On Thu, 27 Oct 2005 12:23:46 -0400, TRABEM wrote:
I'm not sure what the 'abysmal match to the antenna whose Q will be
buried in potter's field' statement is about. With my antenna being in
the 2 ohm impedance range, and the receiver being at 10 ohms (I'll use
your figure), how can the match be abysmal?

Even allowing for the values you offer (they are wrong) you have a 5:1
mismatch. Your actual mismatch is Q times that.

Granted, it's not anywhere
near ideal. Have you assumed I was using a parallel tuned loop?

Tom has already carried the water describing what your antenna Z looks
like. It is orders of magnitude greater by virtue of unloaded Q. You
want to aspire to make your receiver match that as closely as possible
- this design does not. What you have is a heavy, heavy load that all
but wipes out the advantage of Q.

Since power transfer is the goal, and the antenna has a lot of ability
to reject out of band signals,

Not any more.

it was my hope to use the antenna
itself as the only (purposely) tuned circuit in the system. Wouldn't
converting the antenna impedance to a more traditional 50 ohms with a
toroid, and then having a second toroid to convert it back down to 10
ohms also be destructive to the antenna Q and lossy?

2 Ohms, 10 Ohms, 50 Ohms are all still very trivial in comparison to
what the antenna has to offer. At the danger of introducing an
analogy, you have a 1.2 liter high performance race car which needs to
turn 14K RPM for 300 HP and you've put it to hauling a 40 foot fifth
wheel trailer. You are not even going to pull that load a foot before
you burn out the clutch.

Although
ideally, each switch should have it's own 1 ohm variable resistor for
absolute best results...the purpose being to swamp out the dynamic
switch series resistance differences.

You are arguing precision at the wrong end of the scale. Any
additional resistors are burning signal up and any appeal to
technicalities has been lifted from other applications that don't even
come close to this situation. I've been designing with these switches
for 25 years, and for very small signals.

Note the receiver has no rf amp, it isn't needed. The gain is provided
by some low noise op amps, and no rf stage is needed. The QRP2001
receiver is designed for 100 KHz to 30 Mhz, but it is only rated down
to 1.8 MHz. However, it's worst case sensitivity is .4uV for 10 db
sinad.

This is comparing apples and donuts when the menu only offers steak.

First, all these glowing accounts of excellent performance come from
European sources where VLF is far more common, and those services pour
up to a MW into the air. Your fillings would work just as well.
Second, your glowing reports are about HF characteristics with
conventionally sized antennas. When you attempt to extrapolate this
to VLF, you are not carrying the decimal point of inefficiency to the
left as you drop down in frequency.

Again, I think you've assumed it would be fed with a parallel loop
resonant antenna.

You haven't described anything else, and no appeal to series resonant
is going to resolve Q going down the toilet. Loss is loss no matter
what topology. This was the point of discussion with ESR.

And, no active components are needed for outstanding performance.

And yet you were the first to offer some builders have experienced
miserable results. You are about to join that pouting crew. On the
other hand, as I've suggested, you may still get WWVB with all these
problems - even wrist watches do. In that eventuality you have no
real basis of comparison and your only feeling would naturally be one
of wonder and awe. An aw shucks glow in the eyes does not translate
to a marvelous DX receiver.

Again, I think you're trying to match a 2 K parallel tuned loop to the
relatively low impedance of the receiver input. I noticed you said
'follower'. Which, I think means unity gain, but is used for impedance
matching.

Exactly.

Any active component before the audio op amp is STRONGLY
DISCOURAGED in this type of receiver. This includes back to back
diodes as the receiver switches can handle 4v p-p. It also includes
varicap tuning diodes. A front end rf amp should be avoided at all
costs, it can only degrade the performance of this type of receiver.

This only applies in the face of strong signals being applied to such
an amp. Yes, you have guaranteed that with the abysmal match and all
these Cassandra forecasts come true.

The receiver has incredible immunity to strong out of band signals,

This comes only from Q. This is a baseband receiver which means it is
open to all frequencies. Thus the necessity of a hi Q passive front
end (you killed it). What you call immunity is a product of dynamic
range capability and what circuits that follow this detector.

The nature of the beast is that the
quadrature detector cancels them out by (effectively) NOT reinforcing
them.

No, you've gotten very poor instruction on the qualities of this type
of detector. I was designing them 35 years ago and they are used in a
bajillion TVs. Absolutely every one of them has front end
electronics.

The detector neither cancels nor reinforces, it provides a phased
output. The detector also has its points of failure too, but when all
the necessary pre-conditions are met, it has many more features and
immunities. You have described none of these.

What you don't see on the schematic is the
incredible sound of the receiver audio which is clean and crisp...it's
not quantifiable by bench measurements however.

This is simply absurd. The very qualities you describe are measured
every day and are the purpose of this style of detection's use.
However, as you describe them, you still give the appearance of not
knowing what to do with the "I" and "Q" channels. Therein lies the
difference.

I would also offer, that among all your attached references, much less
your own discussion, absolutely nothing is said about the "I" and "Q"
channels. These outputs (why two?) are pushed into a black box, and
one AM signal emerges which begins to argue: what is being detected?
and where? I especially like these ace buster questions because it is
overwhelmingly obvious that no one actually knows what the "I" and "Q"
channels are for. They can be put to work without any more fuss than
amplifying them, but instead they are pushed into equations, software
and black boxes.

What is worse, I have yet to see anyone actually offer what forms of
modulation can be detected - there is a serious gap of research in all
these articles you've offered. For so many that have come here to
breathlessly announce the miracle of the Tayloe mixer, to a person,
they don't even know a fourth of what could be done.

I've put some links to web references of this technology at the end of
the message.

All very nice commercials, and one offers the nuts and bolts of the
practical detector that must have missed your attention as it
contradicts with:
"First, the RF input signal is bandpass filtered and applied to
the two parallel mixer channels."

In fact, and as I've experienced through 30+ years of their design,
there are filters all around.

Your having snubbed the antenna Q violates this first premise. Then
we look at the "I" and "Q" channels, the only way to achieve what you
describe as the marvelous characteristic of
The receiver has incredible immunity to strong out of band signals
is achieved by conventional filtering just like in the superhet. In
fact one of your references employs a nine pole Butterworth. Another
design has cascading filters out to eleven poles. This being a
baseband converter has simply shoved the filtering into the AF band.
Out of band performance only applies for those who cannot hear
dog-whistles. In short, a technological shell game. Effective,
certainly, but not unheard of - direct conversion was the original
form of receiver.

Going further, the QEX article accurately describing the Tayloe
detector describes the purpose of the capacitors in the circuit you
sent me. Problem there is that your lowered RC constants are running
out at 1µS for samples being taken at a much slower rate. Result is a
serious droop is occurring.

schematic I sent you by email.

Didn't get it. My Kill filters barely let your last schematic
through.

73's
Richard Clark, KB7QHC