[time-nuts] OCXO and fluctuations after EFC adjustment

[Richard (Rick) Karlquist]

On 4/10/2020 12:51 PM, ed breya wrote:

> looking for. Also, moving the frequency far away from "ideal" changes 
> the tempco, since it's no longer at the ideal center of the turnover 
> point. In reality, this may not matter much, since after all these 
> years, things may have drifted and aged way out of ideal-ness anyway.

SC cut crystals:

1.  Do not necessarily have a turnover
2.  Individual oven set points are not precisely at the
turnover if any.
3.  The crystals without a turnover, instead have a broad
range over which their tempco is extremely small

FWIW, the E1938A oscillators were manufactured using
a procedure whereby the oven set points were individually
set for each oscillator to be exactly on the turnover.
IIRC, the crystal blanks were tweaked slightly to assure
that all E1983A crystals had turnovers.  BTW, we used the
UPPER turnover on the E1938A, as opposed to the lower
turnover on the 10811.

> 
> The EFC bias (varactor leakage) current changes too, which interacts 
> with the external driving voltage source impedance. For lowest noise and 
> loading effects, keep the EFC driving resistance as low as possible.

It is easy to see from the 10811 schematic that the cathode of the EFC
varactor is connected to an avalanche/Zener diode of the temperature
compensated type with a conventional diode in series to do the
compensation.  The voltage of the diode (which is something like
6.2V is a "magic" voltage where the tempco goes to zero).  According
to the 10811 designers, this particular model of diode is particularly
well behaved in terms of noise due to the process by which it was made,
at least 40 years ago.

If you want to do anything serious with EFC, you probably need to bring
out both ends of the diode, etc.

At this time, I will give my usual speech about IMHO the fact that
since the invention of the DDS on a chip, EFC should no longer be used 
for high performance oscillators.

Rick N6RK