[Amps] 3CX1500A7/8877 vs el cheapo GS-35B

[Jim Thomson]

From: "Leigh Turner" <invertech at frontierisp.net.au>
Subject: Re: [Amps] 3CX1500A7/8877 vs el cheapo GS-35B

Good comments here Bob; you've nailed some relevant points here.

The GS35B does indeed need a higher plate voltage / plate load impedance to
achieve parity in efficiency with the 8877; complaining about achieving
circa "only" 55% efficiency and a few extra watts of exciter drive is
splitting hairs; particularly in a GG AB2 SSB linear.

At circa $120 the NOS GS-35B represents the best bang for buck in the 1500W
plate dissipation triode tube class; costing a meagre 10% of that of a new
8877 from CPI/Eimac, or 15% of the cost of an excellent Chinese clone. 

Most affordable key RF components for building QRO amps like tubes, vacuum
capacitors, vacuum relays, doorknob capacitors, roller inductors, etc are
all made by non-USA countries. If it weren't for the China tube companies
producing affordable clones of the popular ham-radio tubes and Russian NOS
surplus parts, then readily available tube amplifier components would be a
thing of the past...or the exclusive domain of the wealthy ham.

Cheers,

Leigh
VK5KLT 

##  gee,  I helped my buddy with his 2 x GS35B  6 meter amp..and it does an easy
5 kw out on 50.125 mhz..... with 200 watts of drive from his FT-2000.  With 250 w
of drive, we had to change the 5 kw slug to a 10 kw slug. 

##  eff on 6m was 61%.  The secret was to toss the T network  for the tuned input,
and replace with a PI net.  Consisting of two air variables + coil between em. 
B+  was 4300 vdc  no load and 4100 under load of  2A  (200w  drive)

##  You screw up the tuned input on a GG amp... and ur eff  will suffer badly. 
You actually have to get in there and measure  the power out of the tuned input. 
Increase the tuned input Q  till the po on the C2  side just starts to drop off a few watts. 

##  with too low a Q, eff  goes to hell, real fast like.  With too high a Q, you lose power
in the tuned input itself.   Tank eff is high,  BUT you require MORE drive.

##  The tuned input is more crucial for eff  than the kw PI  or PI-L  on the amps output side.  

##  On my 3cx3000A7 GG amp, with 200w going into the PI  tuned input, we only had
160 watts  coming out of it  on 15m..   The fix was to increase the uh a tiny bit, just  a quarter
of a turn.   Then the po shot up to 195 watts.     To do this test, we had calibrated bird and coaxial
dynamics wattmeters on each side of the tuned input.... then into a 50 ohm dummy load. 
The hb tuned input for the 3cx3000a7  consist of  2 x broadcast air variables  + a tapped 
6 ga coil.  3.8 uh on 160m..then tapped for the other bands.  Each  4 section broadcast cap
is 17-520pf..with all 4 sections in parallel...via strap.   Both caps  padded on 160m only,
with  4 x 500 pf @ 5 kv door knobs.  

##  The GS35b  needs..or likes.... 3900-4500 vdc on it.   And a real tuned input.  Eff well
over 60% on 160-10m.   Forget the 22 ga wire + tuned slugs or T-50’s ,  used on small tuned inputs. 
  With higher Q, the coils will cook.   With low Q, eff drops like a rock.   The Q is so low on a SB-220
tuned input it’s useless. 

##  I like the variable caps on the hb tuned inputs vs relay switched stuff.... or bandswitched pi nets. 
Hams will use too low a  Q  for tuned inputs, cuz they want low swr across the band,.  With
variable caps, you can have ur cake..and eat it too.  Just tweak the 2 x caps on the lower front panel
for flat input swr.  then u also get max grid current.   Run the Q up to 3-5 on the low bands..and 3-4
on the upper hf bands.. and eff remains good.   

Jim  VE7RF