[time-nuts] "Barnabus" HP 10811

[Richard (Rick) Karlquist]

On 4/9/2020 3:14 PM, Perry Sandeen via time-nuts wrote:
> Learned Gentlemen,
> Rick wrote:If you really need lower far out phase noise than the 10811 offers,you can redesign the 2nd and 3rd bufferamplifier stages. The 10811 designers knowingly degraded the phasenoise in those stages because of requirements to be backwardcompatible with 10544 sockets. They made a one-off demonstrationoscillator coded named "Barnabus" with ultra low noise. Italways seemed to be the proverbial "solution in search of aproblem."
> Well.....this being Time Nuts......
> 
> Is there a circuit that can use different parts and the same PC wiring in the 10811 to make a conversion?
> I have no ability to design anything like that but, OTOH plagiarism would be a viable option,especially since 10811's on Ebay are less than $80 almost all the time so........
> Regards,
> Perrier

I don't remember seeing the actual Barnabus circuit, or if I
did, I didn't make a copy of the schematic.  Having said that,
I can comment that when I designed the crystal oscillator for
the HP 10816 Rb standard, I plagarized :-) the 10811 circuit
as far as the grounded base transistor.  IE, the first stage
of the buffer amplifier.  That drove two more grounded base
amplifiers in cascade and then went to the RF output for the
user.  That part was basically plagarized from numerous NIST
papers written by Fred Walls and the usual suspects.  These
papers showed buffer amplifiers consisting of a degenerated
common emitter amplifier, with it's collector driving a
string of cascading grounded base amplifiers.  I didn't use
a common emitter amplifier at the beginning of the buffer
chain.  There were many of these amplifiers presented at FCS.
I never tried to figure out who first published it, if not Fred.

I think you could make a tiny PC board with SMT transistors
etc that contain the additional grounded base stages.  You don't have
to use exactly two stages, just because I did in the 10816.
You could remove the unneeded components from the 10811
PC board, and then glue the little PC board into the vacated
area.

More stages gives you more reverse isolation, put also more
noise.  Only the collector shot noise process is active.
That's 17.83 pA/root Hz per 1 mA DC, as any time-nut should
be able to rattle off :-)  The noise powers of the N stages
including the first one that is driven by the crystal all add.
You should use transistors that have high fT and high beta at
low DC current, so you don't have to generate any more shot
noise than necessary.  In those days, Motorola made a line of
transistors for pagers that were perfect.  Today you can still
get transistors that are the successors to these pager transistors,
even though pagers are long gone.

I hope that will get you started.  It's not like its rocket
science any more.

Rick N6RK