Amps
[Top] [All Lists]

Re: [Amps] diode junction temps and thermal resistance

To: knormoyle@surfnetusa.com, AMPS@contesting.com
Subject: Re: [Amps] diode junction temps and thermal resistance
From: TexasRF@aol.com
Date: Wed, 20 Oct 2010 08:10:17 EDT
List-post: <amps@contesting.com">mailto:amps@contesting.com>
Hi Kevin, Bob, W6PO, spent many years in the development lab at  
Eimac/Varian and we enjoy the benefits of his work even today. No doubt he  
learned 
early on about zener bias and failures; hence the use of a 50w rated  device 
instead of an on the edge of destruction 10w device.
 
Yes, some of us have/do use the pass transistor/small zener combo and it  
does work very well. The really great part of this is you can get everything  
needed at your local Radio Shack store in case of a failure and need for a  
speedy repair. Keeping a couple of good sized pass transistors and a 
selection  of small zeners on hand allows construction of a bias circuit for 
most 
any  tube/project.
 
There is an added consideration for some of the tubes we use: the bias  
voltage needed is up in the 20 to 60 volt and even higher range. Using cathode  
bias then involves a dissipation in the 20 to 60 watt range and this is not 
 exactly trivial. Attention to mounting and cooling is a requirement, not 
an  option.
 
Some of the larger tetrodes, like the TH347 we use at 432 and 1296 MHz,  
need bias voltages on the order of 85 to 100vdc. Forget a cathode bias circuit 
 in this case. I use a pass transistor/small zener between the cathode and 
grid  (ground), allowing the B- to rise above ground by the amount of bias  
needed.
 
Since there is no grid current without drive, there would be no bias  
voltage generated in this scheme while the amplifier is in the idle condition.  
To keep it all working, a positive voltage somewhat higher than the needed  
bias is connected through a 1K or so resistor to the B- to keep the bias  
circuit "alive". The pass transistor then has to dissipate the small "keep  
alive" current plus grid current.
 
My TH347 amplifier requires about 95vdc bias and uses this circuit with a  
pass transistor having a voltage adjust pot in the base. a 10W zener 
regulates  the voltage at the pot and is loafing along dissipating maybe a half 
watt. The  zener polarity was chosen to allow it to be bolted right to the 
chassis. The  pass transistor is a PNP configuration, allowing the collector to 
be bolted to  the chassis as well.
 
This same circuit would also work for triodes but the cathode zener is the  
usual choice. The grid metering circuit is slightly more complex but no big 
 problem; it connects between the bias circuit and B- instead of the usual 
ground  to B-.
 
Surge protection can be provided by a single (or multiple) diode from B- to 
 ground as the bias polarity is compatible with this simple circuit. I use  
multiple diodes, at B-, cathode and metering just in case, You know  the 
drill: a belt, suspenders and velcro design approach.
 
Yea, I know this is more information than anyone wants but may be useful to 
 someone.
 
73,
Gerald K5GW
 
 
 
In a message dated 10/20/2010 5:25:43 A.M. Central Daylight Time,  
knormoyle@surfnetusa.com writes:

Good info Ian.

I was looking around trying to learn more about the  various bias circuits 
in use.
maybe people can provide historical views on  some things I was curious 
about:

-This 1971 Eimac paper on a 8877  triode for 144Mhz, had a 1N3311 50 W 12V 
zener for the bias.
Makes me  wonder why R. Sutherland decided on 50W for that  diode.
http://www.landfall.net/Radio/8877-1.htm


-The junction to  case C/W is much better for the zeners in the  larger 
DO-5 stud   package..just 2 C/W max (1.5 C/W typical)
compared to 10 C/W   (typical) for the DO-4 stud package. I wonder if 
that's what's  interesting...not that it can 
tolerate 50W but that it has lower junction  temps to start with, for a 
given DC dissipation..because of the  
better/larger package. Note that it doesn't matter how good your heatsink  
is, if the junction to case thermal resistance 
is the problem.

-The  ability to tolerate power spikes depends on junction temperature rise 
in the  zener, right?. So starting from a low 
junction temp adds  margin?


-rfparts.com has the reverse polarity 1N3306B (50W 7.5V)  for a reasonable 
price, and claims to have the mounting hw. So 
I ordered a  pair. It seems easier to drill larger 1/4" holes and use the 
stud diodes, then  use say the flatter TO-3 
packages and say TO-3 heatsinks or something like  that.  So I ordered a 
pair of those to try.
It appears that some amps  do use the TO-3 packages though?



-The 1971 Eimac paper  mentioned using a smaller zener with a 2N3055 
transistor on a heatsink,  instead of the 50W zener.
I thought that was funny, because some web sites  I've read mention 
something similar, like it's a new idea.
Also saw people  mentioning using zener plus mosfet, at least for audio 
amps...

I  guess you could size a pretty large power transistor? How come we don't 
see  more of these simple lower power 
zener+transistor circuits? or do people  do this?

Ah, I'm reading Ian's Triode Board manual. I guess that's an  example of 
taking that to more  extremes.

-kevin
AD6Z
_______________________________________________
Amps  mailing  list
Amps@contesting.com
http://lists.contesting.com/mailman/listinfo/amps

_______________________________________________
Amps mailing list
Amps@contesting.com
http://lists.contesting.com/mailman/listinfo/amps

<Prev in Thread] Current Thread [Next in Thread>