[time-nuts] noisy varactor diodes

[Alan Melia]

>> The HP sceme in the patent refers to an array of series and parallel
>> diodes to reduce noise.
>>
>> Just paralleling them does not reduce the noise!

Might this not depend what the type of noise was ?? JRC use this technique 
in the RF stage of their receivers but I am wondering if this is only valid 
for Gaussian noise??   the faulty diode looked anything but Gaussian. Is 
there a situation here where we are not consider noise power which should 
add(??) but in the case of the JRC receiver S/S+N ......maybe C/C+N ( 
  :-)) ) i.e it is the effect of the noise on the total capacitance in terms 
of the "FM noise" it produces??

The commonly called Zener diodes will work, all diffused diodes have an 
approximately square-law voltage capacitance relationship. Avalanche diodes 
("zeners" above 6v Vbr :-))  ) are made with big junction areas so can give 
high capacitance. The lower the breadown voltage, the narrower the depletion 
layer so the bigger the static capacitance. However the surfaces might not 
be so well controlled as diodes intended for RF use. The noise demonstrated 
was probably a surface state effect rather than a bulk defect.

Alan
G3NYK