[AMPS] Power Supply Diodes

[measures]

>
>At 07:38 PM 2/2/00 -0700, measures wrote:
>
>> >Did you see the precaution against using so-called equalizer resistors in
>> >the post 1994 Handbooks?  What value of R ans watt rating?.
>> >thanks
>
>Rich, I have seen precautions against using resistors and against NOT using 
>resistors. 

The author provides a semi-explanation of the failure mechanism with 
equalizers.  . 

>I have also seen precautions both ways for shunt caps. I have 
>seen no real proof of any of it. 

We have the knowledge that currents are always equal in a series circuit. 
 

>I have hard proof in the form of hundreds 
>of dead diodes that it is equally easy to blow diodes with any form (or 
>lack thereof) shunt equalization you could think of. In the 40 or so years 
>I've been digging into such things, I have noticed that almost without 
>exception commercial power supplies made from discreet diode stacks have 
>some form of equalization. 

€  Doctors used to recommend eating margerine for health reasons.  
However, all margerines contain hydrogenated oil, which a recent study 
showed was a far more effective plugger-upper of human arteries than the 
stuff from cows and goats.  .  .    

> Those supplies that utilize potted rectifier 
>assemblies don't seem to have any shunt R inside the units. 

They used to until the manufacturers awoke from their slumber.  I still 
have one in my junk collection.  .  

>I assume that 
>to mean that if you can't grade out the diodes, it might be a good idea to 
>at least attempt to equalize the PIV.

Most definitely not.  Reverse current is a major killer.  .  The only way 
to unequalize  existing (naturally equal to start with)  reverse current 
is to connect "equalizer" resistors in parallel with the diodes.  

> I don't care a bit about individual 
>diode reverse current as long as it's well within spec.
>
>The resistors I happened to use this time were 390k 3 watt metal film. 

My guess is that these resistors are rated at 500v maximum.   Who knows, 
there could be a problem with 1000v across 'em?.  .  


>I'm tempted to take you advice, and rewire the assemblies without the 
>resistors. That way I can get twice as many diodes in! If it is/was a PIV 
>related problem that should help, but them I might never know why this 
>failure occurred.

€  I have heard about many such failures occurring during a hv arc.  
Unless a suitable glitch R is used to limit peak fault current, my guess 
is that the problem could happen again, even with more diodes.  //  In 
the 8100v FWD supply I built for my 8169 amplifier, I used 24,  
3a/1000piv diodes that were matched for piv using a high-pot tester.  
Despite a hv arc to gnd, no diode failures have occurred.  (a glitch 
resistor was used to limit fault I)  
>
>
>> >You did not state the manufacturer.
>
>That's correct, I did not. I would have if I had any idea who it was.

The brand is normally marked with white paint.  For example, "GI" means 
General Instruments.  
>
>There is step-start in place and functional. 

ok

>There is glitch protection. 

Is the glitch resistor able to survive the "ugly screwdriver short test" 
without blowing off flakes of glass?  Is the fault current limited to 
200a max?  What value/rating?
thanks
.  
>One more time, the failure DID NOT occur on start-up. It DID NOT occur 
>during a big-bang or other disaster. The whole thing was loafing along at 
>about 1/3 design limits when the damn diodes just rolled over and died. 

I have seen very brief hv arcs to gnd that somehow did not result in a 
big bang.  

>Never before have I experienced diode failure at anything but turn-on or 
>big-bang. I have almost convinced myself that it was just a bad batch of 
>cheap diodes, but yet at the level it was running at, anything should have 
>survived.

If you had a 50% or better piv safety factor, this failure is not easy to 
explain.  
>
>So the question remains, if 3 amp diodes blew with only 700 ma. load, what 
>does it take to keep it running at 2 amps?
>
First off, I would use a high-pot tester to make sure the diodes used are 
confirmed good and roughly matched in actual piv.  .  
>
cheers, Larry

-  Rich..., 805.386.3734, www.vcnet.com/measures.  


--
FAQ on WWW:               http://www.contesting.com/ampsfaq.html
Submissions:              amps@contesting.com
Administrative requests:  amps-REQUEST@contesting.com
Problems:                 owner-amps@contesting.com
Search:                   http://www.contesting.com/km9p/search.htm