[Amps] 10 kw CCS on 6M..using the 3CX-6000A7...Part 6.

Sat Oct 1 14:25:04 EDT 2022

<Looks good Jim. That ferrite choke definitely was too much extra C. I
guess you had a hunk of RTV to clean out.
<Question: If the input is so flat, why even have variable caps there? I am
guessing you need to adjust as the tube is not going to be exactly 50 ohms,
and that number is highly dependent on <the cathode current when driven or
idling and a little bit dependent on the cathode to ground rf voltage.

<Are you certain the 0.01 uF ceramic disk caps (3 kV) are not above their
internal resonance and operating as lossy inductors? When I build power
amps I measure each component with an <impedance meter at the RF freq of
operation. The most accurate way is to use a Hp4193A, goes to 110 MHz. They
are hard to find and the probes go for ridiculous amounts. But the older
<analog Hp4815A will work too, and they are cheap if you can find one
working. They are also excellent for testing? Rf chokes.? I think one cap
would be around 0.3 ohms of capacitive <reactance at 6 meters, if it is
below it's own series resonance. If you operate close to series resonance,
then you don't need multiple caps. Another good part to consider is the
large <multilayer ceramic with fat tabs, made for significant RF current.
ATC 100E is one, but they have some other sizes and their datasheet has a
plot of reactance vs freq, so you can see the <series resonance. They have
fat tabs for leads so very low inductance. Once I
 <found that the 820 pF range was perfect short at 200 MHz, and I used one
of them across the filament of an RCA tetrode 4616, 150 KW output, grid
driven. Only one.

<Did you plan on some sort of cathode bias device. Sorry if I missed that
in videos. Keep up the fun work!
<73
<John
<K5PRO

##  I said I would never use fixed caps ever again in a tuned input.  You
only have to be out a few pf, to throw the tuned input..outa whack. Trying
to get a combo of fixed values
 that would = the  air caps would be a daunting task.

## Typ what happens is the tuned input works good with the test fixture,
but the caps are tweaked a bit when in actual usage, with drive applied.
And that depends on the amount of drive,
and also length of coax from IPA to input of 6m tube amp.  Switch IPA to
standby, and then the length  of the coax from xcvr to IPA comes into play.
The bias is adjustable..( tapped string of diodes).  Change the bias, and
more input slight tweaking.  Knowing how finicky tuned inputs are on 10m,
I'm expecting 6m  to be even more tedious.   The C1 cap was new, cheap,
from mfj / ameritron.  Scott bought 2 of em, then we decided to replace the
C2 cap, with a lower value (10-100 pf) type he had on hand....since the
tube plus stay C from the fil chokes was providing bucket loads of...'free
C2'.

##  I also made an error before.   It's 50 pf from EACH fil wire...to
chassis. (bypass caps at cold end temp removed..and ditto with the 6 x
coupling caps at cathode / socket).  With tube plugged in, C from each fil
lug measures 143 pf.  With the tube plugged in, the cathode just bridges
the pair of cables at the socket..and all 3 x stray C's are then in
parallel..and total C was 143 pf.

## with 2 x beads removed, and choke flipped 90 degs..and hanging in mid
air, each wire is 22 pf.  I was expecting a lot less.    With 14 x beads
used on a 160-10m amp, this could be an issue on 10m, but if 12 x used, it
allows for some wiggle room..and 10m will tune.... while providing enough Z
on 160m.

## re: the .01uf caps.  Based on past experience with 4 x previous 6M amps,
it should not be an issue, but  they were not measured, due to lack of test
gear.   We will find out soon enough.
I also made an error when I said  there is RG-400  from RJ1A input relay to
C1 cap. There is no coax. It's a real short, wide piece of copper strap
from the NO contact on the relay to the C1 cap.  Only coax in the video is
from the COM of relay...to input SO-239 jack.  The input RJ1A is
mounted right next to the C1 cap.  Another piece of RG-400  will go from NC
contact...over to the NC contacts of the paralleled RJ2B output relays.
(aka, 'bypass leg').   SFT-600 ( teflon LMR-600)  will go from C2 vac load
cap...to the  NO contacts on the RJ2B's. Then more SFT-600  from  COM of
RJ2B's..to output 7-16 DIN connector.

## almost forgot.  The output of the C2 air variable  cap is strapped
directly to the junction of those 6 x .01uf coupling caps..with a short
piece of wide cu strap..

## one other idea we will try.  Instead of the usual safety choke across
the vac load cap, we will use a 1/4 wave shorted length of SFT-600. That
will do the same job as the conventional safety choke (shorted at far
end).... but also ( maybe) provide for some additional even harmonic
attenuation. It's short in length..and will easily fit inside the RF deck,
coiled up.    Any 1/4 wave stub should really be located a 1/8 wavelength
from the output of the amp...(= 1/4 wave on the 2nd harmonic).   Our
internal shorted stub idea might be a wasted effort. Per K9YC, only time
the 1/4 wave  shorted stub goes at the output of the amp, is when a PI-L is
used.   A 2nd shorted stub could be added outboard..after the
fact..located  1/8 wave from amp.. on 6m.     Between the L-PI output
network..and the shorted stub, and the fact that the 6M ant is not resonant
on it's even harmonics,  it should all work.

## A 2nd RJ1A will be used for cutoff bias...shorting out a 25 k resistor
in the cathode.  ( not installed ..yet).  Also not installed..yet are the
paralleled pair of RJ2B's.   We are going to speed all 4 x vac relays up
with an overvoltage + series drop resistor.  And also a small ( think it
was 50 uf lytic)  across the drop resistor, to speed it up some more.   All
4 x relays are keyed with an opt-isolator (ODC-5A) which has 4 kv isolation
between input and output.  Vdc for the opto will come from rectifying the
7.0 vac at the cold end of the fil choke. ( 2 x grey 22 ga teflon wires).
  Wrong small xfmr showed up that runs the overvoltage supply.  We wanted
dual primaries, and what came was just a single 120 vac primary. Correct
one just showed up yesterday..from Florida no less.

## we also initially tried isolating the front ends of each air variable
from chassis ( ditto with shafts)..and only grounded  the back ends (where
the strap coil is between the caps). With all the trbl we had, isolation
was then removed, and now front + back ends are bonded to chassis...but
shafts still isolated.  Should front ends (closest to front panel) be
isolated or not ?  I'm worried about alternate ground paths on 6m.

## next week the complete output L-PI tank will be mocked up.then tweaked
with test gear. We don't want holes in the front panel, then find out, the
pair of vac caps need greater / lesser spacing.    It's easy to get things
confused, when I'm juggling notes between 4 x different hb amplifiers.
(check out his 4x15 grid driven amp video).

Later.... Jim   VE7RF