>From A.R.S. - W5AMI
Sent: Sunday, January 07, 2007 11:38 AM
I'm looking for some ideas on a negative feedback circuit for my
"Bartellsville KW" transmitter, as the one I installed does not seem
to reduce distortion from the 7581A drivers to my liking.
Brian / w5ami
Reply from Jim, WD5JKO,
You seem to be treating the 7581's with negative feedback in order to
lower their output impedance so that you can better respond to the varying
load imposed by the 849 grids.
I see this as one of 3 options on how to approach this:
1.) Use low mu triodes as a driver instead of beam power tubes. You could
triode connect the 7581's, or switch to something like 2A3's or 6B4's.
2.) Use 7581's as beam power tubes, and load the P-P impedance with a
resistor load that is about 1.5X the desired P-P load for the tubes
3.) Do what you are doing now.
Since you are concerned with distortion, and positive peak expansion (2
separate but related issues), I'd look more in depth with stage by stage
behavior of your audio system first. You will likely find that the 7581
driver output gets progressively non linear as the 849 grid swing rises
above zero volts. Lowering the impedance of the driver, or swamping the
driver with a resistive load can help out. I suspect you will need more peak
drive power when the 849 grid swing is well into the positive swing. The
7581's will respond to higher plate B+ and fixed grid bias instead of
cathode bias. You could also use 4 of them in push pull parallel.
I once made a solid state driver for 808 class B modulator tubes. My goal
was to swing the grids to 100 volts + cleanly without any plate voltage. The
absence of plate voltage increases the drive need substantially, and in this
case the peak grid current was 1 ampere at + 100volts. When we swing the
grid high positive, we like the plate voltage to fall below or equal the
grid voltage for that instant of an audio syllable. So in this instance I
needed 100 watts peak audio drive where the RMS drive was less than 10
watts. You will need to beef up that 7581 driver as described above, or use
a solid state / tube hybrid circuit like Geoff W5OMR does with his
modulator.
The positive peak expansion depends on clean modulated B+, and the RF
output tube to modulate linearly. This requires proper drive, bias,
conduction angle, adequate emission, and proper RF load impedance on the
tube. Class C plate modulation is not hard, but to do it linearly up to say
+ 150% requires a lot of attention to details. Sometimes modulating the
driver in phase at 30% is required. You check for this with the Trapezoid
pattern setup to your scope.
The modulation transformer might handle higher peak audio power at a
lower frequency if you get the Final RF amplifier plate current to go
through a modulation reactor instead. I got a FREED 125 watt modulation
transformer to handle 600 watts audio down to 150 hz by getting the final rf
amplifier current to go through 24 Henry, and then capacitance couple the
audio. I used two 6B4's class Ab1 fixed bias with 375v B+ as a driver for
four 808's in push pull parallel with 2250 volts on the plate. This was the
cleanest modulator I ever had. I sure miss the days before the rule change
where you could tune up to 2500 volts @ 400ma with about 850 watts RF
output, and then modulate the final!
Once you got the audio peak power, and good class C plate modulation,
THEN you can sample some of the RF from the rig, detect it, and take that
audio as a source of negative feedback where 6db gain reduction might clean
up some overall non-linearity's. Too much feedback will likely cause
instability due to low frequency phase shift in the transformers.
Regards,
Jim
JKO
--
No virus found in this outgoing message.
Checked by AVG Free Edition.
Version: 7.5.432 / Virus Database: 268.16.7/618 - Release Date: 1/6/2007
7:47 PM
_______________________________________________
Amradio mailing list
List Home: http://w5ami.net/mailman/listinfo/amradio_w5ami.net
Partner Website: http://www.amfone.net
Post: mailto:[email protected]
