Owen Duffy wrote:
The plots are at http://www.vk1od.net/lost/NN7K.htm .
Comments?

Good stuff, Owen. Can the average SWR meter be
calibrated for the actual Z0 rather than for
50+j0 ohms?
--
73, Cecil http://www.w5dxp.com

On Wed, 06 Sep 2006 12:22:37 GMT, Cecil Moore
wrote:

Owen Duffy wrote:
The plots are at http://www.vk1od.net/lost/NN7K.htm .
Comments?

Good stuff, Owen. Can the average SWR meter be
calibrated for the actual Z0 rather than for
50+j0 ohms?

Typical samplers derive a voltage sample and a current sample in a
small region approaching a point for all intents and purposes. The
current sample is converted to a voltage (by passing it through an
impedance, and added algebraically to the voltage sample. The
resultant RF voltage is rectified and fed to a voltmeter.

If the impedance used to convert the current sample to a voltage is
purely resistive, and assuming the sampling process does not introduce
any differential phase shifts, then such an instrument can only be
nulled for a purely resistive Zo. If that conversion impedance were
complex and adjustable, then it would be possible to null the
instrument on practical values of complex Zo.

However, under a linear RLGC model, Zo varies with frequency, both Xo
and to a lesser extent Ro (see the graphs at
http://www.vk1od.net/tl/tllc.php ).

So, nulling the instrument on nominal Zo (Ro) makes sense, but just
keep in mind a small contribution to error and an explanation for why
measurements might not agree with a Zo=Ro model. I do recall someone
insisting that reflection coefficient cannot be greater than 1 because
a Bird 43 would never measure such. He was right, the Bird 43 should
never indicate such, but not because reflection coefficient cannot be
greater than 1 in the general case, but that the sampler is calibrated
(nulled) for a Zo=Ro scenario.

The discussion about using a 50ohm VSWR meter in a practical line
where Zo is not exactly 50 ohms was just to acknowledge that there is
error introduced, that it can be estimated, and that it is low at low
VSWR values and will be likely to be lost in other measurement noise
due to lots of other variables.

Owen
--