[Amps] Shunt resistors for a bridge made of 6A10 diodes

[Jim Thomson]

Date: Tue, 2 Feb 2016 08:39:40 -0800
From: Bill Turner <dezrat at outlook.com>
To: Amps group <amps at contesting.com>
Subject: Re: [Amps] Shunt resistors for a bridge made of 6A10 diodes

Jim wrote:

>  Dont put  ANY  shunt resistors  across a 6A10.  They dont need them! 

REPLY:

I agree with Jim but I would qualify that a little. The 6A10s must be
identical, i.e. from the same manufacturer and same batch. Do not try
to mix and match. The safest way to purchase them is "tape and reel"
so you can be sure they are all the same. 

Unless they are truly identical, the reverse voltage will not divide
equally across each diode and one or more may exceed their PIV rating.
If you can not do this, then by all means use equalizing resistors. 

73, Bill W6WRT

##  I usually buy em from mouser,,,, 200-250 in a box, all with the same lot number.
The issue is not when they are frwd biased, its when they are in the non conduction
mode.  Even then, I have tried 6A10s  from different manufacturers, and different lot numbers,
mixed into a string, on a test jig, for a small FWB, and still see very little variance between 
diodes.  My conclusion is..its  not that big an issue these days, but to be on the safe side,  build
your doubler /or FWB  using diodes from the same bag /box /reel.   Most 6A10s will have a piv
rating of way higher than 1 kv.  They are typ 1250-2400 v, when u test them. 

##  You can get around most of these issues by using enough diodes in the 1st place.  For a 3 kv no load
B+ supply,  you could use  4 per leg in a FWB, (16 in total).   Diodes are cheap, so I use more,  like 9 per leg
for a  3 kv no load FWB.   I see in most commercial B+ diode board assys, they all appear to use  triple the
diodes per leg.  IE:  for say a 8 kv no load supply, they will use  24 x 6A10s for each leg of a FWB, this is
for the single phase version.   They also appear to all be installing a single  MOV  across  each  6A10.   The
MOV will conduct long b4 the PIV is exceeded.  Why they  don’t just install a MOV across the pri of the 
xfmr,  or at the main breaker panel, is beyond me.   You can also get movs rated as high as 1 kv..and they can
be placed across the sec of a xfmr too. Mov’s  can also be paralleled for greater joule rating. 

##  what the commercial board makers are doing with their typ triple the piv idea is...they are concerned about
transients /spikes /surges coming in from the street.   The typ mov wont even begin to  conduct till the mov rating
is at least double.   One would think the HV caps  would absorb any spikes /transients /surges ?   Depending on the
freq of the incoming spike /transient etc, it may well not even make it through the xfmr.  My B+K  875B  LCR meter
measures ac resistance.... vs  dc resistance in the older B+K  875A  version  ( or any fluke dvm , like my 87A). 

##  the 875B version uses a 1 khz tone, square wave....and it wont pass through the dahl xfmr, so I cant  use it to measure
the pri or sec  dc resistance of the xfmr. 

##  KM1H and a few others  will  use a single disc ceramic cap, like a 4700 pf @  5/10 /15 kv variety...across  each leg
of a FWB, so any transient  will simply go through the cap...vs the diodes. 

##  so if we use enough diodes to handle a huge  surge /spike /transient, coming in off the street, like say  double or triple
the normal unloaded B+, then we will, by default have plenty of total PIV to handle  any minor discrepancies  in V drop
across each diode...when in non conduction mode.   A  3 kv no load FWB, with say 6 diodes per leg, will only have 500 V
across each diode.  You would have to have a massive  difference between diodes in the same leg, to exceed the  typ
1250-1400 piv rating of one diode. 

Jim  VE7RF