[Amps] filament transformer

[Jim Thomson]

Date: Wed, 22 Sep 2010 17:27:12 -0700 (PDT)
From: Bob Gibson <w5rg at yahoo.com>
Subject: [Amps] filament transformer


I'm looking for a filament transformer for a 3cx3000 or 3cx6000..;My YC156 tube 
went south and I can't see paying $450 for a used tube when I can buy a new tube 
for just a little more..so if anyone needs a 15v at 15amp filament transformer for 
a Yc156 I have one for sale...110 volt..73s Bob

##  Beware, the 3x3 is  7.5 V  @ 50A .  The  3x6 is 7.0 V  @ 80A . 

##  I wound two identical bifilar's,  [ each with  10 ga magnet wire, on 8" long x 1/2" diam
type 43 rod, SSON].     I then paralleled any one winding from 'Rod A'.... with any one
winding from  'Rod B'   and vice versa.   This  way, you get bifilar action [ opposite current flow]
in each rod.    2 x 10 ga's  will easily handle 60A.    I initially wind the 10 ga polyimide wire
on .5" OD AL tubing.  Then install both AL tunes into amp, bend the stiff leads, etc, and install
6 ga  box lugs, which take 2 x 10 ga wires in parallel  nicely.  Then u can fire up the fil and blower
and test the fil.   When satisfied, AL tubes  are removed,  and the  fragile, brittle   type 43 rods
are slid down the insides  of the winding's.    Never wind the 10 ga on the bare rod itself, it's
brittle enough as is.   I found some scrap .5"  OD  x 8"  AL  tubing   [ old wilson yagi tips]. 
This is for a 3CX-3000A7 tube. 

##  For the 3CX-6000A7 /YU-148 / YC-243 [ socketless 3x6]   I used the exact same procedure,
except that  8 ga magnet wire is used.  2 x  8 ga's in parallel  will easily handle 80A .  We  used
the same  1/2"  Od  x 8"  long type 43 ferrite rods.    Same deal, make 2 x identical bifilars. 

## You would think the uh  would drop in 1/2, with two identical windings in parallel... but it doesn't.
You won't get as many turns with the 8 ga wire,  but it has more then plenty of uh  for 160m.

##  Ok, here's the kicker.  I worked out the maths, 16 ways to sunday...and at the most, there should only
be a .2 v  drop  on the 50A  3x3  fil choke.    I used TWO lab calibrated  fluke 87's... one on the input to the 
fil choke, and one at the socket.   I consistently measure  .4 v drop.  All connections are rock solid.  I checked
teh individual V drop across each connection too, and there is none.  Go figure.   The point here  is... IF  you buy
a  7.5 v  @ 50 A  fil xfmr.... it will be  7.5V   UNDER LOAD of 50A.      Since we lose .4 V in the fil choke.... you
really need a 7.9 v  [ under load]   fil xfmr.     My fil xfmr for the 3x3 amp was a surplus  8.25 vac @ 79A  fil xfmr,
with a 208 vac  single phase input.   No big deal.  I use a small  0-240 vac @ 5 A  variac anyway. 

## same deal with the 3CX-6000A7.  To get  7.0 vac at the socket, you really need  7.4 vac  [ UNDER LOAD]
from the fil xfmr.  On the 3x6, I use a dahl    7.5 vac @  75 A fil xfmr  [ made for a 4x10]   Dahl told me himself,
that you easily suck 80A CCS  from his  75A  rated xfmr's.   Most of the xfmr's  either have taps on the primary,
or folks use a 0-270 vac  variac.  The point is... to expect to lose  .4 vac  across the fil choke. 

## on a new/rebuilt  3x3 or 3x6 ,  we run the fil at RATED voltage,  [ measured at the socket]  for  48 hrs...with
just the blower on.. and NO  B+.    The getter in both tubes  is heat activated from the directly heated cathode. 
48 hrs will  burn off any residual gas in the tube.  Ok, now apply B+  and run the tube... and leave it at rated fil V for
the nest 150 hrs. !!     During the 1st 200 hrs, the emmision  wil be increasing steadily.   After 200 hrs, it hits a peak.
OK, at this point [ 200 hrs on your fil  hr meter],  you reduce the fil V with the variac, in tiny increments,  till the power out
on the PEP meter, [using pulse tuning]...JUST drops off a few watts.    Take note of the exact fil V.   THEN  increase
the  fil v  by exactly .1 V      Ok now you will get an easy 30,000 hrs  out of the tube. 

###  To get around the wild V fluctuations  I was having  for a few yrs,  [from the street, from my neighbour who just got busted
for running a grow op]... I installed a  750 va  Sola  constant v xfmr.  It outputs a  highly regulated   240/120 vac  to the input
of the fil variac.  Output of variac feeds the input of the  fil xfmr.  Sola has it's own dedicated 240 line back to main 200 A panel.   

## To meter the fil V.... we installed  100-200 uh chokes [ 100-200ma], one in each leg..right at the socket.   2 x 22 ga wires 
brought to panel, and terminate on some banna plug female test jacks.  Before the wires terminate to the jacks,  we install a 500-1000 ohm
ohm, 2 w mof resistor is EACH leg..right at the banana jack.  We also install a .01uf cap  from jack to chassis.  OK, if jacks
accidently shorted, nothing happens.  7.5 v / 2000 ohms =  3.75 ma  = .014 w  per resistor.   The pair of resistors  will not affect
the fluke 87-A's  accuracy  either, since it's a high Z  dvm.   The chokes, one per leg, at the socket, will keep the drive power out of
the metering circuit.   The pair of .01uf caps  will kill any residual drive  RF.  [ even with 1 kw drive, it doesn't put a dent on a dvm]   

## if you want to see pix of  all this,  contact me off list.   Either tube is a winner.  50 w of drive = 1.5 kw   100 w =2.5 kw
150w = 3.5 kw    200 w = 5.2 kw   700w = 12.5 kw    And no 8 min warmup required, like on a YC-156.   Both the 3x3 and 3x6 have a 
50 ohm input. 

later... Jim   VE7RF