[time-nuts] Ultra Low Phase Noise 5 to 10 MHz Frequency Doubler

[Gerhard Hoffmann]

Am 05.02.2013 22:29, schrieb Rick Karlquist:
> This circuit was already in extensive use 40 years ago at Zeta
> Laboratories, where I used to work.  It's an OK circuit but
> I don't see what advantage it has over a full wave bridge doubler.
> The performance of this circuit is dependent on the source impedance
> of whatever is driving it, and the drive level, and the tuning
> of the output filter.  In most cases, you should provide an
> input filter as well.

I think I have seen similar multipliers with E88CC / 12AX7-like parts,
so it is probably not new, but the Schottky push/pull rectifier has
11 dB loss, an abused ring mixer has even more. That requires a
low noise post amplifier in nearly every case.

I have tried the NIST circuit that has been shown in this group with
2* NXP BF862 and Digikey transformers made by Pulse, that yielded
a doubler gain of +3dB, so one can burn 3 dB in a filter and
proceed to the next doubler without interstage amplifier mess.
Not bad for 1 square-centimeter and parts cost of $5 or so.
(Most is the transformers.)

That worked broadband to f-in = 20 MHz; I'm trying to extend this
to higher frequencies, probably at the cost of tuned circuits, at least
on the output side.  The next board run will be fun.

Phase noise seems to be OK, but we did not spend more than
20 minutes on it, so I'd like to repeat that more carefully and
also compare the Timepod to the E5052B. Source was a 5MHz BVA.

regards, Gerhard