I know that many people think G3LHZ is a little bit off his rocker, but out of
curiosity... what he suggests on slide 15 he
http://frrl.files.wordpress.com/2009...heuristics.pdf -
- is that a valid approach to measuring antenna efficiency? -- Use a thermal
camera to note how much an antenna heats up with a given input power, find out
how much DC power it required to heat it to the same temperature (the
antenna's loss), and -- poof! -- antenna efficiency = (input power-loss)/input
power?

What are the significant loss mechanisms that he's not accounting for? (He
claims his matching network isn't getting at all hot.)

Thanks,
---Joel

On 23 mar, 02:24, "Joel Koltner" wrote:
I know that many people think G3LHZ is a little bit off his rocker, but out of
curiosity... what he suggests on slide 15 hehttp://frrl.files.wordpress.com/2009...f-small-an...-
- is that a valid approach to measuring antenna efficiency? -- Use a thermal
camera to note how much an antenna heats up with a given input power, find out
how much DC power it required to heat it to the same temperature (the
antenna's loss), and -- poof! -- antenna efficiency = (input power-loss)/input
power?

What are the significant loss mechanisms that he's not accounting for? *(He
claims his matching network isn't getting at all hot.)

Thanks,
---Joel

Hello Joel,

As with many questions, the answer to the temperature rise method can
be "yes" or "no".

For this discussion I assume an antenna as a device or system to emit
radio waves to certain directions. Mostly designers try to maximize
radiation intensity over electrical input ratio, or total radiated
power over electrical input power.

For situations where obstacles are wavelengths away from the antenna,
temperature rise can be a means of evaluating antenna efficiency. I
once used temperature rise to accurately measure efficiency of a high
efficiency RF amplifier.

In cases where obstacles are very close to the antenna, just
determining temperature rise of the metallic structure being the
antenna does not satisfy me. You will know the dissipated power inside
the antenna, but not inside the obstacle in the reactive field. When
this obstacle dissipates 90% of the electrical input power, overall
efficiency will not be high.

By using the temperature rise of the antenna only, you will notice
higher efficiency when the (loop) antenna is closer to an obstacle
(for example a thick wall). In case of a loop, the Q-factor drops,
resulting in less reactive currents, hence less dissipated power in
the loop and tuning capacitor. Of course more power is dissipated in
the wall.

When the antenna is close to metallic structures with certain
geometry, the real efficiency (so Prad/Pelec) may increase. The large
structure may extract energy from the loop and reradiate it (instead
of converting into heat). The extraction of energy from the loop
results in lower Q-factor, hence less heat loss in the loop and tuning
capacitor. Theoretically spoken, the temperature rise method is a
good one.

How will you relate temperature rise of arbitrary structures to
dissipation? If this question remains unanswered temperature rise
method will also not solve the antenna efficiency question.

Regarding temperature rise methods in general, it is good way to find
where losses are and whether it is worth to do some redesign to lower
losses.

Best regards,


Wim
PA3DJS
www.tetech.nl
PM will reach me, but don't forget to remove abc.

Joel Koltner wrote:
I know that many people think G3LHZ is a little bit off his rocker, but
out of curiosity... what he suggests on slide 15 he
http://frrl.files.wordpress.com/2009...heuristics.pdf
- - is that a valid approach to measuring antenna efficiency? -- Use a
thermal camera to note how much an antenna heats up with a given input
power, find out how much DC power it required to heat it to the same
temperature (the antenna's loss), and -- poof! -- antenna efficiency =
(input power-loss)/input power?

What are the significant loss mechanisms that he's not accounting for?
(He claims his matching network isn't getting at all hot.)

Thanks,
---Joel

A thermal camera is NOT a good way to do calorimetry. It's a fine way to
look for hot spots. Here are some of the potential problems: 1) the
thermal camera converts long wave IR brightness to temperature using
some assumptions about the emissivity of the surface; 2) convective and
radiative losses to the surroundings will change the surface
temperature; 3) surface temperature may or may not correlate well to
dissipated heat.

It's an RF device, so the physical distribution of the power dissipation
will be different than with DC. In a classic substitution RF power
measurement, a lot of effort is made to try and make sure that the
thermal properties are identical for the DC and RF dissipation cases
(well defined broadband load that is physically small, etc.).

In the subject case here, think of this: say you had a 2cm diameter
copper bar and you run 100 Amps of DC through it. The current is
distributed evenly, as is the power dissipation. Now run 1 MHz RF
through that same bar. The skin depth is about .065 mm, so virtually
ALL the RF current is contained within a layer less than 1/3 mm thick.
That's a very different heat and thermal distribution (sort of like the
difference between putting that thick steak in the 200F oven and
throwing it on the blazing hot grill).

One can calibrate for all this, but, still, it's tough.

A better way to do this measurement is to put the antenna in a suitable
far field test site, accurately measure the power flowing into it,
accurately measure the power flowing out of it (e.g. E & H field
strengths in the far field)

Now, finding a suitable site is difficult, particularly at lower
frequencies: you want to be "many" wavelengths away from the ground, for
instance.

How about hanging it from a balloon with a battery powered transmitter
(or receiver: I assume nobody is claiming that reciprocity doesn't work)
and have the field strength detector also hanging from a balloon.

Hi Jim,

Thanks for the thoughts; I hadn't thought of many of the additional loss
mechanisms you mention.

"Jim Lux" wrote in message
...
In the subject case here, think of this: say you had a 2cm diameter copper
bar and you run 100 Amps of DC through it. The current is distributed
evenly, as is the power dissipation. Now run 1 MHz RF through that same bar.
The skin depth is about .065 mm, so virtually ALL the RF current is
contained within a layer less than 1/3 mm thick. That's a very different
heat and thermal distribution (sort of like the difference between putting
that thick steak in the 200F oven and throwing it on the blazing hot grill).

If you're just looking at surface temperature (i.e., with a thermal camera),
will it take more or power at 1MHz to obtain a given surface temperature
increase than at DC?

At DC, since you're heating up the entire bar, and the only way for the heat
to go is up "out" to the surface... I'm thinking... less power is needed for a
given rise?

That would certainly then overestimate antenna efficiency.

---Joel

Joel Koltner wrote:
Hi Jim,

Thanks for the thoughts; I hadn't thought of many of the additional loss
mechanisms you mention.

"Jim Lux" wrote in message
...
In the subject case here, think of this: say you had a 2cm diameter
copper bar and you run 100 Amps of DC through it. The current is
distributed evenly, as is the power dissipation. Now run 1 MHz RF
through that same bar. The skin depth is about .065 mm, so virtually
ALL the RF current is contained within a layer less than 1/3 mm thick.
That's a very different heat and thermal distribution (sort of like
the difference between putting that thick steak in the 200F oven and
throwing it on the blazing hot grill).

If you're just looking at surface temperature (i.e., with a thermal
camera), will it take more or power at 1MHz to obtain a given surface
temperature increase than at DC?

At DC, since you're heating up the entire bar, and the only way for the
heat to go is up "out" to the surface... I'm thinking... less power is
needed for a given rise?

That would certainly then overestimate antenna efficiency.


And one would need to be careful about when you've reached thermal
equilibrium (if ever)

On Mar 22, 9:24*pm, "Joel Koltner"
wrote:
I know that many people think G3LHZ is a little bit off his rocker, but out of
curiosity... what he suggests on slide 15 hehttp://frrl.files.wordpress.com/2009...f-small-an...-
- is that a valid approach to measuring antenna efficiency? -- Use a thermal
camera to note how much an antenna heats up with a given input power, find out
how much DC power it required to heat it to the same temperature (the
antenna's loss), and -- poof! -- antenna efficiency = (input power-loss)/input
power?

What are the significant loss mechanisms that he's not accounting for? *(He
claims his matching network isn't getting at all hot.)

With some feedlines and frequencies, feedline radiation can become an
issue. For example, using 4" ladder line at UHF.

I think his method, especially for physically compact antennas and
feed systems which tend to have very low radiation resistance at HF
frequencies, is a great check on theoretical calculations. There has
to be a meeting point between mathematical models/NEC and reality and
he is working at one such point. There are of course other points too
(e.g. near field and far field measurements).

Tim.

Hi Tim,

"Tim Shoppa" wrote in message
...
On Mar 22, 9:24 pm, "Joel Koltner"
wrote:
I think his method, especially for physically compact antennas and
feed systems which tend to have very low radiation resistance at HF
frequencies, is a great check on theoretical calculations. There has
to be a meeting point between mathematical models/NEC and reality and
he is working at one such point.

Agreed -- the controversy comes into play in that he ends up computing
electrically-small loop antennas as being upwards of 70-90% efficient, when
everyone "knows" that such antennas are typically 10% efficient. He even
goes after Chu/Wheeler/McLean/etc. in suggesting that the fundamental limits
for the Q of an ESA are orders of magnitude off (slide 47), and that's pretty
sacrosanct terriority (see, e.g., www.slyusar.kiev.ua/Slyusar_077.pdf -- even
the Ruskies buy into the traditional results :-) ).

Hence, while I don't really have the background to know precisely how much of
what Underhill promotes is true or not, it's definitely intriguing to me, and
I'm looking around for various rebuttals by those more skilled in the art than
I am.

One link I found: http://qcwa70.org/truth%20and%20untruth.pdf (but this was
written before the PowerPoint presentation I originally linked to).

---Joel

Joel Koltner wrote:
Hi Tim,

"Tim Shoppa" wrote in message
...
On Mar 22, 9:24 pm, "Joel Koltner"
wrote:
I think his method, especially for physically compact antennas and
feed systems which tend to have very low radiation resistance at HF
frequencies, is a great check on theoretical calculations. There has
to be a meeting point between mathematical models/NEC and reality and
he is working at one such point.

Agreed -- the controversy comes into play in that he ends up computing
electrically-small loop antennas as being upwards of 70-90% efficient,
when everyone "knows" that such antennas are typically 10% efficient.
He even goes after Chu/Wheeler/McLean/etc. in suggesting that the
fundamental limits for the Q of an ESA are orders of magnitude off
(slide 47), and that's pretty sacrosanct terriority (see, e.g.,
www.slyusar.kiev.ua/Slyusar_077.pdf -- even the Ruskies buy into the
traditional results :-) ).


One wants to be careful about "Q" and Chu, etc. If you haven't
actually read the paper, you might think that Chu is talking about Q as
in filter bandwidth (e.g. center frequency/3dB bandwidth), but it's not.
It's the ratio of energy stored in the system to that radiated/lost.
For some systems, the two are the same, but not for all.

"Jim Lux" wrote in message
...
One wants to be careful about "Q" and Chu, etc. If you haven't actually
read the paper, you might think that Chu is talking about Q as in filter
bandwidth (e.g. center frequency/3dB bandwidth), but it's not.

I read it well over a decade ago. I like to think I've learned a fair amount
since then, so I should probably go back and do it again some time...

I had McLean as a professor as an undergraduate -- he was already ruminating
about Chu not having the full story back in the early '90s, several years
prior to his (apparently pretty regularly referenced) paper on the topic on
'96
(http://www.physics.princeton.edu/~mc...44_672_96.pdf).
(He was also a fan of Goubau antennas and wanted me to help him figure out
just how they worked... I never managed to contribute anything of use towards
that end and graduated and moved, but I did visit him a few years later at
which point he told me it'd really been rather more difficult to figure out
then he'd first thought. Harumph! I do think it's cool that it eventually
ended up on a cover of a book:
http://www.amazon.com/Electrically-S.../dp/0471782556 )

It's the ratio of energy stored in the system to that radiated/lost. For
some systems, the two are the same, but not for all.

Something like... it's exactly true of a simple RLC network (2*pi*total stored
energy/energy lost per cycle)... but one can concoct fancy, higher-order
networks where it isn't exactly correct?

---Joel

Joel Koltner wrote:
"Jim Lux" wrote in message
...
One wants to be careful about "Q" and Chu, etc. If you haven't
actually read the paper, you might think that Chu is talking about Q
as in filter bandwidth (e.g. center frequency/3dB bandwidth), but it's
not.

I read it well over a decade ago. I like to think I've learned a fair
amount since then, so I should probably go back and do it again some
time...

I had McLean as a professor as an undergraduate -- he was already
ruminating about Chu not having the full story back in the early '90s,
several years prior to his (apparently pretty regularly referenced)
paper on the topic on '96
(http://www.physics.princeton.edu/~mc...44_672_96.pdf).
(He was also a fan of Goubau antennas and wanted me to help him figure
out just how they worked... I never managed to contribute anything of
use towards that end and graduated and moved, but I did visit him a few
years later at which point he told me it'd really been rather more
difficult to figure out then he'd first thought. Harumph! I do think
it's cool that it eventually ended up on a cover of a book:
http://www.amazon.com/Electrically-S.../dp/0471782556
)

It's the ratio of energy stored in the system to that radiated/lost.
For some systems, the two are the same, but not for all.

Something like... it's exactly true of a simple RLC network (2*pi*total
stored energy/energy lost per cycle)... but one can concoct fancy,
higher-order networks where it isn't exactly correct?



or, an antenna, for which the approximation of an RLC is only true in a
limited frequency range.

There's a fairly good literature out there about the limitations of Chu
(after all, he was only the first shot, and modeled it as a single
spherical mode). Harrington was the next bite at the apple, and then
there's a whole raft, particularly when you get into superdirective
arrays or antennas/systems which have non-reciprocal devices in them.
R.C. Hansen and McLean (as you note) are others. When you start talking
about antennas directly coupled to active devices, that's another thing..

Consider that the low impedance of a small loop is a good "match" to the
low output impedance of semiconductor devices in RF applications.. Now
you've got a reactive load hooked to a reactive source.