[Alex Pogossov]

There seems to be some interest in an AM detector, so I decided to write
something up on this subject.
We can skip more obvious things related to high level carrier operation,
such as:
- effect of the internal resistance (permeance);
- effect of the RF ripple;
- effect of the diode capacvitance.
By these specs semiconductor diodes outperform tubes.

Rather we focus on the less obvious and more controversial issues like:
1. Sensitivity to low signal;
2. Handling deep modulation;
3. Handling high frequency modulation (slew rate issues).

==== PART 1 =====
Which diode is better?

Common perception is that a silicon diode is "very bad" for small signal
because it has a knee as high as 0.6V. Germanium diode, which conducts at
lower bias of 0.1...0.3V, is better, people say. However, some complain
about leakage of germanium diodes which sometimes kills a detector.

Opinions about vacuum diodes greatly vary. Some say that a vacuum diode
detector is the best for a small signal, since the diode is always slightly
conducting even at no signal because of the thermionic emission. Thus, it
has no "knee" and a detector they say is extremely sensitive. Yet others
blame a tube diode for it reverse biases itself! Hot electrons reaching the
plate deposit negative charge on it which reverse biases the diode
preventing it from demodulating weak signals. A load resistor, which is
rather large (500K) is not sufficient to fully "drain" this emission
current.

Where is the truth?

Well, small signal sensitivity of a detector diode is determined by
sharpness, or curvature of the Volt-Amp characteristic of the diode where
transition occurs from a non-conduction (horizontal flat line) to conduction
(rising curve). Position of the knee is not so important if we assume that
the diode can be biased right to the knee point or to the point of the
maximum curvature. After that, the sharpness of the bend is all that
matters.

Conduction of a diode, whether vacuum or semiconductor, depends on the
electrons overcoming a certain energy barrier, whether a work function of a
cathode or bandgap. Even if the holes are the major carriers, still it is
the electrons which actually carry current. If temperature was low (close to
0K) and the electrons did not have thermal movement, then once the bias
(field) exceeded a certain level, all the electrons would start moving. That
would be an ideal diode with the extra sharp transition from non conduction
to conduction.

In reality the electrons are in thermal motion, some faster, others slower.
With the bias increasing, higher percentage of the electrons, assisted by
their thermal kinetic momentum, are able to overcome the barrier. Thus a
real diode gradually goes fron non-conductance to conductance. The higher
the temperature, the less distinctive the transition.

Physisists have shown that the conduction assosiated with overcoming a
barrier is exponential, approximately:

I = Io * EXP ( eV / kT)

where Io -- some kind of a coefficient;
e -- charge of the electron;
V -- bias voltage;
k -- Boltzmann constant;
T -- absolute temperature in degrees Kelvin.

At room temperature (T = 300K) conduction current in a diode would increase
by e times (approx. 2.7 times) per 25mV bias increase. In other words,
within 25mV of bias change we see nearly 3 times difference between
"forward" and "reverse" conduction. This is enough for a detector to rectify
25mV signal reliably. Thus let us assume that low signal sensitivity of a
semiconductor diode is 25mV.

In the tubes the catode runs at 1200...1300K -- about 4 times hotter.
Obviously, for a tube diode it will take about 100mV bias increase for the
current to increase by 2.7 times. Thus a vacuum diode theoretically 4 times
less sensitive than a p-n diode. Low signal sensitivity of a vacuum diode is
100mV. Some people try to improve the vacuum diode by running its cathode at
lower temperature. It helps, but not that much, because you can not reduce T
to 900...1000K, otherwise the rising internal resistance will make the diode
useless at strong signals and cathode poisoning will increase. Perhaps you
can squeeze 75mV sensitivity of a vacuum tube.

As I said, it does not matter what the VA curve does at large signal. In a
semiconductor diode it remains exponential, and in a vacuum diode it turns
into a law of 1+1/2. Important is that the very beginning, the bend itself
is always exponential, sharpness of which is temperature dependant.

Note that the exponential law applies only to the very beginning of the bend
of a V-A curve which is essential to rectification. At higher currents the
V-A curve will change into the power 1 1/2.

So the conclusion is that a vacuum diode is worse than a semiconductor. Note
that the semiconductor diode needs to be properly biased right to the bend.
This bias is critical and must be temperature compensated. I will cover this
in the third part of the article. Also the issues of source loading have
been ignored. In other words, a perfect cathode follower is assumed.

A grid detector (grid leak detector) has the same sensitivity as the vacuum
diode. Those who are into regenerative radio design -- try to reduce the
cathode temperature as much as practical. So called "anode-bend" and
"infinite impedance cathode" detectors are in theory as sensitive as vacuum
diodes. In practice however, they are much worse. If someone is interested,
I can explain. But now I will just give a hint. Can you expect a good bend
if you are using a remote cut-off tube? Of course, not! There will be no
distinct bend at all. Now consider that practically any tube, even a sharp
cut-off one, has a certain component of "remotecutoffness" because of
inconsistency of its grid pitch.

====================
In the next part I am planning to discuss benefits and drawbacks of biasing
the detector. As an example, for that analysis I will be using a famous
Patrick's detector biased to 30...50V with a weak pull-down.

[NT]

On Jul 14, 11:46*am, "Alex Pogossov" wrote:
There seems to be some interest in an AM detector, so I decided to write
something up on this subject.
We can skip more obvious things related to high level carrier operation,
such as:
- effect of the internal resistance (permeance);
- effect of the RF ripple;
- effect of the diode capacvitance.
By these specs semiconductor diodes outperform tubes.

Rather we focus on the less obvious and more controversial issues like:
1. Sensitivity to low signal;
2. Handling deep modulation;
3. Handling high frequency modulation (slew rate issues).

==== PART 1 =====
Which diode is better?

Common perception is that a silicon diode is "very bad" for small signal
because it has a knee as high as 0.6V. Germanium diode, which conducts at
lower bias of 0.1...0.3V, is better, people say. However, some complain
about leakage of germanium diodes which sometimes kills a detector.

Opinions about vacuum diodes greatly vary. Some say that a vacuum diode
detector is the best for a small signal, since the diode is always slightly
conducting even at no signal because of the thermionic emission. Thus, it
has no "knee" and a detector they say is extremely sensitive. Yet others
blame a tube diode for it reverse biases itself! Hot electrons reaching the
plate deposit negative charge on it which reverse biases the diode
preventing it from demodulating weak signals. A load resistor, which is
rather large (500K) is not sufficient to fully "drain" this emission
current.

Where is the truth?

Well, small signal sensitivity of a detector diode is determined by
sharpness, or curvature of the Volt-Amp characteristic of the diode where
transition occurs from a non-conduction (horizontal flat line) to conduction
(rising curve). Position of the knee is not so important if we assume that
the diode can be biased right to the knee point or to the point of the
maximum curvature. After that, the sharpness of the bend is all that
matters.

Conduction of a diode, whether vacuum or semiconductor, depends on the
electrons overcoming a certain energy barrier, whether a work function of a
cathode or bandgap. Even if the holes are the major carriers, still it is
the electrons which actually carry current. If temperature was low (close to
0K) and the electrons did not have thermal movement, then once the bias
(field) exceeded a certain level, all the electrons would start moving. That
would be an ideal diode with the extra sharp transition from non conduction
to conduction.

In reality the electrons are in thermal motion, some faster, others slower.

[John Byrns]

In article ,
"Alex Pogossov" wrote:

There seems to be some interest in an AM detector, so I decided to write
something up on this subject.
We can skip more obvious things related to high level carrier operation,
such as:
- effect of the internal resistance (permeance);
- effect of the RF ripple;
- effect of the diode capacvitance.
By these specs semiconductor diodes outperform tubes.

Rather we focus on the less obvious and more controversial issues like:
1. Sensitivity to low signal;
2. Handling deep modulation;
3. Handling high frequency modulation (slew rate issues).

Isn't "Sensitivity to low signal" more of a crystal set issue? If we are using
active amplifier devices, why can't we amplify the signal to the level required
for proper operation of the detector?

What is the fidelity of a diodes detection at low signal levels?

==== PART 1 =====
Which diode is better?

Common perception is that a silicon diode is "very bad" for small signal
because it has a knee as high as 0.6V. Germanium diode, which conducts at
lower bias of 0.1...0.3V, is better, people say. However, some complain
about leakage of germanium diodes which sometimes kills a detector.

It would be worthwhile discuss the effects of germanium diode leakage.

Opinions about vacuum diodes greatly vary. Some say that a vacuum diode
detector is the best for a small signal, since the diode is always slightly
conducting even at no signal because of the thermionic emission. Thus, it
has no "knee" and a detector they say is extremely sensitive. Yet others
blame a tube diode for it reverse biases itself! Hot electrons reaching the
plate deposit negative charge on it which reverse biases the diode
preventing it from demodulating weak signals. A load resistor, which is
rather large (500K) is not sufficient to fully "drain" this emission
current.

Where is the truth?

Well, small signal sensitivity of a detector diode is determined by
sharpness, or curvature of the Volt-Amp characteristic of the diode where
transition occurs from a non-conduction (horizontal flat line) to conduction
(rising curve). Position of the knee is not so important if we assume that
the diode can be biased right to the knee point or to the point of the
maximum curvature. After that, the sharpness of the bend is all that
matters.

Conduction of a diode, whether vacuum or semiconductor, depends on the
electrons overcoming a certain energy barrier, whether a work function of a
cathode or bandgap. Even if the holes are the major carriers, still it is
the electrons which actually carry current. If temperature was low (close to
0K) and the electrons did not have thermal movement, then once the bias
(field) exceeded a certain level, all the electrons would start moving. That
would be an ideal diode with the extra sharp transition from non conduction
to conduction.

In reality the electrons are in thermal motion, some faster, others slower.
With the bias increasing, higher percentage of the electrons, assisted by
their thermal kinetic momentum, are able to overcome the barrier. Thus a
real diode gradually goes fron non-conductance to conductance. The higher
the temperature, the less distinctive the transition.

Physisists have shown that the conduction assosiated with overcoming a
barrier is exponential, approximately:

I = Io * EXP ( eV / kT)

where Io -- some kind of a coefficient;
e -- charge of the electron;
V -- bias voltage;
k -- Boltzmann constant;
T -- absolute temperature in degrees Kelvin.

At room temperature (T = 300K) conduction current in a diode would increase
by e times (approx. 2.7 times) per 25mV bias increase. In other words,
within 25mV of bias change we see nearly 3 times difference between
"forward" and "reverse" conduction. This is enough for a detector to rectify
25mV signal reliably. Thus let us assume that low signal sensitivity of a
semiconductor diode is 25mV.

In the tubes the catode runs at 1200...1300K -- about 4 times hotter.
Obviously, for a tube diode it will take about 100mV bias increase for the
current to increase by 2.7 times. Thus a vacuum diode theoretically 4 times
less sensitive than a p-n diode. Low signal sensitivity of a vacuum diode is
100mV. Some people try to improve the vacuum diode by running its cathode at
lower temperature. It helps, but not that much, because you can not reduce T
to 900...1000K, otherwise the rising internal resistance will make the diode
useless at strong signals and cathode poisoning will increase. Perhaps you
can squeeze 75mV sensitivity of a vacuum tube.

As I said, it does not matter what the VA curve does at large signal. In a
semiconductor diode it remains exponential, and in a vacuum diode it turns
into a law of 1+1/2. Important is that the very beginning, the bend itself
is always exponential, sharpness of which is temperature dependant.

Note that the exponential law applies only to the very beginning of the bend
of a V-A curve which is essential to rectification. At higher currents the
V-A curve will change into the power 1 1/2.

So the conclusion is that a vacuum diode is worse than a semiconductor. Note
that the semiconductor diode needs to be properly biased right to the bend.
This bias is critical and must be temperature compensated. I will cover this
in the third part of the article. Also the issues of source loading have
been ignored. In other words, a perfect cathode follower is assumed.

A grid detector (grid leak detector) has the same sensitivity as the vacuum
diode. Those who are into regenerative radio design -- try to reduce the
cathode temperature as much as practical. So called "anode-bend" and
"infinite impedance cathode" detectors are in theory as sensitive as vacuum
diodes. In practice however, they are much worse. If someone is interested,
I can explain. But now I will just give a hint. Can you expect a good bend
if you are using a remote cut-off tube? Of course, not! There will be no
distinct bend at all. Now consider that practically any tube, even a sharp
cut-off one, has a certain component of "remotecutoffness" because of
inconsistency of its grid pitch.

====================
In the next part I am planning to discuss benefits and drawbacks of biasing
the detector. As an example, for that analysis I will be using a famous
Patrick's detector biased to 30...50V with a weak pull-down.

I suspect that the benefit of using bias as in "a famous Patrick's detector" is
that bias reduces or evel eliminates slewing distortion.

--
Regards,

John Byrns

Surf my web pages at, http://fmamradios.com/

[Patrick Turner]

On Jul 15, 5:11*am, NT wrote:
On Jul 14, 11:46*am, "Alex Pogossov" wrote:





There seems to be some interest in an AM detector, so I decided to write
something up on this subject.
We can skip more obvious things related to high level carrier operation,
such as:
- effect of the internal resistance (permeance);
- effect of the RF ripple;
- effect of the diode capacvitance.
By these specs semiconductor diodes outperform tubes.

Rather we focus on the less obvious and more controversial issues like:
1. Sensitivity to low signal;
2. Handling deep modulation;
3. Handling high frequency modulation (slew rate issues).

==== PART 1 =====
Which diode is better?

Common perception is that a silicon diode is "very bad" for small signal
because it has a knee as high as 0.6V. Germanium diode, which conducts at
lower bias of 0.1...0.3V, is better, people say. However, some complain
about leakage of germanium diodes which sometimes kills a detector.

Opinions about vacuum diodes greatly vary. Some say that a vacuum diode
detector is the best for a small signal, since the diode is always slightly
conducting even at no signal because of the thermionic emission. Thus, it
has no "knee" and a detector they say is extremely sensitive. Yet others
blame a tube diode for it reverse biases itself! Hot electrons reaching the
plate deposit negative charge on it which reverse biases the diode
preventing it from demodulating weak signals. A load resistor, which is
rather large (500K) is not sufficient to fully "drain" this emission
current.

Where is the truth?

Well, small signal sensitivity of a detector diode is determined by
sharpness, or curvature of the Volt-Amp characteristic of the diode where
transition occurs from a non-conduction (horizontal flat line) to conduction
(rising curve). Position of the knee is not so important if we assume that
the diode can be biased right to the knee point or to the point of the
maximum curvature. After that, the sharpness of the bend is all that
matters.

Conduction of a diode, whether vacuum or semiconductor, depends on the
electrons overcoming a certain energy barrier, whether a work function of a
cathode or bandgap. Even if the holes are the major carriers, still it is
the electrons which actually carry current. If temperature was low (close to
0K) and the electrons did not have thermal movement, then once the bias
(field) exceeded a certain level, all the electrons would start moving. That
would be an ideal diode with the extra sharp transition from non conduction
to conduction.

In reality the electrons are in thermal motion, some faster, others slower.
With the bias increasing, higher percentage of the electrons, assisted by
their thermal kinetic momentum, are able to overcome the barrier. Thus a
real diode gradually goes fron non-conductance to conductance. The higher
the temperature, the less distinctive the transition.

Physisists have shown that the conduction assosiated with overcoming a
barrier is exponential, approximately:

I = Io * EXP ( eV / kT)

where Io -- some kind of a coefficient;
e -- charge of the electron;
V -- bias voltage;
k -- Boltzmann constant;
T -- absolute temperature in degrees Kelvin.

At room temperature (T = 300K) conduction current in a diode would increase
by e times (approx. 2.7 times) per 25mV bias increase. In other words,
within 25mV of bias change we see nearly 3 times difference between
"forward" and "reverse" conduction. This is enough for a detector to rectify
25mV signal reliably. Thus let us assume that low signal sensitivity of a
semiconductor diode is 25mV.

In the tubes the catode runs at 1200...1300K -- about 4 times hotter.
Obviously, for a tube diode it will take about 100mV bias increase for the
current to increase by 2.7 times. Thus a vacuum diode theoretically 4 times
less sensitive than a p-n diode. Low signal sensitivity of a vacuum diode is
100mV. Some people try to improve the vacuum diode by running its cathode at
lower temperature. It helps, but not that much, because you can not reduce T
to 900...1000K, otherwise the rising internal resistance will make the diode
useless at strong signals and cathode poisoning will increase. Perhaps you
can squeeze 75mV sensitivity of a vacuum tube.

As I said, it does not matter what the VA curve does at large signal. In a
semiconductor diode it remains exponential, and in a vacuum diode it turns
into a law of 1+1/2. Important is that the very beginning, the bend itself
is always exponential, sharpness of which is temperature dependant.

Note that the exponential law applies only to the very beginning of the bend
of a V-A curve which is essential to rectification. At higher currents the
V-A curve will change into the power 1 1/2.

So the conclusion is that a vacuum diode is worse than a semiconductor. Note
that the semiconductor diode needs to be properly biased right to the bend.
This bias is critical and must be temperature compensated. I will cover this
in the third part of the article. Also the issues of source loading have
been ignored. In other words, a perfect cathode follower is assumed.

A grid detector (grid leak detector) has the same sensitivity as the vacuum
diode. Those who are into regenerative radio design -- try to reduce the
cathode temperature as much as practical. So called "anode-bend" and
"infinite impedance cathode" detectors are in theory as sensitive as vacuum
diodes. In practice however, they are much worse. If someone is interested,
I can explain. But now I will just give a hint. Can you expect a good bend
if you are using a remote cut-off tube? Of course, not! There will be no
distinct bend at all. Now consider that practically any tube, even a sharp
cut-off one, has a certain component of "remotecutoffness" because of
inconsistency of its grid pitch.

====================
In the next part I am planning to discuss benefits and drawbacks of biasing
the detector. As an example, for that analysis I will be using a famous
Patrick's detector biased to 30...50V with a weak pull-down.

Maybe one day we'll use relays, they may be fast enough on the
nanometer scale

NT- Hide quoted text -

- Show quoted text -

Well, ya could just amplify the whole antenna signal say +20dB, then
digitise the whole darn lot and apply algorithyms in a chip which has
a schematic that covers the dining room table, and hey presto, out
comes any signal you'd ever want!

Its been done for years now, virtual radios on the screen.

There are many ways to skin a fox.

Patrick Turner.

[Patrick Turner]

On Jul 15, 7:58*am, John Byrns wrote:
In article ,
*"Alex Pogossov" wrote:

There seems to be some interest in an AM detector, so I decided to write
something up on this subject.
We can skip more obvious things related to high level carrier operation,
such as:
- effect of the internal resistance (permeance);
- effect of the RF ripple;
- effect of the diode capacvitance.
By these specs semiconductor diodes outperform tubes.

Rather we focus on the less obvious and more controversial issues like:
1. Sensitivity to low signal;
2. Handling deep modulation;
3. Handling high frequency modulation (slew rate issues).

Isn't "Sensitivity to low signal" more of a crystal set issue? *If we are using
active amplifier devices, why can't we amplify the signal to the level required
for proper operation of the detector?

What is the fidelity of a diodes detection at low signal levels?

==== PART 1 =====
Which diode is better?

Common perception is that a silicon diode is "very bad" for small signal
because it has a knee as high as 0.6V. Germanium diode, which conducts at
lower bias of 0.1...0.3V, is better, people say. However, some complain
about leakage of germanium diodes which sometimes kills a detector.

It would be worthwhile discuss the effects of germanium diode leakage.

Our dear Alex sees problems with diode use where the circuit impedance
is high, as it is in a typical tube radio IFT secondary circuit. My
solution is to ALWAYS convert this high Z to low Z with a CF **if
possible** but it not always possible if there's no room to drop in a
12AU7. But a darlington pair of BJT can be used, high base input Z,
low emitter output Z. Ic should be about 4mA. And the transistors can
work off a 50Vdc supply with the IFT coil biased at say +25V. Then my
type of detector can be used so you have an SS signal diode with say
0.2mA of idle current flow via 125k from diode cathode to 0V. Thus the
diode is always turned on with a slight forward voltage. The C used
has to be selected so 455kHz ripple voltage is kept low as possible,
but allowing say 8kHz of AF at full Vo of say 5Vrms to be free of slew
distortion. In practice the amount of HF above 5kHz is very small
compared to LF made at say 200Hz. The lower the Vripple, the less
distortion one gets as one increases modulation %. The variation in
current discharging from C after each 455kHz charge cycle is
negligible for all levels of modulation until one gets close to 100%
mod. Please draw the wave forms, observe the wave forms on the CRO and
ALL will become clear !!!

There will always be some 455kHz ripple on the C and thus one might
add an RC filter, say 100k plus 50pF which filters this away. Then a
following grid or base of a follower type circuit can be connected to
send the audio out on its merry way without being affected by the
dreaded cut off distortion caused by RC coupled loads after the usual
feeble highZ detecor circuits found in most crummy old radio sets.

snip
====================
In the next part I am planning to discuss benefits and drawbacks of biasing
the detector. As an example, for that analysis I will be using a famous
Patrick's detector biased to 30...50V with a weak pull-down.

I suspect that the benefit of using bias as in "a famous Patrick's detector" is
that bias reduces or evel eliminates slewing distortion.

My detector circuit is simple and easy for anyone to reproduce and
fine tune, if only they would, some time during the rest of this
century.

Meanwhile they'll sit and cogitate and mumble forwards/backwards and
produce no working circuits, or research backed schematics.

Patrick Turner.



--
Regards,

John Byrns

Surf my web pages at, *http://fmamradios.com/- Hide quoted text -

- Show quoted text -

[John Byrns]

In article ,
Patrick Turner wrote:

Our dear Alex sees problems with diode use where the circuit impedance
is high, as it is in a typical tube radio IFT secondary circuit. My
solution is to ALWAYS convert this high Z to low Z with a CF **if
possible** but it not always possible if there's no room to drop in a
12AU7. But a darlington pair of BJT can be used, high base input Z,
low emitter output Z. Ic should be about 4mA. And the transistors can
work off a 50Vdc supply with the IFT coil biased at say +25V. Then my
type of detector can be used so you have an SS signal diode with say
0.2mA of idle current flow via 125k from diode cathode to 0V. Thus the
diode is always turned on with a slight forward voltage.

I still don't understand this "diode is always turned on" thing, how does the
diode detect if it is always "turned on"?

The C used
has to be selected so 455kHz ripple voltage is kept low as possible,
but allowing say 8kHz of AF at full Vo of say 5Vrms to be free of slew
distortion.

It looks like the Turner Audio "Bean Counters" got the best of you here. If you
were really interested in doing the best detector, you would use a full wave
detector which would increase the ripple frequency to 910 kHz, easing the
compromise between keeping the ripple voltage low and the detector free of
slewing distortion.

In practice the amount of HF above 5kHz is very small
compared to LF made at say 200Hz. The lower the Vripple, the less
distortion one gets as one increases modulation %. The variation in
current discharging from C after each 455kHz charge cycle is
negligible for all levels of modulation until one gets close to 100%
mod. Please draw the wave forms, observe the wave forms on the CRO and
ALL will become clear !!!

There will always be some 455kHz ripple on the C and thus one might
add an RC filter, say 100k plus 50pF which filters this away. Then a
following grid or base of a follower type circuit can be connected to
send the audio out on its merry way without being affected by the
dreaded cut off distortion caused by RC coupled loads

Instead of a simple RC, why not use the follower as a second order active low
pass filter?

--
Regards,

John Byrns

Surf my web pages at, http://fmamradios.com/

[Patrick Turner]

On Jul 16, 9:31*am, John Byrns wrote:
In article ,
*Patrick Turner wrote:

Our dear Alex sees problems with diode use where the circuit impedance
is high, as it is in a typical tube radio IFT secondary circuit. My
solution is to ALWAYS convert this high Z to low Z with a CF **if
possible** but it not always possible if there's no room to drop in a
12AU7. But a darlington pair of BJT can be used, high base input Z,
low emitter output Z. Ic should be about 4mA. And the transistors can
work off a 50Vdc supply with the IFT coil biased at say +25V. Then my
type of detector can be used so you have an SS signal diode with say
0.2mA of idle current flow via 125k from diode cathode to 0V. Thus the
diode is always turned on with a slight forward voltage.

I still don't understand this "diode is always turned on" thing, how does the
diode detect if it is always "turned on"?

The Si diode has a small DC current flow which means that it is
"turned on", but only very slightly, and there is minimal variation of
voltage across the diode at any time. The DC flow through the R which
discharges the charge in C is substantially constant. If you have the
IFT secondary biased at say +40Vdc, the CF grid is also at 40Vdc, and
without any 455kHz signal the CF cathode might be at 46Vdc. Now,
there's +45.8Vdc from cap to 0V, and if the R after C is say 680k, the
idle current through diode = 48.8 / 680k = 0.072mAdc.
Let us suppose a 455kHz carrier appears with 2Vpk value, then the V
across C will rise by 2Vdc less the slight loss due to 455kHz ripple
voltage. This ripple V might be only 0.1Vpk, and you might DRAW THE
WAVEFORMS you might expect to see, and then have a look at your sample
circuit with CRO and make the appropriate measurements to see how well
you drew the waveforms, thus educating yourself without having to ask
me about how my detector circuit works.
Now, let us suppose we modulate the 2pk V carrier with 100% mod. The V
pk-pk of AF = 4V, and when you examine with your CRO, the ripple
voltage hardly changes at all at any part in the AF wave because the
slight pk to pk AF wave change does not substantially the current
discharging the cap after it has been charged by the CF and diode. So
the recovered AF wave is remarkably free of THD except for the
negative crests of the AF wave. If the ripple is kept low, but no so
low to cause slew rate Dn at high AF, then you should be able to crank
up the Mod % to 95% before any AF THD appears. I leave you to observe,
measure, analyse, and make your own conclusions.

My detector works best for strong signals worth listening to. Short
wave signals are perhaps extremely low at the detector, and their
strength varies considerably, but because the noise is so darn high
with feint SW stations then the detector Dn may not matter. For good
SW with tubes one needs a good antenna, and a tuned RF stage ahead of
the mixer and IF amp. Many sets I have for repair have SW coils but
no RF stage, and rely solely on night listening with a good antenna.
And those with the added RF stage often don't track very well, and one
gets the idea that with many AM sets the SW was included as a sales
gimick and not something offering any more than novelty value.
Headphones might improve matters, but the best solution is to have a
good RF stage. This needs gang tuning caps. In budget sets, there is
no room for the RF amp and its far too difficult to replace a 2 gang
tuning C.

But there have been some remarkable reflex sets with a little PFB re-
gen to liven up the SW reception. These are all prone to whistles and
crap. PFB and reflexing where the IF amp also acts as the first AF amp
was always cheaper to make than a normal set, because one whole tube
could be omitted. Every possible trick was tried to reduce parts count
in old radios. I have a reasonable reflex set made in Oz which a
customer failed to pick up after I spent 4 days ****ing around with it
to get it to sound well. I only charged $100, but that's enough for
customers to abandon their enthusiam to repair grandad's old junk.
This concoction has global shunt NFB applied from OPT sec speaker
connection to the detector circuit with the volume control as part of
the FB network. This works much better than having no NFB at all, as
it tends to correct the dreadful Dn of the AF output stage.

The C used
has to be selected so 455kHz ripple voltage is kept low as possible,
but allowing say 8kHz of AF at full Vo of say 5Vrms to be free of slew
distortion.

It looks like the Turner Audio "Bean Counters" got the best of you here. *If you
were really interested in doing the best detector, you would use a full wave
detector which would increase the ripple frequency to 910 kHz, easing the
compromise between keeping the ripple voltage low and the detector free of
slewing distortion.

I tried full wave, but not so easy at all. I've tried a 2 diode
voltage doubler, also not worth the slight increase in Vo.

There is utterly no need for any more than what I have, IMHO.

In practice the amount of HF above 5kHz is very small
compared to LF made at say 200Hz. The lower the Vripple, the less
distortion one gets as one increases modulation %. *The variation in
current discharging from C after each 455kHz charge cycle is
negligible for all levels of modulation until one gets close to 100%
mod. Please draw the wave forms, observe the wave forms on the CRO and
ALL will become clear !!!

There will always be some 455kHz ripple on the C and thus one might
add an RC filter, say 100k plus 50pF which filters this away. Then a
following grid or base of a follower type circuit can be connected to
send the audio out on its merry way without being affected by the
dreaded cut off distortion caused by RC coupled loads

Instead of a simple RC, why not use the follower as a second order active low
pass filter?

One could, but seems like there's no need with IF = 455kHz.

But with a FM radio where the sub-carrier is an abysmally pathetic low
38kHz ( should have been 76kHz ),
one has to worry about removing 19kHz and 38kHz remnants and switching
noise and then a second order filter and perhaps a 19kHz notch filter
is wise. A radio set should not be able to amplify 455kHz anywhere
after the AM detector, and if there is a small amount of 455kHz ripple
applied to the AF amp, it should not matter at all because tubes will
cope, without detectable Dn created.

Patrick Turner.

--
Regards,

John Byrns

Surf my web pages at, *http://fmamradios.com/

[Don Pearce[_3_]]

On Sun, 17 Jul 2011 04:14:48 -0700 (PDT), Patrick Turner
wrote:

I tried full wave, but not so easy at all. I've tried a 2 diode
voltage doubler, also not worth the slight increase in Vo.

There is utterly no need for any more than what I have, IMHO.

Have you tried a synchronous detector? Generate a well-limited square
wave at the IF frequency and use it to switch a full wave diode
bridge. This will recover - with perfect linearity - any AM signal,
even one modulated well beyond 100% (provided it has been carried out
properly).

d

[Alex Pogossov]

"Don Pearce" wrote in message
...
On Sun, 17 Jul 2011 04:14:48 -0700 (PDT), Patrick Turner
wrote:

Have you tried a synchronous detector? Generate a well-limited square
wave at the IF frequency and use it to switch a full wave diode
bridge. This will recover - with perfect linearity - any AM signal,
even one modulated well beyond 100% (provided it has been carried out
properly).

Alex:
I have with success. In fact I fit a synchronous PLL based detector module
into old radios. It allows to detune the radio so that a carrier sits on the
slope of the IF responce curve, suppressed and working only as a pilot tone,
so to speak. Demodulation is performed by CMOS switches. Virtually only one
sideband is used. This sideband now almost completely falls into a 6...7kHz
IF bandwidth. The sound is amazing -- like on FM, and even better because
there is no this "screeching" sometimes caused by multipath propagation. By
detuning the radio one way or another you can choose a sideband less jammed
by interference.

On crowded SW it is still better not to detune the radio, but uave the
carrier in the middle. It gives more loud reception of weaker stations.
Resilience to fading is great. SW becomes listenable. For fun I listen to
"Radio Russia" in Russian and also to Chinese propaganda intended for Russia
(also in Russian).

Unfortunately for the tubefanatics on this list, this synchro module CAN NOT
BE BUILT using TUBES. It uses ICs: TLE2704, 74HC74, 74HC4046, MCP602, LM339,
LM78M05 and several p-JFETs in auto PLL lock band control.

[Patrick Turner]

On Jul 17, 9:27*pm, (Don Pearce) wrote:
On Sun, 17 Jul 2011 04:14:48 -0700 (PDT), Patrick Turner

wrote:
I tried full wave, but not so easy at all. I've tried a 2 diode
voltage doubler, also not worth the slight increase in Vo.

There is utterly no need for any more than what I have, IMHO.

Have you tried a synchronous detector? Generate a well-limited square
wave at the IF frequency and use it to switch a full wave diode
bridge. This will recover - with perfect linearity - any AM signal,
even one modulated well beyond 100% (provided it has been carried out
properly).

d

I'd have to build a 455kHz square wave oscillator, and its all a lot
of tubes and work and I already get excellent AM reception I'm happy
with. What you say is easier said than done. If you don't agree,
please post fully worked out schematics somewhere at your website to
prove I am uneducated. Ball's in your court now.

Patrick Turner.

[Alex Pogossov]

"Alex Pogossov" wrote in message
...

"Don Pearce" wrote in message
...
On Sun, 17 Jul 2011 04:14:48 -0700 (PDT), Patrick Turner
wrote:

Have you tried a synchronous detector? Generate a well-limited square
wave at the IF frequency and use it to switch a full wave diode
bridge. This will recover - with perfect linearity - any AM signal,
even one modulated well beyond 100% (provided it has been carried out
properly).

Alex:
I have with success. In fact I fit a synchronous PLL based detector module
into old radios. It allows to detune the radio so that a carrier sits on
the slope of the IF responce curve, suppressed and working only as a pilot
tone, so to speak. Demodulation is performed by CMOS switches. Virtually
only one sideband is used. This sideband now almost completely falls into
a 6...7kHz IF bandwidth. The sound is amazing -- like on FM, and even
better because there is no this "screeching" sometimes caused by multipath
propagation. By detuning the radio one way or another you can choose a
sideband less jammed by interference.

On crowded SW it is still better not to detune the radio, but uave the
carrier in the middle. It gives more loud reception of weaker stations.
Resilience to fading is great. SW becomes listenable. For fun I listen to
"Radio Russia" in Russian and also to Chinese propaganda intended for
Russia (also in Russian).

Unfortunately for the tubefanatics on this list, this synchro module CAN
NOT BE BUILT using TUBES. It uses ICs: TLE2704, 74HC74, 74HC4046, MCP602,
LM339, LM78M05 and several p-JFETs in auto PLL lock band control.

I forgot the most important -- the switches DG445. Two are used for
quadrature demodulation, and the other two for automatic switching between
synchro and ordinary AM when PLL unlocks. To switch this module on and off I
use a pot with shaft push-pull switch (like in the guitar stuff). I fit it
instead of the tone control. Thus the "original" look of the radio is
preserved, but synchro detector is added.

Also forgot to mention that this synchro module has a delayed amplified AGC
(integrator) which equalises the levels of all the stations, no matter weak
or strong. Apart from anything else, it is important for having PLL
bandwidth stable.

In fact I strip the crappy booooring AM detector / AGC stuff from a
boooooring radio (what sits around 6AV6) and replace it with the module.

[Don Pearce[_3_]]

On Sun, 17 Jul 2011 05:15:19 -0700 (PDT), Patrick Turner
wrote:

On Jul 17, 9:27*pm, (Don Pearce) wrote:
On Sun, 17 Jul 2011 04:14:48 -0700 (PDT), Patrick Turner

wrote:
I tried full wave, but not so easy at all. I've tried a 2 diode
voltage doubler, also not worth the slight increase in Vo.

There is utterly no need for any more than what I have, IMHO.

Have you tried a synchronous detector? Generate a well-limited square
wave at the IF frequency and use it to switch a full wave diode
bridge. This will recover - with perfect linearity - any AM signal,
even one modulated well beyond 100% (provided it has been carried out
properly).

d

I'd have to build a 455kHz square wave oscillator, and its all a lot
of tubes and work and I already get excellent AM reception I'm happy
with. What you say is easier said than done. If you don't agree,
please post fully worked out schematics somewhere at your website to
prove I am uneducated. Ball's in your court now.

I designed a synchronous AM detector back in the days when I was at TV
designer for the Rank Organization. I have no intention of doing it
again because AM radio is by now a busted flush. And of course you
don't use an oscillator, you just apply severe limiting to the IF
signal to strip off the modulation and leave a square wave.

d

[John Byrns]

In article ,
(Don Pearce) wrote:

On Sun, 17 Jul 2011 04:14:48 -0700 (PDT), Patrick Turner
wrote:

I tried full wave, but not so easy at all. I've tried a 2 diode
voltage doubler, also not worth the slight increase in Vo.

There is utterly no need for any more than what I have, IMHO.

Have you tried a synchronous detector? Generate a well-limited square
wave at the IF frequency and use it to switch a full wave diode
bridge. This will recover - with perfect linearity - any AM signal,
even one modulated well beyond 100% (provided it has been carried out
properly).

Neglecting the problems with the limiter around the zero carrier point, your
scheme doesn't work for signals "modulated well beyond 100%" in the negative
direction. The problem is that when modulation exceeds 100% in the negative
direction the phase of the carrier flips causing the switching of the diode
bridge to be out of phase with the original carrier, causing serious distortion
in the receiver.

Adding a phase detector and a VCO to your circuit will fix this problem.

Extra credit quiz question, what tube era broadcast transmitter could generate
negative modulation peaks beyond 100% if the negative peak clipper in the
transmitter was disabled?

--
Regards,

John Byrns

Surf my web pages at, http://fmamradios.com/

[Alex Pogossov]

"John Byrns" wrote in message
...
In article ,
(Don Pearce) wrote:

On Sun, 17 Jul 2011 04:14:48 -0700 (PDT), Patrick Turner
wrote:

I tried full wave, but not so easy at all. I've tried a 2 diode
voltage doubler, also not worth the slight increase in Vo.

There is utterly no need for any more than what I have, IMHO.

Have you tried a synchronous detector? Generate a well-limited square
wave at the IF frequency and use it to switch a full wave diode
bridge. This will recover - with perfect linearity - any AM signal,
even one modulated well beyond 100% (provided it has been carried out
properly).

Neglecting the problems with the limiter around the zero carrier point,
your
scheme doesn't work for signals "modulated well beyond 100%" in the
negative
direction. The problem is that when modulation exceeds 100% in the
negative
direction the phase of the carrier flips causing the switching of the
diode
bridge to be out of phase with the original carrier, causing serious
distortion
in the receiver.

Adding a phase detector and a VCO to your circuit will fix this problem.

Absolutely correct! One can make a detector consisting of a high-speed
comparator (MAX941 is a good choice) and a CMOS switch (74HC4066 or even
74LVC1G66), and it would work down to very low levels of signal, to about
2mV, but it will never "properly" demodulate suppressed carrier signals with
modulation 100%. Phase reversals in the signal immediately trannslate into
phase reversals on the comparator output.

Extra credit quiz question, what tube era broadcast transmitter could
generate
negative modulation peaks beyond 100% if the negative peak clipper in the
transmitter was disabled?

Alex:
No transmitter will do that, but a carrier might lose its amplitude due to:
- fading;
- multi-path propagation (often in the vicinity of some large objects like
street wiring, transmission lines, etc.);
- double peak tuning of the IFT.


--
Regards,

John Byrns

Surf my web pages at, http://fmamradios.com/

[John Byrns]

In article ,
"Alex Pogossov" wrote:

"John Byrns" wrote in message
...
In article ,
(Don Pearce) wrote:

On Sun, 17 Jul 2011 04:14:48 -0700 (PDT), Patrick Turner
wrote:

I tried full wave, but not so easy at all. I've tried a 2 diode
voltage doubler, also not worth the slight increase in Vo.

There is utterly no need for any more than what I have, IMHO.

Have you tried a synchronous detector? Generate a well-limited square
wave at the IF frequency and use it to switch a full wave diode
bridge. This will recover - with perfect linearity - any AM signal,
even one modulated well beyond 100% (provided it has been carried out
properly).

Neglecting the problems with the limiter around the zero carrier point,
your
scheme doesn't work for signals "modulated well beyond 100%" in the
negative
direction. The problem is that when modulation exceeds 100% in the
negative
direction the phase of the carrier flips causing the switching of the
diode
bridge to be out of phase with the original carrier, causing serious
distortion
in the receiver.

Adding a phase detector and a VCO to your circuit will fix this problem.

Absolutely correct! One can make a detector consisting of a high-speed
comparator (MAX941 is a good choice) and a CMOS switch (74HC4066 or even
74LVC1G66), and it would work down to very low levels of signal, to about
2mV, but it will never "properly" demodulate suppressed carrier signals with
modulation 100%. Phase reversals in the signal immediately trannslate into
phase reversals on the comparator output.

Extra credit quiz question, what tube era broadcast transmitter could
generate
negative modulation peaks beyond 100% if the negative peak clipper in the
transmitter was disabled?

Alex:
No transmitter will do that, but a carrier might lose its amplitude due to:
- fading;
- multi-path propagation (often in the vicinity of some large objects like
street wiring, transmission lines, etc.);
- double peak tuning of the IFT.

John:
Actually there is at least one type of AM transmitter that will do this, the
outphasing transmitters, a great French design from the 1930s. The RCA in the
USA popularized this transmitter design in the 1950s through the 1970s.

The famous British Radio Caroline station used one of these transmitters.

Descriptions of the Ampliphase principle can be found on these web pages.

http://www.rossrevenge.co.uk/tx/ampli.htm

http://www.rossrevenge.co.uk/tx/scope.htm

http://www.fmamradios.com/Ampliphase_104.html

http://www.fmamradios.com/Ampliphase_111.html

http://www.fmamradios.com/Ampliphase_119.html

Disconnecting the negative peak clipper and feedback rectifier will allow these
transmitters to modulate more than 100% negative. Disconnecting the feedback
rectifier will cause the transfer function to become non linear, this can be
corrected either by employing a piecewise linear transfer function compensator
as in the KOH transmitter, or by replacing the feedback rectifier with a
synchronous demodulator so the feedback doesn't become positive when the
modulation exceeds 100% negative. Simply injecting some extra carrier into the
existing feedback rectifier will create a poor man's synchronous demodulator.

--
Regards,

John Byrns

Surf my web pages at, http://fmamradios.com/

[Patrick Turner]

On Jul 18, 7:21*am, (Don Pearce) wrote:
On Sun, 17 Jul 2011 05:15:19 -0700 (PDT), Patrick Turner





wrote:
On Jul 17, 9:27 pm, (Don Pearce) wrote:
On Sun, 17 Jul 2011 04:14:48 -0700 (PDT), Patrick Turner

wrote:
I tried full wave, but not so easy at all. I've tried a 2 diode
voltage doubler, also not worth the slight increase in Vo.

There is utterly no need for any more than what I have, IMHO.

Have you tried a synchronous detector? Generate a well-limited square
wave at the IF frequency and use it to switch a full wave diode
bridge. This will recover - with perfect linearity - any AM signal,
even one modulated well beyond 100% (provided it has been carried out
properly).

d

I'd have to build a 455kHz square wave oscillator, and its all a lot
of tubes and work and I already get excellent AM reception I'm happy
with. What you say is easier said than done. If you don't agree,
please post fully worked out schematics somewhere at your website to
prove I am uneducated. Ball's in your court now.

I designed a synchronous AM detector back in the days when I was at TV
designer for the Rank Organization. I have no intention of doing it
again because AM radio is by now a busted flush. And of course you
don't use an oscillator, you just apply severe limiting to the IF
signal to strip off the modulation and leave a square wave.

One could make a synchronised or "locked oscillator" no?

I've done that with an FM tuner, one doubles the 19kHz tone and uses
it to synchronise a 38kHz oscillator.
Its important that the produced 38kHz carrier is free of any amplitude
variations.

But with AM, you can't just limit an IF signal because the modulation
used often goes to 100% so there is no 455kHz during the 100% mod to
lock an oscillator, which would spring off the actual IF frequency
everytime the mod goes to near 100%. So a PLL might be better because
the oscillator is controlled by a stored charge in a cap which takes
time to change so the oscillator runs on at the IF despite the IF
becoming zero sometimes. But I have not ever built a PLL, and I see no
need to, I've already got a good simple method adaptable to all old
radios and which uses just one 12AU7 or a few bjts, and it sounds far
better than the old type circuits designed by accountants. Not many
schematics of tubed PLL around.

I ain't in any hurry to change. A simple superhet with a good detector
works just fine.

Patrick Turner.

d- Hide quoted text -

- Show quoted text -

[Don Pearce[_3_]]

On Sun, 17 Jul 2011 17:06:09 -0500, John Byrns
wrote:

In article ,
(Don Pearce) wrote:

On Sun, 17 Jul 2011 04:14:48 -0700 (PDT), Patrick Turner
wrote:

I tried full wave, but not so easy at all. I've tried a 2 diode
voltage doubler, also not worth the slight increase in Vo.

There is utterly no need for any more than what I have, IMHO.

Have you tried a synchronous detector? Generate a well-limited square
wave at the IF frequency and use it to switch a full wave diode
bridge. This will recover - with perfect linearity - any AM signal,
even one modulated well beyond 100% (provided it has been carried out
properly).

Neglecting the problems with the limiter around the zero carrier point, your
scheme doesn't work for signals "modulated well beyond 100%" in the negative
direction. The problem is that when modulation exceeds 100% in the negative
direction the phase of the carrier flips causing the switching of the diode
bridge to be out of phase with the original carrier, causing serious distortion
in the receiver.


That is precisely the circumstance in which the synchronous detector
scores. Over-100% modulation is recovered correctly. With a simply
diode detector, of course, it simply clips.

Adding a phase detector and a VCO to your circuit will fix this problem.


Phase detector and VCO are a good idea, but not needed unless the
carrier genuinely vanishes occasionally. I would hope most
broadcasters try to avoid this. Actually, if the carrier has vanished
there is nothing to be gained with a VCO - there is nothing to detect
in the gap.

Extra credit quiz question, what tube era broadcast transmitter could generate
negative modulation peaks beyond 100% if the negative peak clipper in the
transmitter was disabled?

Any that used a double balanced modulator and linear output stage
rather than an anode current modulator.

d

[John Byrns]

In article ,
Patrick Turner wrote:

On Jul 18, 7:21*am, (Don Pearce) wrote:
On Sun, 17 Jul 2011 05:15:19 -0700 (PDT), Patrick Turner

wrote:
On Jul 17, 9:27 pm, (Don Pearce) wrote:
On Sun, 17 Jul 2011 04:14:48 -0700 (PDT), Patrick Turner

wrote:
I tried full wave, but not so easy at all. I've tried a 2 diode
voltage doubler, also not worth the slight increase in Vo.

There is utterly no need for any more than what I have, IMHO.

Have you tried a synchronous detector? Generate a well-limited square
wave at the IF frequency and use it to switch a full wave diode
bridge. This will recover - with perfect linearity - any AM signal,
even one modulated well beyond 100% (provided it has been carried out
properly).

I'd have to build a 455kHz square wave oscillator, and its all a lot
of tubes and work and I already get excellent AM reception I'm happy
with. What you say is easier said than done. If you don't agree,
please post fully worked out schematics somewhere at your website to
prove I am uneducated. Ball's in your court now.

I designed a synchronous AM detector back in the days when I was at TV
designer for the Rank Organization. I have no intention of doing it
again because AM radio is by now a busted flush. And of course you
don't use an oscillator, you just apply severe limiting to the IF
signal to strip off the modulation and leave a square wave.

One could make a synchronised or "locked oscillator" no?

I've done that with an FM tuner, one doubles the 19kHz tone and uses
it to synchronise a 38kHz oscillator.
Its important that the produced 38kHz carrier is free of any amplitude
variations.

But with AM, you can't just limit an IF signal because the modulation
used often goes to 100% so there is no 455kHz during the 100% mod to
lock an oscillator, which would spring off the actual IF frequency
everytime the mod goes to near 100%. So a PLL might be better because
the oscillator is controlled by a stored charge in a cap which takes
time to change so the oscillator runs on at the IF despite the IF
becoming zero sometimes. But I have not ever built a PLL, and I see no
need to, I've already got a good simple method adaptable to all old
radios and which uses just one 12AU7 or a few bjts, and it sounds far
better than the old type circuits designed by accountants. Not many
schematics of tubed PLL around.

Actually there are pant & shirt loads of tubed PLL schematics around. Back in
the tube era tubed PLLs were a common fixture in the living or lounge rooms of
many people. Many Television receivers of the era used tubed PLLs to control
the horizontal sweep oscillator. Tubed PLLs were also used in some FM broadcast
transmitters, for example the RCA BTE-10B FM exciter, the manual for this is
available on the web, I can provide a link if you are interested. PLLs were
also used in the tube era synchronizing signal generators used by Television
broadcast stations. These old synchronizing generators typically had three
modes, one running off a crystal in which case no PLL was required. The second
mode used a PLL to lock the field frequency to the local power line. The third
mode used a PLL to lock the local sync generator to the incoming network, or
remote, signal. I'm sure I have a book that includes a schematic of one of
these synchronizing generators that I could scan if you are interested.

If you can still find a library with tube ara books, you should be able to find
plenty of schematics for tubed PLLs of various sorts.

--
Regards,

John Byrns

Surf my web pages at, http://fmamradios.com/

[John Byrns]

In article ,
(Don Pearce) wrote:

On Sun, 17 Jul 2011 17:06:09 -0500, John Byrns
wrote:

In article ,
(Don Pearce) wrote:

On Sun, 17 Jul 2011 04:14:48 -0700 (PDT), Patrick Turner
wrote:

I tried full wave, but not so easy at all. I've tried a 2 diode
voltage doubler, also not worth the slight increase in Vo.

There is utterly no need for any more than what I have, IMHO.

Have you tried a synchronous detector? Generate a well-limited square
wave at the IF frequency and use it to switch a full wave diode
bridge. This will recover - with perfect linearity - any AM signal,
even one modulated well beyond 100% (provided it has been carried out
properly).

Neglecting the problems with the limiter around the zero carrier point, your
scheme doesn't work for signals "modulated well beyond 100%" in the negative
direction. The problem is that when modulation exceeds 100% in the negative
direction the phase of the carrier flips causing the switching of the diode
bridge to be out of phase with the original carrier, causing serious
distortion
in the receiver.

That is precisely the circumstance in which the synchronous detector
scores. Over-100% modulation is recovered correctly. With a simply
diode detector, of course, it simply clips.

No, you are completely wrong here! First with negative modulation greater than
100% a diode doesn't "simply clip", what it does, assuming a perfect diode
detector, is recover the envelope of the modulated waveform. What happens
assuming a sine wave test signal, is that the portion of the signal that exceeds
100% negative modulation flips upside down as seen by an envelope detector.

Second what you are describing is what I believe is called a pseudo synchronous
detector. This type of detector responds in a similar manner as the diode
detector does, although for slightly different reasons.

Finally a true synchronous detector using a PLL will correctly recover AM
modulation that exceeds 100% negative modulation.

Try drawing the waveforms involved and you will see what is really happening.

It should be noted that most AM transmitters can't modulate more than 100%
negative, and simply clip at that point. The RCA Ampliphase transmitters could
modulate more than 100% in the negative direction if the audio negative peak
clipper and feedback rectifier are disconnected.

It is possible that some of the modern Digital Transmtters can also do this trick

Adding a phase detector and a VCO to your circuit will fix this problem.

Phase detector and VCO are a good idea, but not needed unless the
carrier genuinely vanishes occasionally. I would hope most
broadcasters try to avoid this.

Yes, negative modulation is often limited to 95%.

Actually, if the carrier has vanished
there is nothing to be gained with a VCO - there is nothing to detect
in the gap.

That is only true with a traditional transmitter that clips negative peaks. A
modified Ampliphase transmitter, a suitable digital transmitter, or a
transmitter using a balanced modulator feeding a linear amplifier, will fill the
gaps correctly. An example of the latter type of transmitter would be an SSB
transmitter with some DC added to the audio signal to generate a traditional AM
signal.

Extra credit quiz question, what tube era broadcast transmitter could
generate
negative modulation peaks beyond 100% if the negative peak clipper in the
transmitter was disabled?

Any that used a double balanced modulator and linear output stage
rather than an anode current modulator.

That would work, were there ever any broadcast transmitters built with that
architecture?

--
Regards,

John Byrns

Surf my web pages at, http://fmamradios.com/

[Patrick Turner]

On Jul 19, 7:17*am, John Byrns wrote:
In article ,
*Patrick Turner wrote:





On Jul 18, 7:21*am, (Don Pearce) wrote:
On Sun, 17 Jul 2011 05:15:19 -0700 (PDT), Patrick Turner

wrote:
On Jul 17, 9:27 pm, (Don Pearce) wrote:
On Sun, 17 Jul 2011 04:14:48 -0700 (PDT), Patrick Turner

wrote:
I tried full wave, but not so easy at all. I've tried a 2 diode
voltage doubler, also not worth the slight increase in Vo.

There is utterly no need for any more than what I have, IMHO.

Have you tried a synchronous detector? Generate a well-limited square
wave at the IF frequency and use it to switch a full wave diode
bridge. This will recover - with perfect linearity - any AM signal,
even one modulated well beyond 100% (provided it has been carried out
properly).

I'd have to build a 455kHz square wave oscillator, and its all a lot
of tubes and work and I already get excellent AM reception I'm happy
with. What you say is easier said than done. If you don't agree,
please post fully worked out schematics somewhere at your website to
prove I am uneducated. Ball's in your court now.

I designed a synchronous AM detector back in the days when I was at TV
designer for the Rank Organization. I have no intention of doing it
again because AM radio is by now a busted flush. And of course you
don't use an oscillator, you just apply severe limiting to the IF
signal to strip off the modulation and leave a square wave.

One could make a synchronised or "locked oscillator" no?

I've done that with an FM tuner, one doubles the 19kHz tone and uses
it to synchronise a 38kHz oscillator.
Its important that the produced 38kHz carrier is free of any amplitude
variations.

But with AM, you can't just limit an IF signal because the modulation
used often goes to 100% so there is no 455kHz during the 100% mod to
lock an oscillator, which would spring off the actual IF frequency
everytime the mod goes to near 100%. So a PLL might be better because
the oscillator is controlled by a stored charge in a cap which takes
time to change so the oscillator runs on at the IF despite the IF
becoming zero sometimes. But I have not ever built a PLL, and I see no
need to, I've already got a good simple method adaptable to all old
radios and which uses just one 12AU7 or a few bjts, and it sounds far
better than the old type circuits designed by accountants. Not many
schematics of tubed PLL around.

Actually there are pant & shirt loads of tubed PLL schematics around. *Back in
the tube era tubed PLLs were a common fixture in the living or lounge rooms of
many people. *Many Television receivers of the era used tubed PLLs to control
the horizontal sweep oscillator. *Tubed PLLs were also used in some FM broadcast
transmitters, for example the RCA BTE-10B FM exciter, the manual for this is
available on the web, I can provide a link if you are interested. *PLLs were
also used in the tube era synchronizing signal generators used by Television
broadcast stations. *These old synchronizing generators typically had three
modes, one running off a crystal in which case no PLL was required. *The second
mode used a PLL to lock the field frequency to the local power line. *The third
mode used a PLL to lock the local sync generator to the incoming network, or
remote, signal. *I'm sure I have a book that includes a schematic of one of
these synchronizing generators that I could scan if you are interested.

I've never worked on TV sets. Never ever seen any working tubed PLL.

I'll have to search further -- if I get time.

I've read a few amateur radio books and all of them DON'T have PLL
which operate at 455kHz and with tubes.

I dimly recall seen a tubes PLL in the 1960s in Electronics Australia
magazine, but then I had no use for it.


If you can still find a library with tube ara books, you should be able to find
plenty of schematics for tubed PLLs of various sorts.

I'll check Google before venturing to a library. I been to libraries
at unis and tech colleges before in 1990s, but but never ever seen
useful PLLs for FM or AM detection with tubes. Maybe accountants
banned their use unless they were absolutely necessary, as they may be
in a TV set, for which ppl were prepared to pay a lotta money for to
keep the accountants employed.

Patrick Turner.

--
Regards,

John Byrns

Surf my web pages at, *http://fmamradios.com/- Hide quoted text -

- Show quoted text -

[Don Pearce[_3_]]

On Mon, 18 Jul 2011 16:20:55 -0500, John Byrns
wrote:

In article ,
(Don Pearce) wrote:

On Sun, 17 Jul 2011 17:06:09 -0500, John Byrns
wrote:

In article ,
(Don Pearce) wrote:

On Sun, 17 Jul 2011 04:14:48 -0700 (PDT), Patrick Turner
wrote:

I tried full wave, but not so easy at all. I've tried a 2 diode
voltage doubler, also not worth the slight increase in Vo.

There is utterly no need for any more than what I have, IMHO.

Have you tried a synchronous detector? Generate a well-limited square
wave at the IF frequency and use it to switch a full wave diode
bridge. This will recover - with perfect linearity - any AM signal,
even one modulated well beyond 100% (provided it has been carried out
properly).

Neglecting the problems with the limiter around the zero carrier point, your
scheme doesn't work for signals "modulated well beyond 100%" in the negative
direction. The problem is that when modulation exceeds 100% in the negative
direction the phase of the carrier flips causing the switching of the diode
bridge to be out of phase with the original carrier, causing serious
distortion
in the receiver.

That is precisely the circumstance in which the synchronous detector
scores. Over-100% modulation is recovered correctly. With a simply
diode detector, of course, it simply clips.

No, you are completely wrong here! First with negative modulation greater than
100% a diode doesn't "simply clip", what it does, assuming a perfect diode
detector, is recover the envelope of the modulated waveform. What happens
assuming a sine wave test signal, is that the portion of the signal that exceeds
100% negative modulation flips upside down as seen by an envelope detector.

Quite so, my bad.

Second what you are describing is what I believe is called a pseudo synchronous
detector. This type of detector responds in a similar manner as the diode
detector does, although for slightly different reasons.


No, what I am describing is an entirely synchronous detector - maybe I
described it poorly.

d

[John Byrns]

In article ,
(Don Pearce) wrote:

On Mon, 18 Jul 2011 16:20:55 -0500, John Byrns
wrote:

In article ,
(Don Pearce) wrote:

On Sun, 17 Jul 2011 17:06:09 -0500, John Byrns
wrote:

In article ,
(Don Pearce) wrote:

On Sun, 17 Jul 2011 04:14:48 -0700 (PDT), Patrick Turner
wrote:

I tried full wave, but not so easy at all. I've tried a 2 diode
voltage doubler, also not worth the slight increase in Vo.

There is utterly no need for any more than what I have, IMHO.

Have you tried a synchronous detector? Generate a well-limited square
wave at the IF frequency and use it to switch a full wave diode
bridge. This will recover - with perfect linearity - any AM signal,
even one modulated well beyond 100% (provided it has been carried out
properly).

Neglecting the problems with the limiter around the zero carrier point,
your
scheme doesn't work for signals "modulated well beyond 100%" in the
negative
direction. The problem is that when modulation exceeds 100% in the
negative
direction the phase of the carrier flips causing the switching of the
diode
bridge to be out of phase with the original carrier, causing serious
distortion
in the receiver.

That is precisely the circumstance in which the synchronous detector
scores. Over-100% modulation is recovered correctly. With a simply
diode detector, of course, it simply clips.

No, you are completely wrong here! First with negative modulation greater
than
100% a diode doesn't "simply clip", what it does, assuming a perfect diode
detector, is recover the envelope of the modulated waveform. What happens
assuming a sine wave test signal, is that the portion of the signal that
exceeds
100% negative modulation flips upside down as seen by an envelope detector.

Quite so, my bad.

Second what you are describing is what I believe is called a pseudo
synchronous
detector. This type of detector responds in a similar manner as the diode
detector does, although for slightly different reasons.

No, what I am describing is an entirely synchronous detector - maybe I
described it poorly.

I thought you described it perfectly well. I suppose our difference stems from
how you define a "synchronous detector". I define a "synchronous detector" as
using a locally regenerated carrier that is essentially identical to the carrier
that was used in the transmitter. The local carrier in your pseudo synchronous
detector doesn't meet my criterion because the phase of the locally regenerated
carrier flips every time the modulation crosses the 100% negative modulation
line.

--
Regards,

John Byrns

Surf my web pages at, http://fmamradios.com/

[Don Pearce[_3_]]

On Tue, 19 Jul 2011 19:43:57 -0500, John Byrns
wrote:

In article ,
(Don Pearce) wrote:

On Mon, 18 Jul 2011 16:20:55 -0500, John Byrns
wrote:

In article ,
(Don Pearce) wrote:

On Sun, 17 Jul 2011 17:06:09 -0500, John Byrns
wrote:

In article ,
(Don Pearce) wrote:

On Sun, 17 Jul 2011 04:14:48 -0700 (PDT), Patrick Turner
wrote:

I tried full wave, but not so easy at all. I've tried a 2 diode
voltage doubler, also not worth the slight increase in Vo.

There is utterly no need for any more than what I have, IMHO.

Have you tried a synchronous detector? Generate a well-limited square
wave at the IF frequency and use it to switch a full wave diode
bridge. This will recover - with perfect linearity - any AM signal,
even one modulated well beyond 100% (provided it has been carried out
properly).

Neglecting the problems with the limiter around the zero carrier point,
your
scheme doesn't work for signals "modulated well beyond 100%" in the
negative
direction. The problem is that when modulation exceeds 100% in the
negative
direction the phase of the carrier flips causing the switching of the
diode
bridge to be out of phase with the original carrier, causing serious
distortion
in the receiver.

That is precisely the circumstance in which the synchronous detector
scores. Over-100% modulation is recovered correctly. With a simply
diode detector, of course, it simply clips.

No, you are completely wrong here! First with negative modulation greater
than
100% a diode doesn't "simply clip", what it does, assuming a perfect diode
detector, is recover the envelope of the modulated waveform. What happens
assuming a sine wave test signal, is that the portion of the signal that
exceeds
100% negative modulation flips upside down as seen by an envelope detector.

Quite so, my bad.

Second what you are describing is what I believe is called a pseudo
synchronous
detector. This type of detector responds in a similar manner as the diode
detector does, although for slightly different reasons.

No, what I am describing is an entirely synchronous detector - maybe I
described it poorly.

I thought you described it perfectly well. I suppose our difference stems from
how you define a "synchronous detector". I define a "synchronous detector" as
using a locally regenerated carrier that is essentially identical to the carrier
that was used in the transmitter. The local carrier in your pseudo synchronous
detector doesn't meet my criterion because the phase of the locally regenerated
carrier flips every time the modulation crosses the 100% negative modulation
line.

Yes, clearly the phase must not be allowed to flip with the signal or
the detector won't work. A high-Q tuned circuit will do the job
nicely, carrying the original phase through the inverted peak.

A PLL can also suffer the same problem, depending on the time constant
of the loop. The effective Q must be the same whether resonator or
PLL.

d