[Amps] Zener Diodes

[Jim Thomson]

Date: Sun, 28 May 2017 16:22:09 -0600
From: MU 4CX250B <4cx250b at miamioh.edu>
To: amps at contesting.com
Subject: Re: [Amps] Zener Diodes

Mike needed this as a replacement cathode bias diode in an old
homebrew 8877 amp with 2600V on the plate.  Manfred's suggestion is
the more contemporary solution, but in any case a 50W zener is way
overkill. A 10W zener would work just fine, and these are still
available without spending a fortune.
73,
Jim w8zr

##  I tried  a 6.3 volt at  10 watt zener.......with the correct heat sink. Peak
plate current is  800 ma.   Lasted 2.5 years b4 it failed..shorted. That was in
a drake L4B amp. 

##  Plate current  X zener voltage =   zener diss.   8 V X  1 amp =
8 watts.  You  need at least a 25 watt zener for the job.    A  50 watt
zener requires a 50 watt heat sink.  Zeners will not handle a  surge.

##  Most reliable bias scheme is a simple series string of  1N5408 diodes..which have 
a 200 A  surge rating.  A  6A10 diode has a 400A surge rating.    You can also parallel
diodes to increase both surge rating and also their  CCS current rating. 

##  1N5408 is good for  3A  CCS.   6A10 is good for 6A  CCS.   With any string of diodes
you can easily vary the bias Voltage, by using more or fewer diodes.   IE:  you can easily 
adjust in aprx  .7 vdc increments.    For some applications, like more bias for  CW or FM, and
less bias for SSB, a  simple  SPDT or  SPDT-center off  miniature toggle switch on the rear apron 
will provide either  2 positions of bias.....or  3 positions if a center off  toggle is used. 

## A 3 A CCS rated diode will diss aprx  .7vdc   X 3A = 2.1 watts  CCS..which will run hot. 
A  6A  CCS diode will diss .7 X 6 = 4.2 watts CCS.   Tested on the bench, I found that a 1N5408
will handle 1 A  CCS  for hrs on end..and run barely warm.   A  6A10  will handle 2A  CCS for hrs
on end..and run barely warm.   Dont use a 1N400X, like a 1N4007.  Surge rating is through the floor,
they run too hot.   5408 and 6a10 are dirt cheap.   

##  if you really want to go crazy, a simple rotary switch will provide all the taps  you want. 

##  For the purists, you can also wire a big electrolytic  across the entire string of diodes.   Like
5000 uf  to  47,000 uf.  A  25-50 vdc lytic works fine.   The bias Vwill  not budge. 
The V across  each diode will increase  by aprx  5-6% when  going from normal
idle current  to  full bore key down plate current.    So as you suck more plate current, the V drop
across  each diode will increase a tiny bit...which then tries to charge the cap up higher. 

##  IMO,  50 watt zeners and mating heat sinks  are way too unreliable, non adjustable,  failure prone,
and way too expensive. 

##  henry radio used strings of diodes in both their 3 k uktra  and also 8 k uktra.     I have used the same 
scheme on everything from  3-500Z to  4-400 to 4-1000,  3x3, and every other GG triode out there..or triode
connected tetrode. .

## On a side note, for CW, you can reduce the idle current by a huge amount, like down to just 20-40 ma, and the
amp will  run just fine..... with way less idle power dissipation and slightly higher  tank eff.  However, you cant  reduce 
the idle current down to zero, on any GG triode amp....... or you will end up with key clix on CW.   25 ma idle
produces way less heat vs 150-250 ma....when on cw mode, vox or qsk. 

###  Zeners have gone the way of the do-do bird. 

Jim   VE7RF