On Tue, 16 Nov 2010 19:19:09 -0600, "amdx" wrote:

You have the diode to drive before the audio section and you want to match
the
diodes impedance.

Hi Mike,

This seems to wander the field when there are conflicting agendas on
the table. The wandering is due in large part to the absence of
specific numbers, and in this particular case, even the sense of
scale.

We have an audio Z match. We have a diode Z match. There is no sense
that if this is a high-Z or a low-Z for either/both/neither.

To this point, you have hewed to the commonplace magnetic speaker, and
from that I can only imagine that the problem with audio match is
transforming from a high (diode) Z to a low (speaker) Z. But from the
numbers, it would seem that both diode and speaker Zs are on par. I
could be wrong because this drift takes me into guessing.

Solve your problem and get a piezo headphone (or make one, this is the
tradition of Xtal radio, n'est pas?). They have to exhibit at least
10KOhm if not a bajillion ohms. If there isn't enough current to fire
the diode, Guess What? Slap a sacrificial resistor across the piezo
leads!

Radio Shack sells (or used to the last time I bought one) two or three
inch barium titanate ceramic disks that are offered as buzzer elements
(or rob a clock alarm for its buzzer - these things are a drug on the
market).

* * * * * * * * * * *

Now, if you only return to complain that the Hi-Z phones cause a
problem somewhere else; then I would recommend you abandon circuit
design and take up the study of math employing figures of merit. Such
assignments of merit are subjective, of course, but then you get to
own your personal solution. Discussion that follows this satisfactory
conclusion then enters the appropriate arena of philosophical or
religious controversy. This will certainly garner far more
participation in this group.

73's
Richard Clark, KB7QHC

"Richard Clark" wrote in message
...
On Tue, 16 Nov 2010 19:19:09 -0600, "amdx" wrote:

Why start with a 1.5dB deficit when the Tank is already there to do
the work of matching at no loss?

You have the diode to drive before the audio section and you want to
match
the
diodes impedance. Tapping down will allow you to match diode impedance,
but
I'm not sure tapped down is where you will find the best diode efficiency.
Your response to my "not sure " is anticipated. :-)

Hi Mike,

Unfortunately your response is not an answer. Your response suffers
equally for the proposed additional transformer - and you have added
loss for no net forward movement to the solution.

Second,
Tapping down on the tank coil may not work as well as first thought.

You are going to have to explain that better, because what follows
doesn't.


You still have diode characteristics to overcome, and as you tap down
the
voltage also decreases.

"Here's a quote from Ben Tongue's webpage;
Many times the question is asked, "What is the best diode to use?" The
answer depends on the specific RF source resistance and audio load
impedance
of the Crystal Set in question.

At first blush I have to ask, "How many source resistances are there
to be found for a Xtal radio?" I am not under the impression you have
much flexibility in that regard.

Every time you change frequency you change source resistance.
XL times Q equals source resistance, Q changes with frequency,
and XL changes with frequency.
Also when you change to a different station with a different signal strength
you change the current through the diode, this changes the resistance the
diode
presents to the tank.

At low signal levels the RF input
resistance and audio output resistance of a detector diode are equal to
25,700,000*n/Is Ohms (current in nA).

25 million WHAT?

I don't know what that is derived from.


For minimum detector power loss at
very low signal levels with a particular diode,

If it were a "particular" diode, it would seem giving it a part number
would lend authenticity to this report.

Well, the rest reads about like an oath to motherhood and apple pie.

Do they offer a case for fumbling through a selection of diodes? At
first blush, the best is going to be the best - hands down.

See above.
And yes they sometimes multiple diodes on there radio.

I am less than whelmed.

Still hyave concern about diode characteristics.

Name one characteristic that presents a concern.

Saturation current and
axis-crossing resistance equal to Rr
Sorry that's two, but their related.

See Ben's Page http://www.bentongue.com/xtalset/4opd_xfr/4opd_xfr.html

Frankly, it looks like spaghetti math. If it suits you, then there is
a number that will supply the optimal solution - that is the nature of
math, after all. However, too much of this looks circular, especially
when I see complexity piled on top of loss just to return to the same
problem.

However, what it all boils down to is that you have to live with
whatever diode you select, and you shift the tap for the best
performance (not many variables left in the game of Xtal radios, is
there?).

There are dozens of diodes to pick from, depends on the tank Z
and signal strength. I don't know what you mean when you say there aren't
many variables, there are many, many!
Coil Type (solenoid, spider, 3D solenoid)
Coil Form Material, paper, PVC, polypropylene, etc.
Wire type, solid, Litz
Antenna Match technique, Tap, cap tune, inductance coupling
Diodes, Diodes, Diodes
Audio, Transformer matching , high impedance piezo, sound powered
headphones, crystal earphone, horn driver.

Here's some more info on the diode selection and why.
Richard I'm well aware of your great knowledge (not being facetious)
and my lack of math and knowledge on this, but I'm to the point where
it seems you are not willing to learn new information. It seems you already
know it all. This especially in regards to the idea that the diodes
characterists
are very important to losses in the radio.
I think one of these pages has a graph showing a diode not matched to the
current available has a 33db loss.
Here's more diode info.
http://www.crystal-radio.eu/endiodes.htm
http://www.bentongue.com/xtalset/17Is_n/17Is_n.html
http://www.klimaco.com/HAMRADIOPAGES/xtal_how_to.htm

I do appreciate your time and the discusion, I did learn from it.
I'm going to take a few days off.
Sincerely, MikeK




73's
Richard Clark, KB7QHC

In article ,
Richard Clark wrote:

We have an audio Z match. We have a diode Z match. There is no sense
that if this is a high-Z or a low-Z for either/both/neither.

Richard-

We have a diode detector, which is nonlinear. If you allow a filter
capacitor on its output, there is current flow only on alternate RF
voltage peaks, in order to charge up the capacitor for the amount lost
to the audio load during the cycle. Its effect on circuit Q may be a
fraction of the total, also consisting of coil Q and input circuit
matching.

Therefore you want the diode connected to the high end of the coil, not
transformed to a lower impedance point on the coil.

Load impedance will affect audio level, so should be kept high for that
reason. However there may be some value of load that produces the
greatest audio output voltage. If too high, the filter capacitor would
charge to the peak of the modulation waveform and stay there.

Worst case is a low impedance audio load and no filter capacitor. Half
of the RF waveform is essentially cut off, leaving a peak-to-peak value
of one half the open circuit value. Since energy is extracted from the
circuit during the conduction cycle, the effect on Q should be
approximately the same as if a resistor having the same peak-to-peak
effect was connected across the circuit, disregarding the RMS value of a
half-sine waveform.

Fred
K4DII

On Wed, 17 Nov 2010 09:17:51 -0600, "amdx" wrote:

At first blush I have to ask, "How many source resistances are there
to be found for a Xtal radio?" I am not under the impression you have
much flexibility in that regard.

Every time you change frequency you change source resistance.
XL times Q equals source resistance, Q changes with frequency,
and XL changes with frequency.

Hi Mike,

There are a number of problems here. Q is strictly a function of
RESISTANCE and nothing else.

What you call source resistance, when using the math you provide, is
actually source impedance. Impedance embraces resistance and is
solely resistance when reactance is gone - such as at resonance.
However, it would seem that the Xtal Radio community treats reactance
as resistance interchangeably.

That is fine to a point, but it introduces serious problems in the
usage of the term "match" as there are actually several types of
"matching" and each has its peculiarity (only peculiar when you are
unfamiliar with the other types).

It is a given that you are going to have to change either Xc or Xl if
you are going to tune to another frequency. This alters the circuit
Z, certainly; and for the AM band you have a 3:1 variation from band
end-to-end.

Also when you change to a different station with a different signal strength
you change the current through the diode, this changes the resistance the
diode
presents to the tank.

Without numbers, this is like anticipating worry about your MPG
driving up a hill in comparison to driving down the same hill. Yes, a
wild variation, but do you give up cars and start walking instead?
Some of these "concerns" sound like they come from the AVC needy. I
thought Xtal Radio folks would man up and get with the program.



However, what it all boils down to is that you have to live with
whatever diode you select, and you shift the tap for the best
performance (not many variables left in the game of Xtal radios, is
there?).

There are dozens of diodes to pick from, depends on the tank Z
and signal strength.

No, actually you've either mis-read the intent (which is certainly a
problem with that author's hodge-podge writing style) or you've read
too many written sources that each only see part of the elephant.
"It's a snake! We need something bulky."
"It's a wall! We need something slim."
"It's thin and flappy! We need something sturdy."

The problem with the diode is in its forward current and reverse
current. If they are about the same, you don't have a rectifier. This
would be a weak signal problem. The forward current defines the
series resistance of the diode and how much that is in comparison to
the load defines the efficiency. You want more current to lower the
diode resistance to increase the efficiency. Unfortunately this
demands you have a higher source voltage that offers less current
(Catch-22).

However, this problem is strictly for a low Z load like a magnetic
speaker - and thus enters the consternation over competing agendas
that gives everyone the fits that seems they enjoy agonizing over when
the solution is obvious. I will touch on that next:

I don't know what you mean when you say there aren't
many variables, there are many, many!
Coil Type (solenoid, spider, 3D solenoid)
Coil Form Material, paper, PVC, polypropylene, etc.

Those are non-starters for those who love to struggle in hope against
reality.

Wire type, solid, Litz

Ah yes, the magic Litz wire that is the salvation. The solution for
the problem that doesn't matter. You do this because you can, not
because you need to. And then you return to life with the car still
in the ditch (but you changed the air in the tires).

Antenna Match technique, Tap, cap tune, inductance coupling

This has been long figured out. For nearly a century.

Diodes, Diodes, Diodes

Whose solution is found in:

Audio, Transformer matching , high impedance piezo, sound powered
headphones, crystal earphone, horn driver.

Whose choice is obvious in hi impedance.

Like I said, there are not that many variables, until you find someone
on a soap box with a stump speech about the invidious dereliction of
design introduced with the Philips head screw. The advantages of the
Bristol head anyone?

Do not confuse the multitude of competing and contradicting factors as
the multitude of variables.

Introduce figure of merit to the discussion and argue about that. If
that gets resolved, the solution will pop right out.

Here's some more info on the diode selection and why.
Richard I'm well aware of your great knowledge (not being facetious)
and my lack of math and knowledge on this, but I'm to the point where
it seems you are not willing to learn new information.

There is no new information. It is a niche application you are
talking about, not new science. The niche has its concerns and they
are impacted by issues that are rarely a problem for the standard
usage of the components you mention, but there have been no new
devices introduced in this discussion. I see nothing innovative. In
fact, the discussion is so in-bred it lacks coverage outside of a
clubbish mainstream. I will offer something new below that is
actually quite old.

It seems you already
know it all. This especially in regards to the idea that the diodes
characterists
are very important to losses in the radio.

I have designed at the extreme edges of component characteristics,
true. More often in regard to temperature variation, mechanical
vibration, noise, accuracy, resolution, speed, bandwidth, linearity -
all that went towards continuous service in a largish orange box in
some jets.

I have studied diodes to no end, especially LEDs and photodetectors.
Talk about marginal signal detection and loss. I started with 1n34s
being used as varactors back in the late 60s. PIN and Tunnel diodes
followed. There is a world of non-linear devices out there. I
especially like the surgistor and posistor. Any reports of their use
in Xtal sets?

I have maintained a number of audio standards and measurement devices.
I have had to repair detectors that measured absolute value voltage
out to 7 places, the same issues arose there too. Small currents?
Like those that rise continuously through the air in femto-ampere
levels? You would need polonium treated probes to measure them (or
just use the anti-static record brushes of the 1970s).

I think one of these pages has a graph showing a diode not matched to the
current available has a 33db loss.
Here's more diode info.
http://www.crystal-radio.eu/endiodes.htm
http://www.bentongue.com/xtalset/17Is_n/17Is_n.html
http://www.klimaco.com/HAMRADIOPAGES/xtal_how_to.htm

This stuff merely shows the agony of counter trending data. In other
words select the optimal Is and you get trash Rd - and vice versa.
Reporting the numbers doesn't solve the problem however.

* * * * * * * *

In this offered discussion there is nothing of how the pioneers
managed it with galena crystals and cat whiskers. That data for
comparison is starkly absent in this and their discussion.

My Fourth Edition of "Standard Handbook for Electrical Engineers"
(1915) offers Coherers and Magnetic detectors. The coverage is
likewise absent of these alternatives.

"Section 21-289 Contact Rectifiers. The rectifiers most frequently
consist of a contact between a fine wire and some variety of mineral.
Among the minerals frequently use are iron pyrites, galena, silicon,
and molybdenite. In other forms, two crystals are used in contact,
such as zinkite with chaleopyryte or bornite, or silicon with metallic
arsenic.
"For detecting the direct-current pulses, head telephones of from
1,000 to 3,000 Ohms resistance are ordinarily used."

Returning to the obscure references, for magnetic detectors, consider:
http://en.wikipedia.org/wiki/Magnetic_detector
for the coherer, consider
http://en.wikipedia.org/wiki/Coherer

* * * * * * *

You might want to reflect that this thread has not been about matching
an antenna to a crystal radio at all. That topic was rather easily
and quickly answered by J. Todd. Loss is defined by wire resistance
and the proximity of objects to the tank.

The match from the Tank is similarly simple - once you define the
load. Loss is defined by wire resistance and the proximity of objects
to the tank.

The load is presented as its two native components: the detector and
the headset.

The detector is the variable of concern, it is defined by the current
necessary to drive the headset, everything surrounding the detector
has to conform to its choice. The loss is defined by the ratio of the
detector resistance to the load resistance plus that diode resistance.

The headset is defined by a power level you identified as one picowatt
delivered to the ear. I think we can agree that was very generous.
The loss is defined by the transducer efficiency.

This is the genesis of figure of merit, and that is your solution.

73's
Richard Clark, KB7QHC

At low signal levels the RF input
resistance and audio output resistance of a detector diode are equal to
25,700,000*n/Is Ohms (current in nA).

25 million WHAT?

Ok One more, with a little help I figured it out.
It has to to with the Thermal Voltage of the diode.
Vt=KT/q
with k=1.38E-23 and q=1.6E-19. T is absolute temperature in degrees Kelvin,
k is Boltzmann's constant and q is the charge of an electron. VT is close
to 0.025 volts at 20 degrees Celsius.

With a slightly increased temperature the .025 is raised to .0257,
and .0257 / 1 nanoamp =25,700,000
Thanks, MikeK

On Wed, 17 Nov 2010 14:57:22 -0500, Fred McKenzie
wrote:

In article ,
Richard Clark wrote:

We have an audio Z match. We have a diode Z match. There is no sense
that if this is a high-Z or a low-Z for either/both/neither.

Hi Fred,

The subtext of that observation was the lack of numbers.
Quantification will aid immensely and give perspective to the direness
of the problem.

We have a diode detector, which is nonlinear. If you allow a filter
capacitor on its output, there is current flow only on alternate RF
voltage peaks, in order to charge up the capacitor for the amount lost
to the audio load during the cycle. Its effect on circuit Q may be a
fraction of the total, also consisting of coil Q and input circuit
matching.

This is dangerously close to decoupling the problem by looking at
minutia. Q is the ratio of power in to power consumed.

Therefore you want the diode connected to the high end of the coil, not
transformed to a lower impedance point on the coil.

Your detector placement serves the consumption side and consumption -
listening to program content - is the whole point of detection.

Preserving Q does not serve that goal.

Load impedance will affect audio level, so should be kept high for that
reason.

Audio level "should" follow power applied, be it low voltage high
current (a speaker); or high voltage low current (a piezo). Preserving
power levels while changing its form is the purpose transduction.

However there may be some value of load that produces the
greatest audio output voltage. If too high, the filter capacitor would
charge to the peak of the modulation waveform and stay there.

What you describe is a clamp, not a filter; but point is well taken.
What you have is the wrong balance of charge time to discharge time.
In other words, you have selected the wrong detector for the chosen
transducer, or vice-versa. Choose both to complement each other.

Consider, what would be lost if you clipped out that pesky cap?

73's
Richard Clark, KB7QHC

"Richard Clark" wrote in message
...
On Wed, 17 Nov 2010 09:17:51 -0600, "amdx" wrote:

At first blush I have to ask, "How many source resistances are there
to be found for a Xtal radio?" I am not under the impression you have
much flexibility in that regard.

Every time you change frequency you change source resistance.
XL times Q equals source resistance, Q changes with frequency,
and XL changes with frequency.

Hi Mike,

There are a number of problems here. Q is strictly a function of
RESISTANCE and nothing else.

Ya, loss resistance of the tank.
Can you agree that it varies with frequency?


What you call source resistance, when using the math you provide, is
actually source impedance. Impedance embraces resistance and is
solely resistance when reactance is gone - such as at resonance.

The tank is used at resonance, so yes, source resistance.

However, it would seem that the Xtal Radio community treats reactance
as resistance interchangeably.

Not so!


That is fine to a point, but it introduces serious problems in the
usage of the term "match" as there are actually several types of
"matching" and each has its peculiarity (only peculiar when you are
unfamiliar with the other types).

Are you getting at conjugate match?

It is a given that you are going to have to change either Xc or Xl if
you are going to tune to another frequency. This alters the circuit
Z, certainly; and for the AM band you have a 3:1 variation from band
end-to-end.

So your in agreement that Rp is not fixed?

Also when you change to a different station with a different signal
strength
you change the current through the diode, this changes the resistance the
diode
presents to the tank.

Without numbers, this is like anticipating worry about your MPG
driving up a hill in comparison to driving down the same hill. Yes, a
wild variation, but do you give up cars and start walking instead?
Some of these "concerns" sound like they come from the AVC needy. I
thought Xtal Radio folks would man up and get with the program.

No pick a different diode.

However, what it all boils down to is that you have to live with
whatever diode you select, and you shift the tap for the best
performance (not many variables left in the game of Xtal radios, is
there?).

There are dozens of diodes to pick from, depends on the tank Z
and signal strength.

No, actually you've either mis-read the intent (which is certainly a
problem with that author's hodge-podge writing style) or you've read
too many written sources that each only see part of the elephant.
"It's a snake! We need something bulky."
"It's a wall! We need something slim."
"It's thin and flappy! We need something sturdy."

The problem with the diode is in its forward current and reverse
current. If they are about the same, you don't have a rectifier. This
would be a weak signal problem. The forward current defines the
series resistance of the diode and how much that is in comparison to
the load defines the efficiency. You want more current to lower the
diode resistance to increase the efficiency. Unfortunately this
demands you have a higher source voltage that offers less current
(Catch-22).

However, this problem is strictly for a low Z load like a magnetic
speaker - and thus enters the consternation over competing agendas
that gives everyone the fits that seems they enjoy agonizing over when
the solution is obvious. I will touch on that next:

I don't know what you mean when you say there aren't
many variables, there are many, many!
Coil Type (solenoid, spider, 3D solenoid)
Coil Form Material, paper, PVC, polypropylene, etc.

Those are non-starters for those who love to struggle in hope against
reality.

Wire type, solid, Litz

Ah yes, the magic Litz wire that is the salvation. The solution for
the problem that doesn't matter. You do this because you can, not
because you need to. And then you return to life with the car still
in the ditch (but you changed the air in the tires).

Can you wind a coil with a Q over 1000 with solid wire?

Antenna Match technique, Tap, cap tune, inductance coupling

This has been long figured out. For nearly a century.

Diodes, Diodes, Diodes

Whose solution is found in:

Audio, Transformer matching , high impedance piezo, sound powered
headphones, crystal earphone, horn driver.

Whose choice is obvious in hi impedance.

Like I said, there are not that many variables, until you find someone
on a soap box with a stump speech about the invidious dereliction of
design introduced with the Philips head screw. The advantages of the
Bristol head anyone?

Do not confuse the multitude of competing and contradicting factors as
the multitude of variables.

Introduce figure of merit to the discussion and argue about that. If
that gets resolved, the solution will pop right out.

Here's some more info on the diode selection and why.
Richard I'm well aware of your great knowledge (not being facetious)
and my lack of math and knowledge on this, but I'm to the point where
it seems you are not willing to learn new information.

There is no new information. It is a niche application you are
talking about, not new science. The niche has its concerns and they
are impacted by issues that are rarely a problem for the standard
usage of the components you mention, but there have been no new
devices introduced in this discussion. I see nothing innovative. In
fact, the discussion is so in-bred it lacks coverage outside of a
clubbish mainstream. I will offer something new below that is
actually quite old.

It seems you already
know it all. This especially in regards to the idea that the diodes
characterists
are very important to losses in the radio.

I have designed at the extreme edges of component characteristics,
true. More often in regard to temperature variation, mechanical
vibration, noise, accuracy, resolution, speed, bandwidth, linearity -
all that went towards continuous service in a largish orange box in
some jets.

I have studied diodes to no end, especially LEDs and photodetectors.
Talk about marginal signal detection and loss. I started with 1n34s
being used as varactors back in the late 60s. PIN and Tunnel diodes
followed. There is a world of non-linear devices out there. I
especially like the surgistor and posistor. Any reports of their use
in Xtal sets?

I have maintained a number of audio standards and measurement devices.
I have had to repair detectors that measured absolute value voltage
out to 7 places, the same issues arose there too. Small currents?
Like those that rise continuously through the air in femto-ampere
levels? You would need polonium treated probes to measure them (or
just use the anti-static record brushes of the 1970s).

I think one of these pages has a graph showing a diode not matched to the
current available has a 33db loss.
Here's more diode info.
http://www.crystal-radio.eu/endiodes.htm
http://www.bentongue.com/xtalset/17Is_n/17Is_n.html
http://www.klimaco.com/HAMRADIOPAGES/xtal_how_to.htm

This stuff merely shows the agony of counter trending data. In other
words select the optimal Is and you get trash Rd - and vice versa.
Reporting the numbers doesn't solve the problem however.

* * * * * * * *

In this offered discussion there is nothing of how the pioneers
managed it with galena crystals and cat whiskers. That data for
comparison is starkly absent in this and their discussion.

My Fourth Edition of "Standard Handbook for Electrical Engineers"
(1915) offers Coherers and Magnetic detectors. The coverage is
likewise absent of these alternatives.

"Section 21-289 Contact Rectifiers. The rectifiers most frequently
consist of a contact between a fine wire and some variety of mineral.
Among the minerals frequently use are iron pyrites, galena, silicon,
and molybdenite. In other forms, two crystals are used in contact,
such as zinkite with chaleopyryte or bornite, or silicon with metallic
arsenic.
"For detecting the direct-current pulses, head telephones of from
1,000 to 3,000 Ohms resistance are ordinarily used."

Returning to the obscure references, for magnetic detectors, consider:
http://en.wikipedia.org/wiki/Magnetic_detector
for the coherer, consider
http://en.wikipedia.org/wiki/Coherer

* * * * * * *

You might want to reflect that this thread has not been about matching
an antenna to a crystal radio at all. That topic was rather easily
and quickly answered by J. Todd. Loss is defined by wire resistance
and the proximity of objects to the tank.

The match from the Tank is similarly simple - once you define the
load. Loss is defined by wire resistance and the proximity of objects
to the tank.

The load is presented as its two native components: the detector and
the headset.

The detector is the variable of concern, it is defined by the current
necessary to drive the headset, everything surrounding the detector
has to conform to its choice. The loss is defined by the ratio of the
detector resistance to the load resistance plus that diode resistance.

The headset is defined by a power level you identified as one picowatt
delivered to the ear. I think we can agree that was very generous.
The loss is defined by the transducer efficiency.

This is the genesis of figure of merit, and that is your solution.

73's
Richard Clark, KB7QHC

I think you finally got it Richard.
Thanks, MikeK :-)

*Use a fork to scoop water out of a glass to satisfy your thirst.
I will work, just not well.
* * * * * * * * MikeK


True, but operating crystal radios is much like sailing.

Maybe multiple taps is not that bad an idea.

- I never saw a crystal radio optimized for such a small tuning ration
that the LC had a fixed impedance. A 1-to-3 tuning ratio is quite
normal, and tank impedance will therefore vary considerably.

- On medium wave the impedance of the electrically very short wire
antenna will vary, but not that much. It will stay probably in the
order of a kiloohm or more. On shortwave it will swing considerably.

- If the receiver has to resolve different stations or at least
clearly separate different SW broadcast bands (to listen to the most
powerful broadcaster in each band, more is not doable), it may make
sense to reduce the loading below optimum energy transfer at the
resonant frequency.

This suggests that a coil with multiple taps could be useful. If the
radio has a tapped coil and 3 croc clips (one each for the tuning cap,
antenna, and detector) the user has to optimize over 3 dimensions in,
and 2 dimensions out - wanted signal level and interference.

Great fun, but not easy.

The human ear is not linear and tends to be a bad level measuring
instrument. It helps if the radio also has a microamp-meter to make
relative linear comparisons in total signal level. A 50uA meter may
have a DC resistance of 2kohm and load about the same as a magnetic
headset, but a 500uA may be sensitive enough and load less.

The combination of separate improvements that the ear alone could not
judge may in the end be considerable.

A digital voltmeter would be so high impedance as to be practically
invisible even at very high impedances too, but it has a battery
powered DC amplifier. IMO it is aesthetically incongruent with a RF-
powered radio.

On Nov 15, 1:26*am, Wimpie wrote:
On 15 nov, 02:59, "amdx" wrote:

Hi Guys,
*Assuming I have a tank circuit on a crystal radio with a Z at resonance of
1.5 megaohms.
How would I make an antenna and extract maximum signal and keep the Z at
750,000 ohms.
* If don't think that's what I want to do, tell me that too. :-)
* * * * * * * * * * * * * * * * * * *MikeK

Hello Mike,

Assuming you have a long wire outside and a ground provision, you may
use an inductive coupling. *By changing the distance between the
antenna coil and the receiver coil, you modify the impedance
transformation.

Other method is capacitive coupling. *Probably positioning the end of
the antenna wire close to the high impedance side of the tank circuit
will give the desired effect. Changing the distance changes the
coupling.

For inductive coupling, your receiver circuit can be floating, for
capacitive coupling, the receiver should be grounded.

By changing the coupling you can optimize for maximum selectivity
(with reduced sensitivity) or maximum sensitivity (with reduces
selectivity).

Regarding the antenna, assuming LW and AM reception, long combined
with high gives strongest signal, hence you can reduce the coupling to
get best selectivity.

Best regards,

Wim
PA3DJSwww.tetech.nl
without abc, PM will reach me very likely

I like your very practical advice here, Wim. For Mike, as Wim noted
in another post, you can simulate this in Spice quite easily. While
you're playing with it in Spice, you might also look at two high-Q
tanks, tuned to the same frequency, with a non-zero coupling
coefficient between the inductors. Excite the first with a voltage
generator in series with the RLC tank, or a current source across it,
and observe the frequency response at the second tank. Vary the
coefficient of coupling between the coils and notice how small it is
to get critical coupling. If you use LTSpice, you can use a .step
statement to run a set of simulations over a range of k values, for
example.

The flip side of this is that in LC filters that are very narrow-band
that you design assuming no coupling among the resonators, expect to
have to work some to insure that there really is no coupling among
them in your implementation! Sometimes it gets difficult to shield
well enough between adjacent resonators to get the performance you
want.

Cheers,
Tom

On Nov 15, 1:59*am, "amdx" wrote:
Hi Guys,
*Assuming I have a tank circuit on a crystal radio with a Z at resonance of
1.5 megaohms.
How would I make an antenna and extract maximum signal and keep the Z at
750,000 ohms.

I did take part in the discussion of matching, but I am afraid you are
off by several orders of magnitude regarding the impedance value.

Say you shoot for resonance at 1 MHz: a 2.5 mH inductor (already
impractically large) and a 10 pf capacitor (already impossibly small)
will resonate close to 1 MHz but have an impedance of 16 kilohm or so,
100 times lower than your assumption.

For 10 MHz, it would be 1 pF and 250uH, again for 16 kohm.

In practical circuits the impedance is typically in the order of
300-2000 ohm.

A typical old-style mediumwave tuning capacitor has a maximum
capacitance of 360 pF to tune about 520 kHz at 850 ohm with a 260 uH
coil.

Assuming you use a much smaller than usual 100pF variable, say 130 pF
including tray capacitance in the coil, at the bottom of the band you
will need 760hH, and get 2.5 kiloohm.

On shortwave, let's say yo want to tune a lowest frequency of 6MHz
with only 60pF all included - again, nobody does that - you need 12uH
inductance but only get 450 ohm impedance.