A lot depends on what the ALC is supposed to be for. In the 'good ole days'
of a pair of 6146s, ALC was supposed to be adjusted so the tubes were on the
verge (or maybe JUST - a few microamps) into grid current. Some rigs
achieved this by having the operator vary the mic gain when tuning the rig.
That gives a carrier suppression dependent on the gain distribution down the
amplification chain. The better ones varied the RF chain gain.
It seems that the majority of today's solid state rigs use ALC as a 'power
setting device', where the 'delay' (if you think of it as an AGC system)
sets the power. Depending then where in the chain the power control is done,
you can get the IF/RF chain running full gain even when the power is turned
down, so the S/N ratio decreases - JUST what you want for transverter
applications! ALC is popular in speech only SSB where it is used to flatten
out variations in frequency response over the HF band. If you want to use
any of the reduced or AM compatible systems, then there are problems because
the carrier pumps up and down in level depending on the amount of
modulation.
Probably the best approach is to fix the mic gain to give the correct S/N
and S/D ratios, set the gain of the IF to get the correct output to the
final mixer and drive stage, and to vary the gain of the driver stage so
that on tune up, the correct output is obtained. To avoid the overshoot etc,
the implementation of the ALC can be done by a suitable switched attenuator
driven from circuitry that switches the attenuator while on tune, and
remembers where it is to be set until the next tune sequence is initiated. I
saw an analogue version of this in the 1960's, where in the tune position
the gain was varied by a motor driven rotary attenuator. You could use an
arrangement where the DC level fed back to the ALC control circuit was
digitised, memorised, and converted back using a DAC to give a fixed DC
level stable with time - that wouldn't spike. But this implementation has
about as much relationship to current types of ALC as aardvarks have to
zygotes. It is in fact an automatic gain set arrangement, rather gain
control.
It is possible to produce a gain control that doesn't overshoot, but it
isn't easy. The GSM standard for cellular requires the transmit power to be
set within a +/-1dB wide window, but allows up to 4dB overshoot - and that
is considered a difficult spec to meet. The US IS137 standard doesn't
specify overshoot, but has a tight reqirement on adjacent channel spurious,
which amounts to the same thing.
Another method that has possibilities with a Class AB1 tube amp is to use a
detector on the grid, with suitable voltage delay so that it starts to
generate an ALC voltage when the grid gets more positive than about -10 or
-20 volts. With that ALC fed back through a loop with a bandwidth of 5 to
10kHz ( which will probably require a balanced RF control stage to prevent
the control input passing transients down the RF chain), and also some
careful control of phase shifts to make sure it was unconditionally stable,
it should work. Will mean that the output power that you can get is a fair
bit less than that available by really driving the tube in AB1 to where Vg
reaches zero volts. Or you could use a PIN diode attenuator at the amplifier
input to prevent the overdrive, but similar problems apply. Note that it
needs to be a constant input impedance attenuator, so as to avoid
mismatching the driver.
As Rich said, ALC is historical. It had its advantages where it was
originally applied, but it's come a long way in application and intent from
there, bringing a different set of problems, not all of which appear to have
been adequately addressed in many rigs.
73
Peter G3RZP
--
FAQ on WWW: http://www.contesting.com/ampsfaq.html
Submissions: amps@contesting.com
Administrative requests: amps-REQUEST@contesting.com
Problems: owner-amps@contesting.com
Search: http://www.contesting.com/km9p/search.htm
|