About Us

August 11th 03, 11:25 AM

(Tom Bruhns) wrote in message om...
(Dr. Slick) wrote in message . com...
...
What do you mean by "calibrated to the line"?

The SWR meter should read zero reflected power when connected to a
load whose impedance is equal to the line's. Does it? If not, it's
not properly calibrated. Putting it another way, what's the
directionality of the bridge?

You mean the directionality of the couplers inside the meter? I
don't know.
You can adjust the needle on the analog Daiwas to read zero watts with
nothing hooked up to them.

...

Not so much a surprise as a disappointment! A difference of 70
watts incident power is totally unacceptable with only 8 feet of coax
length added.

But the load presented to the amplifier is totally different in the
two cases, most likely. Only if the amplifier's output impedance were
the complex conjugate of the line's, and the line were lossless, and
the amplifier behaved as a linear time-invariant system would you
(should you) expect the power to remain unchanged.

I'm not sure i agree with your statement as you have written it.

The power would remain the same only if the load and the line
were perfect 50 Ohms and also lossless, then the length of the coax
should not matter at all.

This is certainly not the case on our bench, which is a problem,
because i want to have an approximation of how it will perform
attached to the antenna before i ship it out to the customer. A
difference of 70 watts is not acceptable.

Tam brought up the suggestion that i try a really long piece of
RG-58 from meter to dummy load, to make the Cantenna more like 50
Ohms. Gonna try it.

Or just tune the load to zero reflected power after you're sure the
meter is properly calibrated...but that can be a catch-22 situation
that the really long RG-58 can help with. Be aware, though, that
"50-ohm" line seldom is -- it can be off 5 ohms or more. It's a cause
of some consternation to those of us involved in calibration of
precision RF test equipment.

Oh, definitely, especially the cheap stuff. But if the RG-58 went
around the world a couple of times, the end could be open or shorted,
and you would still measure 50 Ohms, most likely.

Load-pull techniques are commonly used to characterize RF source
impedances. You make known incremental changes to the load, and
deduce from the change in output power what the source impedance is,
assuming it's a linear time-invariant system. Note that adding length
to a mismatched line is one way to make an incremental change to the
load... I wouldn't necessarily say that the output impedance of a
class C amplifier is meaningless, but it may well not be constant for
all loads. and may depend on parameters you'd have trouble controlling
from day to day.

Cheers,
Tom

Right, they did plenty of load-pulling on FETS at my former place
of employment. I think they may have been more meaningful for class A
linear CDMA PAs. But finding the max power transfer impedances with
small signal class A measurements doesn't guarantee that these will be
the optimum impedances the transistor wishes to see in class C mode.

But i don't think there is a practical way to measure the S22 of a
high powered Class C amp. Most companies don't really seem to care
how close or far from 50 Ohms it really is, all they care about is the
%PAE and Pout and harmonics dBc going INTO a 50 Ohm system. But even
ensuring that your bench will match your customer's bench is a
difficult thing.

Slick

August 11th 03, 05:54 PM

(Dr. Slick) wrote in message . com...
....
You mean the directionality of the couplers inside the meter? I
don't know.
You can adjust the needle on the analog Daiwas to read zero watts with
nothing hooked up to them.

:-) The meter-needle zero generally not a very big part of the
problem! Yes, it's the coupler (bridge) inside that you want to make
sure is calibrated for your reference impedance. If it's not
adjustable, at least learn what it is calibrated to, or how good it is
for your reference impedance.

....

But i don't think there is a practical way to measure the S22 of a
high powered Class C amp. Most companies don't really seem to care
how close or far from 50 Ohms it really is, all they care about is the
%PAE and Pout and harmonics dBc going INTO a 50 Ohm system. But even
ensuring that your bench will match your customer's bench is a
difficult thing.

So you've come back around to what some of us here have been saying
for years. A transmitter or amplifier may be designed to work
optimally into some particular impedance (say 50 ohms), but have a
source impedance vastly different from that. Whether it's a
measurable impedance or not is moot: what matters is that you present
a reasonable load to it. There ARE cases where a matched source
impedance is important, but they are rare in ham transmitter work.

The easy way to insure proper operation is to present the
transmitter/amplifier with the right load. A properly calibrated SWR
meter (or as Reg would have it, a "transmitter loading indicator")
will tell you when you have such a load. Yours is telling you that
you do not. In such a case, I'd work on a match to the antenna (or
the Cantenna or whatever) first--well, right after making sure the
meter was telling me the truth or something close to it. When you get
that correct load, then the length of coax won't matter noticably,
assuming the SWR meter is calibrated to the line impedance.

You generally don't have control of your customer's bench, but you can
specify some particular load impedance and then make sure you test
with that, and leave it to the customer to insure that they provide
that load. It's reasonable (if not common) to specify a range of load
impedances over which some perhaps slightly reduced performance is
achieved. You may, for example, get more power but also higher
distortion products, at a non-optimal load.

Cheers,
Tom

August 11th 03, 06:12 PM

On 11 Aug 2003 09:54:47 -0700, (Tom Bruhns) wrote:

In such a case, I'd work on a match to the antenna (or
the Cantenna or whatever) first--well, right after making sure the
meter was telling me the truth or something close to it.

Hi Tom,

This is kind of a chicken and egg problem.

Calibrate the meter against the load.

Calibrate the load against the meter.

It would seem that discussion would eventually get around to bringing
in a known and stable reference. The length of lossy line would
suffice for a rough cut solution but still leaves a bench full of
rather sloppy instruments.

73's
Richard Clark, KB7QHC

August 11th 03, 10:31 PM

(Tom Bruhns) wrote in message om...

:-) The meter-needle zero generally not a very big part of the
problem! Yes, it's the coupler (bridge) inside that you want to make
sure is calibrated for your reference impedance. If it's not
adjustable, at least learn what it is calibrated to, or how good it is
for your reference impedance.

It's all supposted to be for 50 Ohms, or course. And that is the
only adjustment on the meter.

The easy way to insure proper operation is to present the
transmitter/amplifier with the right load. A properly calibrated SWR
meter (or as Reg would have it, a "transmitter loading indicator")
will tell you when you have such a load. Yours is telling you that
you do not. In such a case, I'd work on a match to the antenna (or
the Cantenna or whatever) first--well, right after making sure the
meter was telling me the truth or something close to it. When you get
that correct load, then the length of coax won't matter noticably,
assuming the SWR meter is calibrated to the line impedance.

I've never seen an SWR meter that wasn't designed for 50 Ohms...
:/

If you mean adjusting the meter so that the forward power is full
scale deflection, well, that's not needed for a cross-needle SWR
meter.

I'm gonna try the long RG-58 trick...

Slick

Thread Tools	Search this Thread
Show Printable Version	Search this Thread: Advanced Search
Display Modes
Switch to Linear Mode Hybrid Mode Switch to Threaded Mode

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
50 Ohms "Real Resistive" impedance a Misnomer?	Dr. Slick	Antenna	255	July 29th 03 11:24 PM
The Cecilian Gambit, a variation on the Galilean Defense revisited	Richard Clark	Antenna	11	July 24th 03 07:30 PM

Menu

About Us