[Alex Pogossov]

"John Byrns" wrote

The claim is that the pull voltage is proportional to the carrier, i.e.
good
for all carrier levels, unlike in Partick's biased detector (with a fixed
pull) which is good only for a certain range of RF carrier.

Are you saying that Partick's biased detector also combats AC/DC loading
distortion?

Sure. Patrick's heavily biased detector combats both slew rate and AC/DC
distortion. But since it is heavily biased it has very low input impedance
at low signal (1.7K, as I showed in part 2). Therefore, Patrick's detector
must have a cathode follower in front.

In this circuit from RDH4 this extra pull needs to be as small as possible
not to reduce significantly the input impedance at low signal, which would
in turn raise the sensitivity threshold and cause the same distortion by a
different machanism.
Yet the pull must be enough to overcome the effect of the volume control AC
coupled. Thus in this circuit the pull is critical. That is why they made R3
a trimpot.


A problem here is that the pull voltage is very-very "averaged" by the
0.1uF
capacitor across R3, and it will not track signal level well. If a signal
varies quickly or if it composed of a composite multi-tone audiom it will
not work perfectly.

It's not clear to me what you are saying here? What difference does the
modulation make as long as we are talking a normal AM signal? When you
talk
about "not tracking the signal level well" what are you talking about?
Normally
the carrier level of a booooring AM signal is constant and doesn't vary,
are you
talking about fading?

You are right. Tracking is not that importamt. Fast fading is virtually
nonexistant for MW broadcasts, especially local stations.


Besides the pull is taken from the primary LC tank of
the IFT, and the IFT transformer ratio is not fixed, but depends on the
accuracy of the radio tuning.

Sounds right, however I would think it would work OK as long as the
compensation
voltage is at least the minimum level required for any given carrier
level.

This circuit increases the effective load on the secondary of the IFT,
effectively lowering the value of the detector load resistor. An
improvement in
modulation acceptance of the ordinary booooring AM detector could
acomplished by
simply lowering the value of the 500k resistor to provide the equivalent
load,
although it wouldn't get us to 100% as this circuit purports to do.

In the 21-st century you would rather simply use an inverting op-amp
which
would be creating a pull, EXACTLY proportional to the instantaneous
signal
level from the same detector. You need only one diode.

If you have an extra OP Amp why go to all this trouble, why not simply use
the
OP Amp to buffer the output of the detector to provide an AC/DC load ratio
of
unity?

You are right. Use an op-amp as an RF buffer and a biased detector. Or
unbiased detector and an AF buffer, or both. I use an emitter mildly biased
detector and an op-amp buffer after it, another op-amp working as an AGC
with infinite gain.

We need to be honest with ourselves -- SS stuff works so much better in
every and any respect. So much more convenience and versatility.

Yet there are some *ppl* who still try build stuff with tubes. These are
like Commmunists, creating problems (hum, noise, drift, low efficiency,
weight, size, distortion, heat, instability, aging, tricky craftmanship,
etc....) and then overcoming the self inflicted problems with enormous
puffing, blowing steam, slogans and chest-pounding....

Steam engines are fun to watch or ride, but we use electric trains,
thanks...

Sorry if I offended anyone.

Regards,
Alex