[time-nuts] 35601A as stand-in for 11848A ?

John Miles jmiles at pop.net
Sat Jan 5 18:17:22 EST 2008

> Hi guys,
> does anyone have any experience with "home-brew" quadrature-PLL PN
> measurements? Renting/buying a E5052A or TSC5120A does get to  be
> expensive..

Yes, quite a bit.  When you don't use its microwave downconverter or other
extended capabilities, the 11729B/C is just an implementation of the
principles in Wenzel's app note.

If you haven't seen the references I've gathered in the last FAQ entry at
www.ke5fx.com/gpib/faq.htm I'd suggest that as a potential starting point.

> I am using different mixers (Mini Circuits ZFM-150 and ZFM-4) and
> the noise
> floor coming out of these seems much higher than I expected. Any
> suggestions
> on mixers anyone? How about the low pass filter and amplifier?

What were you expecting to see, and what are you getting?

As far as I know, one DBM is as good as another.  Higher-level ones would be
preferable if you want to stay clear of the thermal noise floor at -174
dBm/Hz by using stronger RF/LO signals.

> I built an LC 8th-order 1.5MHz lowpass, and I wonder if the
> reflected  energy
> (at 10MHz, 20MHz, etc) is causing noise and  intermodulation in
> the diodes
> etc.

I didn't notice any change in the noise floor when I added a 1.5 MHz LPF
between the PD mixer and LNA in my 11729C.  This was a 3-pole ellpitic
designed for 50R in, 50R out.  You could have a point about IMD being a
problem, though.  Add a 3-dB pad to your IF output and see if your noise
floor gets >3 dB better?

> I am using an EL2125CS opamp (<1nv/rtHz noise) for the PLL and the Audio
> amp, with 40 or 60dB gain for the Audio amp.

What's the loop bandwidth?  Have you verified that it's as narrow as you're
assuming?  You can rule out PLL problems by testing two well-warmed-up OCXOs
and adjusting for quadrature manually with a dual-trace scope.

> Tried different power into the mixer (0dBm, +10dBm, +13dBm etc),
> this does
> not seem to affect the noisefloor, which I believe to be at least
> 25dB too
> high.

HP's approach with the 11729 was to put about 0 dBm on the RF port (from the
'reference', typically but not necessarily an 8662A) and apply the test
signal at about +11 dBm to the LO port, via an AGC-limited amplifier.  The
IF output is terminated by the 50-ohm LNA input port.

> Thanks for any hints,

I certainly don't claim to be an expert but I'm pretty careful with the test
methodology on my 11729C setup, and have become fairly familiar with it.
Let's hear a little more about what types of sources you're looking to

-- john, KE5FX

More information about the time-nuts mailing list