> If your duplexer presents 50+j0 at its input at the operating
> frequency and
> you are using 50 ohm cable to connect it to the transmitter then the
> transmitter is always going to see 50+j0 at the operating frequency no
> matter what the cable length is. But at some off frequency
> that is not 50+j0
> that impedance is going to get transformed into something yet
> again by the
> time the cable reaches the transmitter. And depending on how
> long or short
> you make that cable will depend on what transformation it
> will make to that
> off frequency impedance all the while still delivering 50+j0 at the
> operating frequency.
> So cable length can be used to change unwanted reactance of
> off frequency
> things.
Of course. But if you stuck a Z-matcher between the duplexer and the PA,
then what started out as 50+j0 on-channel is not going to be 50+j0 at the
PA. That was my point - the Z-matcher - not cable lengths.
> Even if the duplexer is not a perfect 50 ohms, changing cable
> lengths can
> have enough of an effect on off frequency things to make the
> system work
> sometimes.
The "sometimes" is what troubles me...
> Transistors are strange animals. The do not equate to tubes in there
> operation especially when broad band combining devices are used at the
> output of the transistor amplifiers as most do. The devices
> themselves do
> strange things in the presents of reactance. That's mostly why many
> amplifiers have swr shutdown circuits or power roll back.
Yeah, but those foldback circuits usually rely on a directional coupler to
monitor reflected power. If all of the energy (well, almost all of it
anyway) is being produced on-channel, the directional coupler is going to
sense nothing and the transmitter is going to run full bore and be happy.
If the PA were to go spurious, and all of the off-channel garbage came back
down the line, then it would fold back. But if the PA's clean, and it's
looking into a well-matched load, there should be no reason to fold back.
> problems are
> generally not one of excess current because of reflected power but the
> excess current the devices draw is from the unwanted
> reactance that they
> see.
You're talking about reactive currents flowing in the matching networks I
presume. But those reactances are present *inside* the PA, before the point
where it is connected to the outside 50+j0 world.
Most bipolars will operate into an infinite VSWR without giving up the
ghost. Excess reactive currents within the PA circuitry will manifest as
higher current draw, which is yet one more reason why I always recommend
watching current consumption when doing any kind of matching trickery.
> Sometimes a duplexer just doesn't present a perfect 50 ohms to the
> transmitter either. Some transmitters will start cutting back
> the power with
> little reflected power. Others may not be able to put out all
> the power it
> is supposed to if it does not see a perfect 50 ohms. These
> are some of the
> reasons most duplexer manufacturers recommend trying
> different cable lengths
> between the duplexer and transmitter.
You can't "convert" to a perfect 50 ohms using cable lengths if the load
isn't already 50 ohms. I've said before, and I'm saying it again. If your
duplexer 50 ohms load, you can pull all of the 50 ohm cables you want out of
your bag and you'll never get it back to 50 ohms at the PA.
But if the duplexer is designed, built, and tuned right, it should be very
close 50 ohms. I can't think of any duplexer I've ever tuned up on a VNA
that didn't have at least 20 dB return loss at the pass frequency, with many
often being 30 dB or more. Of course, that's only as good as your antenna
load, but let's keep that out of the equation for now since we're talking
about matching between the PA and duplexer only.
> All transmitters are not created
> equal.
Amen to that brother.
> I do agree with you that gross problems that show up are
> signs of problems
> that need to be fixed other than changing cables.
See, we do agree on something!
--- Jeff
