[Amps] Baffling DUAL HV problem ! We need some help.

[Jim Thomson]

Ok,  here's one that has myself and a friend  really baffled. 

We  decided to try and hang  TWO  x FWB's  and  TWO x HV  filter caps... both from the SAME plate xfmr. 

Supply #1  uses  a FWB  [ 10 x 6A10's per each of the 4 x legs]  and a high C  filter.. [ 12 x lytics in series] 

Supply #2     also uses  a  FWB  [ but more  6A10's per leg]   and also a high C filter  [ 20+  x ltyics] 

Dahl plate xfmr  has  sec taps   like  0-2300-2900-3900-4600-5000-5400.

Supply #1   hangs off   the  " 0  and  2900 taps"   

Supply  #2  hangs off the    "0  and   5400 taps" 

RVS connected  6A10's wired between  B-  and chassis.. on both  supplies.

Standard Pri setup, with contactor #1  applying 240 vac.    Contactor  #2  shorts out the step start R, after XXX  seconds.

Good so far.   When testing the supply's ...ONE AT A TIME  [ IE:  fwb on UN-used supply disconnected from  sec]  everything is fine.

## supply #1 =  aprx  4180 vdc          Supply #2 =  aprx 7800 vdc 

 OK, now we tie down BOTH  FWB's  to the sec.     B+  of power  supply #1  is  temp disconnected  from  VHF RF deck.
B-  of  supply #1  IS  connected to  VHF  RF deck.       BOTH  B+  AND  B-  of  supply  #2  IS  connected to HF  RF deck. 

Ok,  both supplies work fine.  B+ meter's everywhere read  what they should.   Plus a  Fluke  0-40 vdc probe available.

Now,  here's the part  when all hell breaks loose.     With  B+ connected to VHF deck... [ at this point, B+  and B-   from each supply, are connected to their
respective VHF /HF  rf decks]   and   B+  switched on... this is what happens.   Voltage as seen on VHF deck  sails  up PAST 5 kv !   This is during the 20 x second  step start delay
sequence.  Before he could hit the off switch on the power supply,   BOTH  the    AC  HV fuse  for  supply #1  AND DC  HV fuse, also for  supply #1...  blow open!
  In both supplies,  a  HV AC fuse is located between sec  and input to FWB  [ one leg only]   A  HV  B+  fuse is located just prior to the respective glitch R. 
Supply #2 isn't even scathed, nothing happens to supply #2  [ no fuses blown]
In addition, the big 240 vac breaker that feeds the input of the plate xfmr, ALSO  trips open. 

To sum up, everything is  fine, UNTIL  the B+  gets plugged into  VHF deck .    We traced this thing 6 x ways to Sunday, and can't  find any back door path  anywhere.
Funny thing though.  With the Fluke HV probe  going... we noticed that  supply #1  showed  4200 vdc, when measuring directly between  B-  and B+ .  However, when probe was between
B+  and CHASSIS... then we measured  5 kvdc !! 

So what are we doing wrong here ???    What am I missing ?   The concept was to obtain two simultaneous, but  different B+  voltages, and not have to buy a 2nd  plate xfmr. 
Only one RF deck was to be  driven,  at any time.  Moot point, since the plate xfmr  can handle the load of both Simultaneously. 

There is NO resistor's  between  B-  and chassis.. on either supply.   [Only rvs connected  6A10 diodes]  Rvs connected diodes are also located in both RF decks, also between  B- and each RF deck chassis.
The rvs connected diodes ensure the B- can never float more than  +/- .8vdc  from chassis potential.  

Individually, EACH  RF deck  works superb.. as long as ONLY the mating B+ supply is hooked up to the sec of plate xfmr.   

Is the  dual supply /one plate xfmr  concept flawed ???     We  really don't want to have to install a 2nd plate xfmr  for  Supply #1.    Plan B was to install vac relays, and wired such that only one FWB is tied to the
plate xfmr sec... at one time.   

The only thing common to both supplies is the chassis.   Also, both  RF deck chassis's are well bonded together.    Only the  '0'  tap of  the plate xfmr is  common to each FWB. 

baffled........ Jim   VE7RF