[Amps] GS35B Amps Common Problem / few more comments

[Jim Thomson]

Date: Thu, 7 Nov 2019 09:01:52 +0900
From: Jeff Blaine <KeepWalking188 at ac0c.com>
To: amps at contesting.com
Subject: Re: [Amps] GS35B Amps Common Problem / few more comments

<The grid on the 3000 is indestructible where the 8877 needs good 
<protection.? On the other hand, the output C of the '3000 is quite a bit 
<higher than the 8877 (about 25 pf vs 10 pf) and for a guy cooking a 
<mulitband amp, it's a factor when it comes to building the output net.? 
<An L-PI will fix it, but I mention that because it's a complicating 
<factor.? Definately not the right choice for a first build.

<73/jeff/ac0c

##  output  C  on  the  3000  is  exactly  24.0 pf  per  eimac. 
I  measured 24.0  pf  on both  my  svetlana 3000s.....on a wooden
bench,  with  my  LCR  meter.    

##  However,  that  increases  to  33  pf....  with  tube in socket.
The extra  9 pf  comes from proximity of lower anode fins ...to  chassis. 

##  a  simple  L-PI  solves that  issue.  Easily  designed on  the  GM3SEK
PI /PI-L spread  sheet.  

##  My  3000  covers 160-15m..... no 10+12M.   I  used  a .66  uh  coil,
made  from 7  turns  of  .375 inch OD  cu  tubing...with  a  1.5 inch ID. Wired
between plate  block  caps....and  C1  tune  cap. 

##  The  33  pf  C  between anode  +  grid....  plus  the  .66 uh  coil,  forms a step
down  L  network  on  15M.   That  reduces the  plate  load  Z...to a much  lower  value.
Low  enough,  that  I  can  use a Q  of  8  on  15M   for the  main   PI  net.   PI  net  is
designed around  the  lower....transformed   plate load Z.  

##  The  15M  tap  on the  main  tank  coil  is   re-used  on  17M.   Q  rises  to  12....on  17M  band. 

##  That  L-PI...... and  also  L-PI-L   trick  has  been  used on  10m,  12m,  6m,  15M.....and
on  tubes  with  a LOT higher   anode to  grid  C.    YC-156,  6000, 10,000, 15,000,  20,000  tubes. 
The worse  case  I designed for was a pair  of 15,000s  which had  120 pf  from anode to  grid. 60 pf
per  tube.... in their  sockets. 

##  RF  parts sells  both  the  socket... and their  newly designed  superb chimney.   Socket  is  good  for  100A....and
also used on  the  6000  tube. 

### Dayton  list  a  small  blower  that is  rated for  80  cfm  at  1.2 inch pressure.   For   4 kw  CCS max  anode  diss,   67  cfm at  1.2  inch
required.  Thats  with an  intake air  temp  of  104  deg  F. 

## For  typ  ssb  +  cw  use...  and  70-75 F   shack  temps...  WAY  less  airflow  required. 
200w  drive  =  5.5  kw  out.    Input Z of  the  3000  is  50  ohms.   PI  tuned input is  simple.  
I  used a pair of  broadcast variables  +  tapped 4 uh  coil.  6:1  jackson  bros  vernier dial drives
on  both  broadcast  caps.   Input  swr is  1:1  from  1.8  to 25  mhz    4 uh  coil  made from
17  Turns of  6 ga  cu  wire....  with a 1.5 inch ID....suspended in  air  between the  caps. Only
2 uh  used  for  80m,  etc.   Again, PI  spreadsheet  can  be  used for PI  tuned inputs.   Tested  with
800w CCS cxr..through  the  hb  tuned  input..then bird..then  DL.   Beware, the  spreadsheet  will  not
handle  2  equal Zs...  so  use  50  ohms on  the  input,  xcvr  side..and  49  ohms  on the output, cathode  side. 

##  Only  grid  protection  I  used  is a fast  3AGC  fuse,  on  rear  panel,  wired  between  chassis..and   negative 
terminal of  grid meter.   IF grid  fuse opens,  no  path  for  DC grid  current,  amp  cant be  driven, PO  drops to  zero.
Input  SWR  rises to  infinity,  xcvr  po  drops to  zero. 

## HV  fuse  in  series with  a 50  ohm  glitch  R....  in  B+.      Grid  diss  is actually  a lot  more when in  GG.  Grid is  
bonded  to  cold chassis..via  finger stock.   6000  also  uses the  same  225 w grid. 

## 3.5  to  6 kv works  good  on a 3000.   Measure the  fil V  via a pair of small  chokes..at the  cathode.   Fil  V  set
for  6.5  vac.   Tube  will  last  forever.

Jim  VE7RF