[time-nuts] Re: E5500 blues

[John Miles]

> So, with a 5 MHz source, a RF generator that is doing DCFM lock of PLL
> and then analyse spectrum of that, that fails to lock this PLL loop, and
> I fail to get a monitored beat note that makes sense. So, I was hoping
> someone had done this and could help me with the teething problems. Once
> that is done I intended to redo this at 180 MHz and 1440 MHz just for fun.
> 
> Thus, 3047, 3048 and E5052 becomes off topic, and not part of this exercise.

If your 5 MHz source is a crystal oscillator, it's probably injection locking.  The phase detector has relatively poor port isolation, maybe 30 dB or so.  You almost always have to use isolation amps (note the plural) for crystal oscillator measurement with these boxes.  

Another thing to check is the VCO gain parameter.  This doesn't have to be spot-on, but if you enter a value that's too far off, the initial beat note lockup process may not work properly.  

If you are using an optional frequency counter as part of the setup, try disabling it and letting the FFT analyzer measure the DUT properties.

It's worth getting the E5500 working IMHO, as it's capable of doing everything the 3048A system could without taking up a whole equipment rack.  Rick's right in that these instruments are sort of a pain in the neck to work with, but I wouldn't say that they're primarily meant for residual tests.  They are very much suitable for general purpose AM/PM noise measurements as long as you have the necessary peripherals and patience.  

-- john