On Sat, 04 Dec 2004 05:24:33 GMT, (Robert Lay
W9DMK) wrote:

Just today, I made a careful measurement on an RG-8/U line of 5.33
meters length at 30 MHz and terminated with a 4700 + j 0 load. The
Matched Line Loss of that line at 30 MHz is 0.9 dB per 100 feet, and
its Velocity Factor is between 0.75 and 0.80 The input impedance was
actually measured at 2.45 -j15 ohms for an SWR at the input of 22.25.
The SWR at the load end was 94. Those two SWR's establish a total loss
on the line of 0.15 dB. If one were to blindly apply the formula in
The Antenna Book on page 24-9, the result obtained would be 4.323 dB.

Surprisingly, it was not until today that I finally made a computation
of the input power to the line for the configuration above.
Specifically, I was able to compute the voltage and current at the
input to the line that would produce a 100 volt reference voltage
across the 4700 ohm line. That is the obvious thing that had to be
done in order to establish a reference power for purposes of computing
losses. That calculation resulted in an applied voltage at the line
input of 29.2 volts at angle -171.5 degrees and a current of 1.917
amps at -90.78 degrees. Computing power into the line as E*Icos(theta)
= 9.024 watts. The power delivered to the load is 100 volts squared
divided by 4700 ohms, which is 2.127 watts.

Therefore the efficiency is 23.6% and the losses in the line are 6.275
dB. In all fairness, I did have to change one assumption in the data
above. I had to revise my attenuation value of 0.9 dB per 100 ft.
upwards to a value of 1.72 dB per 100 ft. in order to get my measured
impedance at the line input to be consistent with that line impedance,
length, load value and velocity factor.

Hi Bob,

I notice that true to form, all the response in this thread were not
to your bench results. I trust you will appreciate hard
correspondence rather than the fluff.

The one thing (actually there are several) I noted was your having to
double the presumed line loss to make the numbers come out. Given
this injection (or removal) of 100% (or %50) of error, it stands to
reason that your bench, method or instruments need a attention.

The details as I've sifted from the postings:
Cable type = Columbia's number 1198 - not 9913.
Open circuit stub length = 5.334 meters
Frequency = 10.6 MHz,
Input Z = 0.57 + j 0.3 ohms.

The above details appear to have shifted in mid-stream to:
Cable type = RG-8/U line
stub length = 5.334 meters
Frequency = 30.0 MHz,
Open end termination = 4700 + j 0
Input Z = 2.45 -j15

I will skip the determinations of loss and SWR as being problematic as
you indicate and simply go with your observations noted above, and
summarized he
Open end Vtermination = 100V @ 0°
Open end Itermination = 0.0213A @ 0°
Fed end Vinput = 29.2V @ -171.5°
Fed end Iinput = 1.917A @ -90.78°

I was puzzled to see what was initially a resonant stub now measured
with an extremely high Reactance until I re-scanned the material to
note the tripling of frequency.

What caught my eye was this load and certainly your leap of faith that
it was wholly resistive, especially at HF for its size. This would be
extremely unlikely even in a standards lab.

To make a measurement, the rule of thumb is to have instrumentation
whose precision and accuracy exceeds the goal of measuring an unknown
by 5 to 10 times. For really difficult tests (and RF is classic in
that regard) 3 times is often the best you can achieve.

Let's look at that 4700 Ohm resistance. It demands that any
instrumentation support a paralleling load of no less than 23.5K Ohms
to 47K Ohms, or worst case, 14K Ohms. Let's simply ask about the
family lineage of that 4700 Ohm resistance. It sounds suspiciously
like a common (hopefully) carbon composition resistor.

If so, such beasties are rare if it is of the commercial variety, to
not exhibit reactances due to spiral cut value trimming, or an
end-to-end capacitance. Let's just say that you obtained a remarkable
resistor, but it cannot escape this common parasitic capacitance which
for the garden variety resistor amounts to 1.5pF.

At 30 MHz, this 1.5pF capacitance represents a reactance of 3.5K Ohm.
This rather sweeps aside your specification for the load and replaces
it with 4700 -j3537 Ohms. This is a big time source of error and does
not even come close to the 3X requirement for the lowest accuracy.
Worse yet, you haven't even added the bridging impedance of your
measuring device which is certain to be on par, if not worse (you
haven't identified your instrumentation).

Hope you are still looking forward to more fun. ;-)

73's
Richard Clark, KB7QHC