Can a crystal radio detect an infinitesimal load at a distant radio station?

[sophiecentaur]

To maximise the power delivered from the feeder into the antenna you need to match at the drive point - that's straightforward enough and you try to minimise your VSWR / reflection coefficient at the join.
To get power actually into the feeder from the transmitter there has to be some degree of matching. A low impedance output stage will not drive much power into a high impedance load without using massive voltages - and vice versa. But, as you imply, conway, there are practical limits to how well you can match a (probably non-linear) amplifier stage and there will be inherent losses in the device in any case. I don't think that there are any particular design problems for 'bog standard' frequency operation, these days. They seem to build big HF transmitters a bit like audio amps these days - class B push pull blah blah.
To avoid echoes, over voltage and or odd frequency response, it is necessary to match at least one end of a feeder.

 

[sophiecentaur]

Not true. Why are the reflector and director elements in a Yagi not terminated? If they were, your Yagi wouldn't work so well...

A Yagi antenna is not steerable - it is designed to have lots of current in its parasitic elements in order to to drive it in one particular direction. It has only one driven element and it is not what I was describing. To steer with two or more driven elements is the problem. Then the mutual impedance between two elements is a nuisance. You can end up chasing your tail if you want to fire in certain directions.

 

[berkeman]

A Yagi antenna is not steerable - it is designed to have lots of current in its parasitic elements in order to to drive it in one particular direction. It has only one driven element and it is not what I was describing. To steer with two or more driven elements is the problem. Then the mutual impedance between two elements is a nuisance. You can end up chasing your tail if you want to fire in certain directions.

I was referring to the Yagi in RX mode, not TX. I thought that's what was in question -- whether there is a transmitted field from a well-matched (terminated) antenna when it is receiving a signal.

 

[sophiecentaur]

I think we must be at cross purposes here. A Yagi has only one feed point (or 'driven point'), whether in transmit or receive mode and, as the two directivity patterns are indistinguishable, the relative currents in the elements must be more or less the same in each mode. The parasitics are, of course, 'short circuited' and not terminated. The Yagi only works with the elements around resonance. In the receive mode, the explanation for its operation must be that the parasitics resonate in similar relative phases to the driven element. Whether transmitting or receiving, this must take a few cycles to build up when the wave is first received or when the TX is turned on. After that, the currents in the elements are mutually set by the self impedances of the elements themselves and their mutual impedances with the other elements.
What has that got to do with the problems associated with multiple fed element antennae, though? The thread may have migrated here and there, as they do, but my comments were in response to comments about steering multi element arrays and the problems of currents flowing in elements affecting the current / input impedances of other driven elements. (The Yagi is not the only directive antenna and it certainly is not steerable - apart from by waving it around)

 

[berkeman]

I was using the example of an RX Yagi to try to make the point that you only get reflections off of unterminated antennas (shorted in this case), not off of terminated antennas (which I think is what conway has been saying. It seemed a convenient example. A Yagi with terminated directors and reflector would not be any more directional than a regular dipole.

And I brought up the example of steerable RX antenna arrays for the same reason. In RX mode, I'm not aware of any interaction between the antennas (no reflected signal bouncing between the elements and needing to be taken into account in the pattern calculation).

Sorry if I haven't been clear about why I was using these examples.

 

[sophiecentaur]

I getcha now. But the fact is that, although there are higher currents in parasitics, there are still currents in driven elements (TX or RX) and these have mutual effects. If you ignore them, your antenna just won't do what you want it to do. If it is a TX array, you can end up with transmitter matching problems as well as a wrong pattern and, if it is an RX array, the pattern goes to pot.

 

[conway]

I hate it when people start off by saying "even if you're right..." because its a big cop-out: you're not even taking a stand on whether he's right or wrong. But in Berkeman's case I don't see that I have any choice: "even if he's right" about how you steer a receiver array, it's a really indirect connection to the original question of a crystal radio, which involves a single electrically short dipole. So whether he's right or wrong I don't think he can claim to have dealt with the question of how much re-radiation there is in a tuned matched short dipole.

 

[sophiecentaur]

However short the dipole is, there will be currents flowing in it - in phase and out of phase with the PD. These currents will radiate. It will only be the real part of the current and PD that will be absorbed by the receiver. Matching the short dipole will involve other reactive components to resonate with the capacity that the short dipole represents and to transform to the receiver input impedance. In practice, you can only go so far down that road and the radiator gets less efficient due to the finite resistance of the conductors for really short radiators.
Just think how hard it is to make a 'stealth' aircraft. If it were simply a matter of making an aircraft surface look like a set of matched dipoles then they would be built like that. They all re-radiate - which is why they have to be made with 'least worst' reflecting shapes.
Very short elements won't re-radiate much - and neither will they absorb much power. Arrays of very short elements can be treated as voltage probes because there is so little scattered power. (This accords with experience - even down to the sky being blue.)
The earlier quoted 2:1 ratio of currents for unmatched and matched termination (way back in the thread) seems to have been forgotten but it should be taken into consideration.

 

[Averagesupernova]

Ya know, it doesn't really make a damn bit of difference. Any power that was reradiated had to be 'drawn in' by the antenna in the first place. For all I care every available amount of power that the transmitting antenna hundreds of miles away radiated could have been absorbed by a receiving antenna which would then have turned around and reradiated all but what we normally see at the feedpoint of an antenna.

 

[GODISMYSHADOW]

I took issue with the notion of a RX antenna "drawing power" from a TX antenna when they are well separated. Naturally, when they are close, their mutual impedance can affect the input impedance of the TX antenna (as in a Yagi type array, with parasitics) but any model which implies the "drawing out" process at great distance should work over any distance - hence, I took it a bit further to where propagation time is a significant factor. As a thought experiment I don't think it was over the top to consider ("take sides with") both RX and TX; after all, they are both involved in the process. Where was the problem with that?.

It makes me feel uneasy, because when you're at RX the latest news
you can possibly have about TX is limited to whatever transmissions
or other electromagnetic radiation you are receiving. This is the most
current knowledge you can possibly have. The same would be true
but in reverse if you were at TX. The region outside the "light-cone" is
called "elsewhere" and you cannot know about that in your "here-now"
with certainty but only as a probability. Even when you look up at the
stars, what you see is the most current knowledge available. Any more
current news would be outside the light-cone and only a probability as
far as you're concerned. Anyway, that's the reason I don't like thought
experiments where they jump around too much.

Perhaps RX and TX taken together can be considered a system and then
wait for the system to stabilize.

 

[conway]

I'm going to support Sophie on this issue. When the receiver is far away from the transmitter, it does not in any way "load down" the transmitter, and basically for the reasons Sophie said. Of course there is a very small reflected wave but that's not what we mean by loading, even in principle. The energy transfer from A to B does not depend on the detection of the reflected wave; the receiver and transmitter would work exactly the same if the reflected wave could somehow be caught and destroyed before it got back to the transmitter.

 

[berkeman]

I'm going to support Sophie on this issue. When the receiver is far away from the transmitter, it does not in any way "load down" the transmitter, and basically for the reasons Sophie said. Of course there is a very small reflected wave but that's not what we mean by loading, even in principle. The energy transfer from A to B does not depend on the detection of the reflected wave; the receiver and transmitter would work exactly the same if the reflected wave could somehow be caught and destroyed before it got back to the transmitter.

And I disagree with your reflected wave concept for a well-terminated RX half-wave dipole antenna, at least for now. We can agree to disagree, until one of us comes up with the math or an experiment to settle that, okay?

 

[sophiecentaur]

I have a feeling that there is some confusion about how an antenna works. It isn't just the metal bits out there in space that are responsible for the power going in or our. An infinitely / very thin wire has more or less the same energy gathering properties as a fat wire. (Width helps to make the bandwidth of the element wider, but that's all) So the antenna must be working due to the fields surrounding it, which can be regarded as being caused by currents flowing in the conducting bit. These fields can be regarded as coupling energy into the feedpoint of the antenna. So why should it be suggested that, just because the antenna is matched, these fields should suddenly have no effect on the passing wave? You can, in principle and in practice, feed a wire anywhere you like - not just its mid point - and have a perfect match.

I was trying to think of an analogous situation and I suggest that a light disc shaped float, floating on the sea would be suitable. If it can float up and down on the waves without taking any energy then, if it were light enough, you would get no significant scattered energy from it. However, if you attempt to take energy from it - using it to pull on a crank, for instance, the float will no longer move directly in step with the water and this will involve waves emanating from it. You could imagine an appropriate value of friction (resistance) applied to the crank that would take a maximum of energy from the system - the waves would still be disturbed; you wouldn't imagine that, suddenly the waves would go past without being affected. So taking energy has involved disturbing the waves. Coupling the float to an appropriate resonating mass / spring system (without loss - representing a resonant length of wire) could produce greater movement of the float and, hence, more scattered waves. But in both cases there will be disturbance. The only time there is none would be when the float is not altering the power flow at all.

 

[conway]

OK, I don't think the math is normally helpful in these kinds of discussion, but here goes:

Take an AM radio station putting out a signal of 377 mV/meter (that's a pretty strong signal, obviously I picked the numbers so that the power density would be ExH=377 microvolts per meter squared). Let the station be at 680 kHz, wavelength approx 440 meters. Let your antenna be an 11 meter dipole. (No ground. Assume there is no ground.)

Now let's go to Wikipedia (http://en.wikipedia.org/wiki/Dipole_antenna) and look up the radiation resistance for a short (L/Lambda=1/40) dipole. Paraphrasing slightly:

R = approx 200 (L/lambda)^2 which gives in this case 0.125 ohms. To match the radiation resistance we use an appropriate coil to cancel the reactance of the antenna and insert a load resistance equal to R. The antenna voltage is ExL = approx 4 volts and the current is therefore 4 volts / 0.25 ohms = 16 amps.

The power to the load is I^2*R(load) = 32 watts and the power re-radiated, substituding R(rad) for R(load), is obviously the same.

 

[sophiecentaur]

The sums have sorted it out, as usual. Thanks conway.

 

[conway]

As usual? It's rare that anything ever really gets sorted out in these kinds of discussions. I wouldn't write this one off yet either...

 

[conway]

I have to say I was wondering if anyone thought my numbers were unrealistic.

 

[sophiecentaur]

Near enough to make the point.

The point seems to be that matching your antenna feed is not the same as replacing an area in space with a 'hole' for energy to flow out of, which is indistinguishable from the rest of the sphere into which the power is radiating. It is more like hanging a load half way along a long, perfectly terminated transmission line. You can get a optimum level of power into the load but that will, inevitably, produce a reflection / mismatch on the line at that point.