[Amps] PIV requirements for identical, individual diodes

[Gene May]

Good question, thanks.  This was done a couple of decades ago, I admit.  I can't remember the part number, but they were rated at 2.5 amp and all the same lot from a major reputable manufacturer.
 
I've heard the situation has improved as you say, but I have gotten mixed lots of other parts in quantity purchases since then and the same problem may exist in such a situation.  The key is not so much how much they leak, but how much the diodes vary in their reverse leakage or "equivalent resistance".  I agree it would be worthwhile to repeat the experiment with current parts, but still, diodes and resistors are cheap. 
 
Gene May
WB8WKU
  



Date: Wed, 25 Apr 2012 17:38:55 -0700
Subject: Re: [Amps] PIV requirements for identical, individual diodes
From: tomc at carneysugai.com
To: gene-may at hotmail.com


Hi Gene

Just curious, when did you do the test with the 200 1K diodes?  My experience is that current production, from the same lot, shows a lot less veriation in leakage current.  

73,

Tom K6EU




On Wed, Apr 25, 2012 at 5:23 PM, Gene May <gene-may at hotmail.com> wrote:




Reference the thread about this:  The PIV across each diode in a bridge will be the peak voltage of the transformer voltage. . . . . , I offer some words of caution:

How the reverse voltage disributes across a chain of diodes is problematic in a way that is similar to the way power supply output voltage distributes across a chain of filter capacitors.  One of the important functions that bleeder resistors serve across such a string of capacitors is to equalize the voltage drops.  Just as electrolytic filter capacitors vary A LOT in how much they "leak" current, meaning how much they vary as resistors across the output voltage, rectifier diodes vary considerably in how much they leak when they are reverse biased.  Once when I had access to the instruments that I could do this with, I got a batch of about 200 1 KV PIV diodes, reverse biased them to about 800V and measured the currents through them.  The currents varied from several microramperes to a mil or so, easily a ratio of 20:1 or more.  This is the reason that you see resistors of about 470K-ohms to about a megohm across a chain of diodes -- to "swamp" the leakage and equalize the voltage drops when they are reverse biased.  I put about a megohm across each diode in a chain or leg of diodes, and use enough diodes to give a total PIV of 5X the output voltage.  Nothing blows.  I concede that I am conservative here, but diodes and resistors are cheap.

Gene May
WB8WKU

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