Robert Casey wrote:

Looks like an envelope detector can perform well. See my post
over in a.b.p.radio with diagrams. Below is the text without
the diagrams:

Unfortunately I cannot see the ABPR post you say you posted.

ABSE seems a more reliable place to post a schematic perhaps.

Patrick Turner.



After reading Brian's web page on modulation acceptance
http://n2.net/k6sti/ma.htm
I decided to try some simulations of the AA5 envelope detector and
got results similar to Brian's page before any mods.
Then I decided to try using that second diode 12AV6's and 12SQ7s
come with. That second diode will do the AVC function, thus
leaving the first diode doing only the audio envelope detection. As
Brian described, the AVC filter cap tends to back bias the envelope
detector in convential AA5 circuits, causing clipping at low valleys of
the modulation. Moving the AVC to its own diode gets rid of that
problem. Also reducing the envelope detector resistance value to 150K
from 500K also helps (you will lose about 3dB of audio, which isn't that
much). And with the AVC now separate this will only mildly affect the
AVC (by about 1dB). And when you move the AVC pick-off cap to
the plate of the IF tube, this drop goes away. This was observed in a real
radio I tried this mod in.

If I can believe my simulation results, I get distortion products nearly
50dB
down at 95% modulation index. IIRC, 95% is an FCC rule for AM stations.

The simulation circuit tried to factor in the loading of the AVC
detector on
the source 455KHz signal on the envelope detector performance. Also
the circuit simulates a little "contact potential" on the vacuum tube
diodes
with the 1V source and 10meg resistor.
To modify an AA5 style AM receiver, unground the extra diode in the 'AV6.
Disconnect the AVC resistor (usually 2.2 to 3.3 megs) from the detected
audio line. Add a 220K resistor across the half meg volume control to
get a load of 150K on the envelope detector. Now connect the free end of
the AVC resistor to the newly freed diode plate. And connect that diode
plate
to ground via a 470K resistor. And from the diode plate to the IF tube
plate
goes a 30pF cap. See diagram. This would cost 2 extra parts in
production,
the 470K resistor and 30pF cap. 'AV6's shipped with the extra diode
already.

This should make for better sound from consoles and bigger table sets. And
AM sections of tube AM/FM receivers where AM was an afterthought. It
probably won't much help that tiny Japanese AA5...