> Jeff, you aren't stepping on my toes at all. Glad to see your
> comments.
OK, good. Since you've never met me, I can assure you, you definately DO
NOT want me stepping on your toes, it would be painful.
> I do have to agree with Kevin that most duplexer
> manufacturers recommend different cable length trials between
> the transmitter and the duplexer when full power can not be
> reached into the duplexer.
Ah, but the crux of the matter is that we're not changing the performance of
the duplexer, we're just getting the transmitter to transfer more power into
the line.
> Over the years I have been a manufacturers rep for TX-RX,
> Sinclair and Telewave. All of them recommend the same thing.
Again, it's a CYA measure as Kevin pointed out. PA won't make power? Don't
blame us, try mucking with the cable length, see if that helps.
> I am not a transmitter expert but it is my understanding that
> the problem is not one of the duplexer not presenting 50 ohms
> at the wanted frequency but the impedance that it presents
> off frequency to the transmitter finals. Some solid state
> devices do not like to see high reactance, even off
> frequency.
But why? If all of the power (or, let's hope, at least 99.999999% of it) is
on-channel, *should* a properly-designed and properly-functioning
transmitter misbehave due to the poor match a duplexer presents at
frequencies far removed from the channel center?
> For one thing the reactance causes them to draw
> more current than normal.
Again, why?
> This may be why you find that
> tuning for minimum pa current and maximum power out don't
> exactly agree with one another.
I can promise you they almost never do, but that's not any great mystery.
> You are probably finding a
> balance between the off frequency reactance and the on
> frequency wanted load that the finals see.
No, that's not it. The off-frequency Z issue is a totally separate topic
from the efficiency vs maximum output subject. Let's keep those two topics
separate for the sake of this discussion.
> If you have the duplexer properly tuned to provide 50 ohms at
> its input port, the transmitter may still not be happy
> because of the off frequency reactance presented by the duplexer.
I disagree. I would accept the notion that the transmitter may not be
"happy" (and I put that in quotes not to mock you, but becuase I can't come
up with a better word either) because it is not *properly matched* when
looking into a 50+j0 load. This indicates a deficiency in the amplifier; if
it were designed and working right, it *should* make rated power when
terminated in a 50 ohm load on-channel.
> Changing the cable length in this case really does nothing
> for the on frequency load between the duplexer and
> transmitter, when the duplexer is presenting 50 ohms, but it
> can change the off frequency impedance transformation that
> the transmitter sees.
Yes, but again, I argue that this all points back to a PA problem. Or the
input Z of the duplexer really isn't 50 ohms and the line is acting as a
transformer.
> Detuning the duplexer and or changing
> cable length to get the transmitter power up is the wrong way
> to go here. First the transmitter should be optimized into a
> 50 ohm load. Then optimize the duplexer input for 50 ohms input.
Yes, yes, yes, amen!
> Someone asked about a "rule of thumb" for transmitter to
> duplexer cable length. There is none!
Yes there is. You take out a tape measure and the distance from the
transmitter to the duplexer. You make the cable at least that length.
> The cable length between multiple cavities is predictable. As
> an example between two notch cavities; the first notch
> presents a very low impedance. With a quarter wave line to
> the next cavity that low impedance is transformed to a high
> impedance at the input to the next cavity. That high
> impedance is then presented with a very low impedance of the
> second cavity. This critical length cable increases the
> ultimate notch depth because the high impedance that the
> cable presents and the low impedance of the cavity form a
> voltage divider. The greater the ratio the better the rejection.
'zactly. When done right, you can pick up close to 6 dB additional net
notch depth when cascading notch (or pass/notch) cavities when the
intra-cavity cables are cut this way.
---- Jeff WN3A
