[Top] [All Lists]

[Amps] Alpha 77D anode choke swap?

To: <>
Subject: [Amps] Alpha 77D anode choke swap?
From: "Jim Thomson" <>
Date: Sat, 28 Apr 2018 12:17:52 -0700
List-post: <>
Date: Tue, 24 Apr 2018 12:32:47 -0500
From: Rob Atkinson <>
To: "" <>
Subject: [Amps] Alpha 77D anode choke swap?

##  You are not after XL.  You are after Z..and there is a big difference.

No pure resistive component to the imaginary number.  Anything else?

## The 50 uh plate choke on the hb  YC-156 amp  is   wound with  20 gauge magnet

It can handle more current.   That doesn't reduce RF voltage to the
power supply.

## Bypassed at the base with a 4700 pf  @ 10 kv disc ceramic capacitor.  Also a
1500 pf
@  15 KV   HEC  doorknob cap.  The  4700 pf cap  does a better bypassing job on
the lower freqs.
The  1500 pf  doorknob does a   better bypassing job on the higher freqs.

If they are paralleled you have 6200 pF.  I don't have time to do the
math but I would still want more inductance.  That much capacitance
might be okay for a linear RF amplifier though.

I find your comment about the disc caps being "better," curious.  With
ceramic capacitors, "better" relative to frequency usually means lower
frequency current rating, and higher current rating means lower or no
value drift.  Disc ceramics don't usually do well on lower
frequencies, especially 160 m.  It makes much more sense, if you need
6200 pF to get that job done with one or more doorknobs rated for a
few amps at 1000 kc.

##   HV filter in the 6700 vdc supply consist of  4 x 47 uf @  4.5 kv oil
filled caps in series parallel.

That's all you need.  Get rid of the electrolytics.  Too much filter
capacitance and stored energy.  Oil much  more reliable.



<The best way to not have to sweat the h.v. power supply plate choke is
<to use a balanced output network with the choke in series with the
<center of the output network coil where the RF v. is already low.



###  easiest way to end this argument is to actually measure the RF current at 
the base  of the plate choke,
using a clamp on  RF ammeter...and also repeat the test on each band the choke 
is being used on. 
A safer method would be to measure with the clamp on RF ammeter, at 
the cold end of each bypass cap in question.   IE:  between chassis and cold 
end of bypass cap. 

## These  ceramite brand   4700 pf  at  15 kv disc ceramics are huge things.   
1.125 inches in diameter,
and   exactly .375 inch  thick.   The 10 kv version is slightly smaller 
diameter, and slightly  thinner.   Trust me, they
will handle one helluva lot of RF  on the lower freqs, like  160M  band.    Typ 
2-3 are used at the base of the plate choke
of any choke thats used on 160M.   Typ 2 used on any choke whose lowest freq 
used is 80M. 

##  Even a plane jane  .01 uf  @  1 kv  disc will handle  1 A or more on  80m.  
A  .01 uf  @ 2 kv disc  ceramic  handles 
a lot more  RF on 80m than I thought it would.   .047 uf  =  4700 pf.  .01 uf = 
10,000 pf    

##  The biggest  value HEC  15 kv door knob  is a HH-57... and   its  1500 pf.  
  Rated for 5.3 A  CCS....  from  100 khz  to 100 mhz. 
I have 6 of em packed away.   I also install some bypass caps right inside the 
HV supply itself. 

##  Those CD  brand  47 uf  @  4.5 kv oil caps are  10 inches tall,  8 inches 
wide, and 4 inches thick.  Steel cased, and weigh exactly 
30 lbs each.   120 lbs   +  crate is a lot of weight to have shipped.  They 
also have to isolated from each other..and chassis, when used
in the series – parallel   config.   One big .25 thick sheet of red micarta 
under all 4 of em. Then .125 thick  sheet of micarta between 
each oil cap.   Those caps originally came from Henry radio, back in the mid 
70s.   I sold the completed  HV supply  to a buddy,
to use with his YC-156 amp.   IMO, oil caps are too heavy for the amount of C  
you get out of them. 

##  No such thing as too much filter capacitance and or  stored energy.  As 
long as a HV fuse +  50 ohm glitch used,  I can 
cro  - bar these  B+  supplies all day long, been there, done that.   XXX  Vdc  
 /   50 ohms  =  the same  value of fault current, regardless
of how much C used.   3-8 kv  will kill you anyway, regardless of total C used. 
 Heck,  the plate xfmr by itself, with no diode board or
HV filter will  also kill you.   They only use  2200 vac for the electric 
chair.  Just use common sense, interlocks, multiple HV meters, etc,
and its all safe. 

##  I have experimented with various values of  filter C.  Aside from  ripple 
being  directly proportional to filter C being used, the dynamic 
regulation when more filter C is used is astounding.   It barely wiggles if 
that.   Loads of filter C also works really good on any doubler setup ,
since the caps only get charged up every 16.6 msecs... instead of the usual 8.3 

##  On these bigger  HV supplies,  we run into a one off unique problem.   The 
V drop  from the street  is  like 6-10 vac.  My  200 amp
service consists of  3-000  Cu from 200 A panel to meter, and more  3-000 CU  
from meter to pot head.   Then its spliced into 1-0 AL..
for the drop wires from the street.  My drop wire terminates on a mid 
span..across the street, then it travels way down the street,   2 x 
poles down, till  it gets to the 50 kva xfmr.  All the  wire out on the street 
between poles is also Al.   AL only conducts  60%  as good as
Cu, then having a drop wire thats   2 x gauges smaller doesnt help regulation.  
 So 9 homes hanging off the same  50 kva xfmr. 

##  That  6-10 vac  sag really mucks things up..esp on a  7-8 kv no load 
supply.  Then toss in the V drop across the 50 ohm glitch R, and
then the normal  drop you get on any HV  supply... and I’m not impressed.   My 
solution was to try to compensate with loads more filter C. 

##  Latest experiment uses  24 x 10,000 uf  @  450 vdc lytics in series.   417 
uf in  total.    I also have another 24 x 10,000  uf caps for
a 2nd bank.  Both banks can be strapped in parallel if required.   Caps were 
dirt cheap  surplus.   3 inches diameter  x 9 inches tall. 
.007  ohm  ESR  per cap.  253 lb Dahl xfmr used.  Ok, now I end up with my 
version of a regulated HV supply, doesnt budge.   

##  who the heck is gonna used a balanced network on the output of a tube amp ? 
  However I have seen it done  on a HF  SW
broadcast TX....which used 2 x tubes . 

###   Getting a plate choke to work is not too tricky.    I may well try the  
VE3s trick of using 2 x chokes, big and small, with cold
ends bonded together, then   toggle between the top ends.  SPDT  vac relay 
wired between  anode and  top of each choke. Then
you can easily optimize the 2 x chokes.  Like say  300 uh for lower bands.... 
and  perhaps   50 uh for  20-10m.

Jim   VE7RF   

Amps mailing list
<Prev in Thread] Current Thread [Next in Thread>