[SignalOne] ALC investigation

[Ulrich Graf]

Hello to the list,

just to let everybody know how I "messed around" with my Milspec 1030:

Since about half a year I have another Milspec 1030 - I sold my first
one (and regretted it...!). Both rigs were not usable in combination
with a transistor amplifier since my modern pa switched off as soon as
it was driven with a voice signal in ssb. I suspected ALC overshoots
due to a delay in ALC circuitry. This phenomenon should now be
analyzed with the second one. This radio stems from the collection of
a friend. He sometimes hands over a new rig to me for repair or
upgrade. This way I got this Milspec already on my table some time ago
with the defect of beeing unable to adjust maximum transmit power with
R104 (power supply board). I increased the voltage swing of the
detected forward power but still the solution was not satisfying.
Now I wanted to dive somewhat deeper into the circuitry to find the real defect.

1) To "see" the rf signals I connected the transciever's output via a
peak power meter and a directional coupler to a dummy load. An rf
sample from the coupler was fed to a detector and the detected rf to
my oscilloscope. Keying the transmitter in CW with slow and very fast
signals and driving it with quick speech syllables in ssb did NOT show
any ALC overshoots. So rf detection and reaction of the pin-diode
controller on the 2nd mixer board A15 was quick enough to control the
transmitter's output. But still voice power peaks exceeded CW output
power (with same output level adjustment on the front - R6).

2) I do not know if there is another defect in this radio why maximumj
output power control by R104 did not work properly. I could not find
any. Therefor I modified the evaluation of forward power until output
maximum could definitely been controlled by R104. This was merely a
limitation of the created voltage swing from U13A (power supply
board). If someone else experiences some effect like this I will be
pleased to publish the details of this mod. Because I limited maximum
output to 150 W (due to thermal limitation of the heatsink even with
blower) I had to readjust the voltage swing of the reflected power (by
antenna mismatch). I inserted an swr meter and an antenna tuner
between tx output and the dummy load and tuned for swr = 1. Then I
tuned to capacitive, inductive and different reactive loads and
modified R99 (U13B) so that every maximum swr=3 decreases output power
to half.

3) I wondered why an ssb voice signal could give more peak output than
a cw pulse. The reason is that the fast detector in the 1030 detects
only average power, and because cw signals have a constant rf envelope
and voice signal amplitudes differ by 40 dB and more, the voltage to
control output peak power differs even if power control R6 is constant
for both modes. That means that voice peaks were only taken into
account to give a same average voltage as with constant level cw
signals. To avoid overdrive of sensitive power amplifiers with voice
peaks you need to have a peak detector to produce the reference
voltage for maximum output power. My solution was simple: I just
connected a 22 uF electrolytic from the summation point of CR18, CR19
and CR20 to ground. Even with this large value alc was quick enough to
avoid any overshoots. With this mod you may speak or jell or whistle
into the mic - maximum output is always equal to that in cw and which
you have adjusted with R6 on the front.

Long mail - hope you're not bored....

73, Uli, DK4SX