[AMRadio] Gonset G-76

Jim candela jcandela at prodigy.net
Sun Feb 12 10:49:22 EST 2006

See responses below......

-----Original Message-----
From: amradio-bounces at mailman.qth.net
[mailto:amradio-bounces at mailman.qth.net]On Behalf Of Brett gazdzinski
Sent: Saturday, February 11, 2006 6:36 PM
To: 'Discussion of AM Radio'
Subject: RE: [AMRadio] Gonset G-76

I did some work on it today.
First off, I replaced the tube I was using for the audio driver,
I had a 6c4 (triode) in there, some past attempt to clean things up,
and strangely enough, only one pin needed to be moved.
That gave me a little more audio drive.

> The stock audio driver is a 6CM6 beam power tube, and it uses a 9 pin
socket. This tube is identical to a 6AQ5 except the 6AQ5 is a 7 pin tube as
is the 6C4. Did you change sockets back when you added that 6C4? Is the 6C4
really driving the 6DQ6's?

I also increased the RF grid drive by bypassing a 4700 ohm resistor
in the B+ line to the rf driver tube. I found my G76 does an odd thing,
if you go to the tune position and peak the grid drive, when you go to
the high power mode, the peak shifts.
What I did was go to drive level 2 and then tune the grid drive to get
minimum plate current on the output tube, measuring the grid voltage
had that position the highest, then go back to the maximum grid drive
position on the drive level switch.

> I think you will find that this is a neutralization issue, and the NC
setting for best stability seems to vary from band to band. I recall that if
you set the NC for 10 meters first, the other bands are all good enough,
with one exception,  40 meters is pretty far off. I'm not sure why. Bypass
capacitors often go through series resonance somewhere between 5-10 Mhz.
Above resonance they look like a inductor. You might also look at the 6DQ5
grid leak resistor, R12 at 27K 1 watt. I wonder if the best class C
operating conduction angle is being achieved for plate modulation? Since the
6DQ5 has a very high pervience, maybe you can increase R12 along with the
drive level so that the "ON" state of the tube is short in duration << 180
degrees, and the "ON" plate voltage is very low < 100 volts. This might
boost plate circuit efficiency, and at the same time allow for higher
positive peaks with better linearity. Only way to tell though is to modulate
with a triangle wave (speech amp must be pretty flat however), or use a sine
wave, and set scope for a trapezoid pattern (X:Y).

I don't know if its something with my G76 but if the grid drive
is low at all, the positive modulation falls off badly.

> This will vary from 6DQ5 to 6DQ5. What I said above applies here too.

Next was to add some neg feedback from the modulated B+ to
the driver tube grid, bypassing it for highs
to prevent phase shift positive feedback (supersonic).
I tried different amounts and parts values and got the rig
sounding not bad, before the feedback, the highs were extreme,
with very little lows at all.
The amount of feedback acts like a tone control, little
is very bright, more feedback rolls that off and makes things
sound good for a D104.

> This is pretty complicated stuff, and a approach taken by many engineers
that design audio amplifiers. Another approach for the folks that rather not
have their head hurt from such complications is to make the circuitry flat
within the feedback loop (with loop open) within say 50-5000 hz, and adjust
the coupling capacitors within the feedback loop for minimum phase shift
down to 50 hertz. Set the NF gain reduction somewhere between 6-20db
depending on the gain available, and how much NF can be tolerated before
ultrasonic oscillations kick in from high frequency phase shift in the
transformers. Sometimes you can add a Bode pole (R-C) to cut the gain down
beginning at say 10 Khz at the rate of 6 DB / Octave to ward off the
oscillation, and still have lots of NF over the 50-5000 hz range. Once this
is done, then you can define the frequency response in the preamp stages
which are outside the feedback loop. One likely problem with this approach
is that with lots of gain down to the 10's of hertz, the power supplies may
not effectively bypass themselves, and the audio will put ripple on the
power. This can lead to instability in the form of motor boating. Always
look at the power supplies, and evaluate their performance. Stiffening the
supplies may be as simple as boosting the bypass capacitor value 5X (caps
are small & cheap these days), or more elaborate means might be required.

The end result was about 50 to 60 watts output, positive
modulation is up about 80% at 100% neg, if I exceed 100% neg
modulation, the positive will go up to 110%.

> Brett you did it! Now you need to worry about speech polarity, and
negative cycle chopping of the carrier. I added a simple diode -resistor in
series with the modulated B+ to ground. I recall using two 1n4007's in
series (cathode ends towards the B+), and then that in series with something
like a 3.9K 5 watt resistor to ground. This doesn't prevent splatter, or
over modulation, but rather insures that the modulation transformer is
loaded (3.9K) when the negative modulation limit is exceeded.

I checked out the receive section, I had removed the resistors
across the IF cans (narrows the IF bandwidth) and the 5pf caps
bridging from primary to secondary (also narrows the bandwidth).
besides the noisy audio IC chip, the receiver works well.

> You might be able to work on that IC circuit to lower the gain somewhere
in the 3-5 Khz region to cut down on the hiss.

I need to go back and change the B+ source to the audio stages,
and take if off the relay so the tubes are not on all the time,
with a stock rig they do the receive audio, but they
don't on mine, no reason to leave them on....

> Good idea if there are relay contacts available. I did something like this
on my Gonset G50 where the parallel Heising modulator / RCV audio output
tubes (2 X 6L6) would literally cook during prolonged RCV time. I increased
the cathode bias resistor during RCV to about 2X stock value, and then on
XMIT I bridged that resistor with another to get it down to a little less
than the stock value.

I also went through my 6DQ5 and 6DQ6 tube collection, testing
them out, some were bad, some were good...

I did not get a chance to get it on 40 today, I had to go out,
but maybe tomorrow I can work someone with it.

I got on 80 meters tonight, a very rare thing for me, and
the 80 meter antenna (if you can call it that) was acting
crazy from the snow. I got to get rid of that thing and
put up something that works.
Back to the loading coil setup I guess, but in a different spot
not directly over the house...

> Maybe we can work you on 20 meters AM sometime?

Fun with radio...


No virus found in this outgoing message.
Checked by AVG Free Edition.
Version: 7.1.375 / Virus Database: 267.15.6/257 - Release Date: 2/10/2006

More information about the AMRadio mailing list

This page last updated 21 Oct 2017.