Well, I did the honorable thing and researched the reflected power question
in the ARRL Handbook (1995 edition) and Bill Orr's Handbook (23rd edition.)
For those who care, here are the results. By the way, some questions are
raised here that are indeed relevant to the subject of high power
amplifiers (see number 3 below.)
1. On a mismatched line (SWR>1:1), what happens to the reflected power?
The ARRL Handbook covers this subject on pages 19.2-19.4. On 19.4, after
describing the initial reflection of power caused by the discontinuity in
impedance between the antenna and transmission line, the Handbook goes on
to ask:
"Now what actually happens to the energy reflected back down the line? This
energy will encounter another impedance discontinuity, this time at the
generator. Reflected energy flows back and forth between the mismatches at
the source and load. After a few such journeys, the reflected wave
diminishes to nothing, partly as a result of finite losses in the line, but
mainly because of absorption at the load. In fact, if the load is an
antenna, such absorption at the load is desirable, since the energy is
actually radiated by the antenna."
I was wrong about the reflected energy being dissipated by the finals (but
see below for a question about what does happen to the finals.) I was
right, however, that the reflected power does go somewhere: it goes out the
antenna.
2. Is it correct to subtract reflected power from forward power in order to
compute actual power output?
There was only one brief reference, in the ARRL Handbook (p 19.4):
"If a line is not matched (SWR>1:1) the difference between the forward and
reflected powers measured at any point on the line is the net power going
toward the load from that point."
So I was wrong about that, too. But then I got to wondering why my test
results seemed to contradict this (meters on both side of the antenna tuner
showed the same forward power.) Then I read the rest of the paragraph from
the ARRL Handbook:
"The forward power measured with a directional wattmeter... on a mismatched
line will thus always appear greater than the forward power measured on a
flat line with a 1:1 SWR."
I guess they're saying that the forward power meter will actually show the
sum of the forward and reflected powers.
So, the fellow who claimed that at 3:1 SWR the meter in front of the tuner
would read 1500 watts out (0 reflected) while the meter on the other side
would read 2500 watts forward and 1000 reflected, for a total of 1500 watts
net output, was right! This made me realize that years ago, when I had
different equipment, I used to see the phenomenon of apparent forward power
increasing as the SWR increased. But for a long time, I have been using
meters that must automatically subtract the reflected power from the
forward power! I'm one of the few who built John Grebenkemper's Tandem
Match back in the late 80's. I went back to the original 1987 QST article
and, sure enough, that's what the wattmeter does. I also have a Nye Viking
RFM-003 that appears to do the same thing, as does the power meter on my
Alpha 87A. I never see forward power rise with SWR on any of these meters,
and believe that it would be incorrect to subtract reflected power to
compute the output power.
So, while it is generally correct to subtract reflected from forward power,
it does depend on the type of wattmeter you are using. You have to make
sure that the wattmeter isn't already doing it.
3. What happens to the amplifier's output tubes when operated into a high
SWR?
I had a lot of trouble finding any kind of definitive answer to this
question. Orr's Handbook points out (on p 21-4) that if the generator's
output impedance just happens to be the same as the characteristic
impedance of the transmission line (say, 50 ohms looking into 50 ohm coax),
and there is a mismatch between the line and the antenna, then the
generator will absorb all of the reflected power. I imagine this would
cause the plates to get hot, but I don't know. Does anybody out there know?
(Note: when the generator's output impedance, the transmission line's
characteristic impedance, and the load impedance are all the same, say 50
ohms, you have a "conjugate match".)
But, in most tube amp circuits, the output impedance is higher, at least
100 to 600 ohms. So the reflected power just bounces off the generator back
towards the antenna. (Note, even though the amp's output impedance is not
50 ohms, it can still be designed to operate at maximum efficiency into a
50 ohm load. In this case, when the transmission line and load are at 50
ohms, you have a Z(sub zero) match, not a conjugate match.)
If this is the nominal case, then what actually happens when a high power
amplifier tube is operated into a high SWR? The ARRL Handbook points out
that most tubes are pretty forgiving of a high SWR, and that's certainly
been my experience with the 3-500Z's in my old SB-221. It's been along time
since I've used it, but I seem to remember that it wasn't bothered by a SWR
of about 3:1 or less. But I also seem to remember that above about 2:1 I
couldn't always drive maximum power.
On the other hand, the 3CX800A7's in my Alpha 87A are heavily protected
from high levels of reflected power. Operating at full power with anything
near about 2:1 or greater will trip the protection circuits. Evidently,
these tubes are not as tolerant of higher SWR values. Why? What would be
damaged?
What actually happens to the tube as SWR rises? Although some tubes seem to
be able to develop full power output with some SWR present, certainly the
ability to develop full power output must diminish as the SWR gets higher.
So, is it a case of the tube not being able to dump the power into the
transmission line and burning it up on the plate? Or is it that the tube
can't generate the power at all?
I guess a radical case would be to operate a tube into a dead short. My
guess is that the plates would melt. Or is it the cathodes? The grids?
4. "... why do we strive to get an antenna to resonate (thus 1:1) swr, why
even bother?"
Ignoring the possibility that the author of this question was pulling our
collective leg, I think it does deserve a quick answer based on what I read
in the handbooks.
All of the literature agrees that, in the HF bands, antennas can still
radiate efficiently at SWR values up to about 6:1. Once you get up to VHF
frequencies, losses in the transmission line become significant. However,
transmitters can't necessarily supply maximum power when the SWR is high.
Transistors will blow up (which is why modern rigs quickly reduce output
power when the SWR is high), and even tubes get to a point where either
maximum power can't be delivered or the tube will be damaged.
But this problem can be eliminated with a matching network that makes the
antenna appear resonant at the operating frequency. Depending on the
application, you can use tuned feeders, matching stubs, traps, loading
coils, top hats, counterpoises, L-C circuits, an antenna tuner, etc. So why
bother to make the antenna self-resonant (without all this junk)?
Well, I think there are two primary reasons. First, the more extreme the
matching job, the more likely it is that the matching network will become a
source of resistive losses. In other words, the chances are that you will
lose power as heat in the matching network. Second, as SWR rises, so does
antenna Q, reducing bandwidth. Compared to a self-resonant antenna, a
heavily matched antenna may exhibit an unacceptably narrow bandwidth,
forcing you to operate on a very limited range of frequencies or constantly
adjust the matching network.
Of course, things aren't so bad at relatively low values of SWR (say, under
10:1). A well-designed matching network can match loads like this with
minimal resistive losses and still leave you with a reasonable bandwidth. I
would guess that a large number of amateur antennas are not self-resonant
at the operating frequency and employ some kind of matching system.
Of course, trying to load a coathanger on 160 meters is asking for trouble.
73, Dick, WC1M
--
FAQ on WWW: http://www.contesting.com/ampfaq.html
Submissions: amps@contesting.com
Administrative requests: amps-REQUEST@contesting.com
Problems: owner-amps@contesting.com
Sponsored by: Akorn Access, Inc. & N4VJ / K4AAA
|