bba8 at mail.gmail.com>
> Content-Type: text/plain; charset=ISO-8859-1
> 
> I am considering getting a new spectrum analyzer so I can make better
> phase noise measurements than with my 8596E.  I've looked at the 8566B
> and the 8562 and 8563 since I need coverage to at least 6 GHz.  The
> 8566 is huge and ancient, though, so I think I'm leaning away from
> that one.  Anybody have other suggestions?  Some of the Advantest
> units seem to be reasonably priced on ebay, but it is hard to figure
> out what their phase noise performance is.
> 
> Also, what is the difference between the A, B, and E models on the
> 8560 series?  The A models are much cheaper on ebay.
> 
> Thanks,
> Matt




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