[time-nuts] Very stable synthesizer, alternative to PTS(Programmed Test Sources) x10 or 040?

Jim Lux jimlux at earthlink.net
Thu Jul 11 03:48:19 UTC 2013


On 7/10/13 12:29 PM, Didier Juges wrote:
> Jim said:
>
> "It's like a HP 8663B (not the modern Agilent E8663).. very low noise,"
>
> The Agilent E8663 has similar SSB phase noise spec as the older HP 8662A
> (-144dBc/Hz @ 10 kHz with option UNY, versus -143 for the 8662). You seem
> to imply they are different. Can you elaborate?
>
> Of course, the Agilent has many more features and 0.001Hz resolution, and
> the 8662 only goes to 990MHz (I think, I should know, I have two thanks to
> JohnM...), but are they that much different in pure phase noise or ADEV?
>
> Didier
>

It's not the phase noise that raised the problems for us. It's that when 
you program them for a sweep, it goes in steps that aren't phase 
continuous AND the behavior when you feed a signal into the FM input 
isn't the same. The HP 8663B was, at the core, a really good phase 
locked VCO, so when a sweep is programmed, the output is phase 
continuous as it sweeps.  This is a huge problem when you are testing a 
very narrow band tracking loop (our deep space transponders have a loop 
bandwidth of a few Hz)

I can't remember the details on the FM input, but it too has some 
behavior that we depended on.

We take the output of the 8663B and run it into a x7 to make the 7150 
MHz uplink and/or the 8450 MHz downlink frequencies.

Part of the reason we do a x7 is so that any leakage from the 
synthesizer isn't in band for our receiver under test. A typical input 
level for test is -150 to -160 dBm, so leakage at the wrong frequency 
can easily be more than the desired signal.




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