Larry,
I chose the 7591 because I have several, and they are about the same Gm
as
a 6AG7. That means the drive requirement is similar for Ab1 operation. I did
have to increase the G1 bias, and to fill out the AB1 range, I removed the
10K resistor across the 6BA7 mixer tank circuit (it's hidden up top), and
the available drive increased noticeably. The mixer circuit has significant
modifications from CE between the early rigs, and the late 20A's. Mine is
the old version. The newer version has cathode bias on the 6BA7, and
significant mods to the 12BH7 cathode follower. I see why because mine has
reduced output drive to the 7591's as you go up to 10 meters. With the
6AG7's in there, I could get 1/2 watt output on 10 meters, and the 7591's
didn't change that. My VFO has the optional 10 meter circuitry, so drive
from the VFO is fine.
I have had no instability unless I have so much RF in the shack that I
get
RF burns from the D-104 screen. I run mine so far out of the case, and
without the screen above the final cage. No tendency to oscillate, or be
unstable. Both the 7591, and 7868 have two G2 connections at the base. This
allows for redundant G2 bypassing. I am very happy with the changes thus
far. I can now drive my Gonset GSB-201 to 150 watts output on AM, whereas
before 30 watts out is all I could muster when the 20A was stock..
With the 6AG7's on 160, 80, 20, I could get about 12 watts out, or
about 3
watts AM. On 40 meters, the VFO must triple, so drive can be a problem.
There is a CE endorsed mod to the VFO to help out there. With that mod, 40
meters comes in the same. The output on 15 meters is a little squirrelly
(rich in spurs). I haven't perused fixing 15 meters however.
Well back to the 7591. I was actually looking at the 7868 with envy
because
of the very low Cgp, but the Novar socket requirement stopped me. Surplus
sales of Nebraska has a nice ceramic Novar socket that isn't too expensive.
The 7591 is electrically similar, but with Octal socket. You must rewire
from the 6AG7 pinout. It's pretty easy... The 7591 is the same size as a
6V6, so the size is kind of nice. There isn't much room in there, so a fat
tube like a 6DQ6 may be a challenge to remove (if it fits at all!). I also
like the linearity of the Hi-Fi audio tube. The EL-34 you mention is a
possible contender, and it is skinny too. A grid driven sweep tube may not
be very linear, or have enough gain. I don't know though, because I never
tried them.
I currently have my 20A P/S solid stated. The capacitor input pi filter
is
still there. I added a 1K 20 watt resistor, and another 40 uf 450 volt
capacitor following the Pi C-L-C filter, so I know have a C-L-C-R-C filter.
I run all the low level stuff off the 1K feed. This gives me about 300v B+
(key up), and 275 (key down max carrier). The output from the C-L-C is 400v
(key up), and 375v (key down max carrier). The output from the C-L-C Pi also
goes to the 7591 screen grids (G2). Here is where I get crazy. I then take
an external 200v (rack mount Kepco regulated) P/S, and add it to the pi
C-L-C output. This gives me 575 volts for the 7591 plates. This is just
above the max rating for the tube, so no big deal. The 7591's give me about
30 watts carrier output on 160-80 meters at 375 v B+, and about 45 watts at
575 v B+. Since the HV B+ increases some at reduced load, I can get close to
60 watts PEP output from my 20A. That means that 15 watts AM is possible. I
like to keep it below that for pos+ peak headroom. If you add an external
P/S like I did, put a diode across it's output (back biased by the 200V).
That way if the 200V is off, you still get the 375-400V to the 7591 plates.
You don't want to ruin these tubes by running with G2 on, and the plates
off. I adjusted the bias (g1) for about 60 ma idle current. This takes about
18-20 volts of bias to get there.
On the subject of pos + expansion, my 20A, and every other one I have
heard
on AM sounds a little "clicky" whenever the neg- mod peak goes beyond 100%.
Instead of just cutting off the carrier like a plate modulated rig, the
phase reverses 180 degrees, and we make nasty DSBRC. I have found a way
around this. On the rear of the 20A there is a RCA phone jack for audio
output. This same signal is used on AM to drive the upper balanced
modulator. The impedance is low, and the 1n34a diodes are non linear. This
results in asymmetrical modulation where the neg- peak hits 100% when the
mod+ peak is only at about 50%. It looks like [EMAIL PROTECTED] on the scope,
and on the
air whenever you crank the audio just a little bit to get through QRN. Here
is what I do. Take a low drop diode, and put the cathode at the RCA center
pin, and return the anode to ground through a 15 ohm resistor. This seems to
balance the circuit, and sine wave testing shows symmetrical audio up to +/-
80%. Beyond that the mod + goes further to beyond 100% (depends on carrier
output power), and the mod- clips at about 85% without ever hitting the
baseline causing DSB. The difference in your station monitor will be night
and day. You can now compete with the plate modulation rigs. In fact if you
cut the power in half (7 watts in my rig), the dam thing looks like a
ultra-modulated rig on the scope. On my 20A I wired the diode into the mode
switch (there are extra contacts) so that it is only there for AM. I started
out using germanium diodes. I have some 1n270's. Two in parallel did pretty
good, but the diode knee is sloppy. Maybe more in parallel would be better.
I abandoned them when I found some 1n5818's. These are Schotty, and one is
all I need.
There is room to explore this subject further..and I welcome your
comments.
please share them with the group.
Well Larry, I hope this helps!
Regards,
Jim
> -----Original Message-----
> From: Larry Szendrei [mailto:[EMAIL PROTECTED]
> Sent: Tuesday, December 17, 2002 9:24 PM
> To: Jim Candela
> Subject: CE20A mods for QRO
>
>
> Hi Jim,
>
> I read one of your posts on your mods to the CE20A to increase output
> power. I have two 20A's, one which works fine that I'll keep stock, and
> one I need to repair (open primaries on both "modulation transformers")
> that I would like to experiment with to increase the output as you've
> done.
>
> I'm interested as to why you chose the 7591A - I know it's a
> single-ended high transconductance beam power tube capable of some poop,
> but it's also quite rare and expensive. I was thinking of a pair of 6DQ6
> sweep tubes (moderately high Gm around 7000, or so, as I recall, and it
> shouldn't be too difficult to adapt the CE20A for this configuration,
> since the plate tank components are above chassis already), or maybe the
> EL34.
>
> Did you have to add neutralization? If so, how did you do it? I think CE
> could get away without neutralization with the 6AG7's because it is a
> "well screened" pentode, but Ggp is typically high enough to require
> neutralization with the beam power tubes. Do you have any problems with
> instabilities on the higher bands?
>
> Thanks, and I'll be very interested to hear the details.
>
> 73,
> -Larry/NE1S
>
> ---------------------------------------------
> Introducing NetZero Long Distance
> 1st month Free!
> Sign up today at: www.netzerolongdistance.com
