Cable impedance upper limit in Radcom Plus?

[G8HQP Dave]

RSGB members may have seen the latest Radcom Plus e-magazine. In it there is an interesting article on transmission lines, but I am not sure I believe everything it says. The author says that the Heaviside Telegrapher's Equation does not tell the whole truth about transmission lines (true - it only applies in the low frequency approximation of EM, otherwise known as circuit theory) so there are other modes apart from the usually dominant TEM (true - but usually irrelevant until you get to very very high frequencies).

The contentious part is where he claims that no transmission line can have a higher characteristic impedance than free space (i.e. 377R), and that if you calculate a higher one assuming TEM then the other modes adjust it downwards to get below this limit. He claims that this explains anomalies seen in precision RF labs. He also says that this new theory explains the 'single wire' feeder.

If 377R is an upper limit then I would expect some anomalies to be seen with a 300R line, and some people claim to have used 450R open wire feeders. I am very sceptical about this particular claim, especially as he does not explain it but merely refers us to other publications - which may or may not say what he thinks they say.

What do others think?

[G3PIJpeter]

According to the open-wire line impedance calculator at
http://www.smrcc.org.uk/tools/OpenWire.htm
1mm radius wires (i.e. 2mm diameter) spaced at 100 mm (aka four inches) gives a characteristic impedance of 552R - as near as 'dammit' to the 600-ohm line that we all used pre-1960s.

- Peter G3PIJ

The GPO used 600ohm overhead wires. Also the classic "Z of a twin feeder" formula has no limits, gets rather large at over a few thousand ohms though.

[G8HQP Dave]

Yes, but that is his very point: if you use the traditional equations you can easily get higher figures, but reality is different (according to him). All it would need to prove him wrong is someone carefully measuring a line at more than 377R - and I suspect that lots of people have done that over the years but not made a fuss because nobody had told them that this is impossible.

[Hartley118]

I guess that much depends on frequency. When I was involved in the design and manufacture of coax cable for the military and for mobile telecoms, it was always asserted that the signal (normally 100- 3000 MHz) actually travelled in the dielectric, the conductors merely acting as boundaries. I guess that model would fit with the maximum possible characteristic impedance being the 377 ohm free space value.

On the other hand, in the audio industry, we regarded the ‘standard’ 600 ohm impedance figure as a quaint and largely irrelevant hangover from the ancient days when the only audio was telephony and their long wires were spaced 9 inches apart in free air. In fact, the most useful impedance model of screened audio cable is a capacitor - easy to deal with.

I suppose that classical ‘textbook’ transmission line theory lies somewhere between these two extremes.

Martin

[Radio Wrangler]

A piece of 600 Ohm line is still 600 Ohm line, whatever its length.

It looks to its driver like a 600 Ohm resistive load right from the instant it is driven. no need for the delay for a signal to get to the far end and back.

It looks to its termination like a 600 Ohm resistive thevenin/norton source. If accurately terminated, there will be no reflection here.

If you can make a 600 Ohm line good for tens of kilometres, you can make one good for thousands. Only the losses and the overall delay scale up.

The calculation of cross-section dimensions is where things go a bit squiffy when you start pushing to extremes because if you make coax with an almost infinite diameter outer, the C per unit lrngth doesn't approach zero. A standard student exercise is to calculate the capacitance of a lone sphere.

Capacitance to what? I hear. It puzzled me too. Capacitance to everything! Whatever its initial potential with respect to anything unchanging, its potential increases by 1 volt per coulomb divided by its capacitance, if you pump charge onto it.

So infinite spacing can't give you Z tending to infinity. Infinitesimal diameter seems a better bet.

David

[russell_w_b]

We used 328 Ohm Zo balanced lines on HF broadcast transmitters, and there would be matching sections inserted of 400 Ohms Zo or so, maybe 440 Ohms Zo on the phasing verticals to the end-fed dipoles, all which actually matched the transmitter to the aerial, satisfying the 'traditional' line equation, the Smith chart, and any VSWR / reflection coefficient metering. And because it all works, theory translating nicely to practice, there is no reason to doubt what we were taught all those years ago.

So have we missed something?

[G8HQP Dave]

He claims that infinitesimal inner diameter does not get you there either. I can't remember why.

The problem with this type of article is that the details of the actual technical claims are either missing or expressed so vaguely that it is difficult to refute. A clear argument gives critics the opportunity to pinpoint the precise point in the argument where a mistake has been made, but 'trust me, I have read the original papers and they agree with me' is not very helpful.

I will continue to assume that he is wrong, unless someone can come up with good reasons why I should believe him. Genuine new theories in radio engineering are more likely to be published by IEEE or IET than RSGB or ARRL.

[GrimJosef]

Originally Posted by Hartley118

... When I was involved in the design and manufacture of coax cable for the military and for mobile telecoms, it was always asserted that the signal (normally 100- 3000 MHz) actually travelled in the dielectric, the conductors merely acting as boundaries ...

It's true that the signal sets up an electromagnetic field in the dielectric and if the geometry is simple enough for you to do the sums then you can work out the power being transported down the line by integrating the Poynting vector over the dielectric's cross-section. However current also flows in the conductors and if they're resistive then voltage can be dropped along them (they can get warm, after all). Perhaps the field contributions due to these currents and voltages account for the line having a higher Z than that of free space ?

Cheers,

GJ

[Radio Wrangler]

Measuring Zo is quite routine.
Making higher Zo than 377 Ohms is also quite routine

So the universe thinks he's got it wrong

Many years ago, RADCOM published an article on a 'toroidal antenna' with an academic prof's name on it. The ARRL asked me if it was a spoof article... It went out in an April edition. I said I had no inside knowledge, but just looking at it, choosing an inductor topology optimum for minimising leaked field must be the very antithesis of the intent of an antenna, so I reckoned it had to be a leg-pull.

It wasn't!

You have to be wary about stuff in magazines, even peer-reviewed ones. The occasional whoopsies do get through. 'scuse me, I have to get the cold fusion reactor going now, I need it to power my N-Ray generator.

David

[dsergeant]

Only skimmed through the new RadCom Plus (after the rather tedious process in getting a PDF version downloaded, but I guess that is off topic for here). I must sit down and have a read of that article, yes 600 ohm transmission lines most definitely exist!

73 Dave G3YMC

[G8HQP Dave]

Yes, I have seen antenna nonsense written by a prof in a real peer-reviewed journal so I guess I should not expect too much of RSGB. The problem is that some people believe what they read.