Stuart, K5KVH, wrote, in part:
> The ARRL web site has a section on testing by their lab
> and how it differs from that of accepted manufacturer's tests.
Thank you for the lead, Stuart. I have just been there and
read the complete method, procedure and examples of
how they do the 2nd and 3 rd order IP tests.
I was shocked!!! Why do they choose to differ completely
from industry standard? They use audio meters to measure
RF performance??? That is nuts. The proper way it is done,
and the way it is done is completely described in the January/
February issue of the ARRL's own QEX magazine in a very
detailed article by Ulrich Rohde, page 21 the piece begins;
actually the 2nd part of a two part paper by him.
The correct test measuring instrument, as used by Rohde,
and every one else is the RF Spectrum Analyzer where the
IM products can be seen and accurately measured in both
frequency and amplitude. Rohde illustrates with a simple
block diagram, his Fig. 30 in the Jan/Feb article. His
actual test equipment arrangement is shown in his Fig. 32,
and the procedure used to measure several rigs --
a couple of Rohde & Schwartz rcvrs, the XK2100L and
the EK985, and a slightly modified Yaesu FT-890 (the
"standard" front end switching diodes were replaced by
Rohde with PIN diodes, a practice Rohde strongly urges
be done with all diodes preceding the mixer stages;
he used MI204 PINs in the tested 890).
He describes the correct Spectrum Analyzer test method
procedure in the two concluding pages of his current article.
My suspicion is that this is exactly the procedure followed in
the Ten Tec engineering labs; it is certainly NOT the procedure
performed by the ARRL labs, if they follow the procedure they
say they do using audio test meters!
BTW, the reason I am so sensitive to this issue is, that so
far as I know, I wrote the first paper ever in which the
intermodulation product performance of linear devices
is analytically discussed and measured: see the September
1965 issue of "The Microwave Journal". The complete
transfer characteristic spectrum of a linear device is
analyzed via a Taylor Series expansion; the expansion
terms define both the frequencies and amplitudes of all
harmonic and intermodulation "error" products arising from
the "near" linear transfer process. In our case of receivers,
the near linear process we are after is from the RF signal
entering the antenna terminals of our rigs, to the output.
Rohde's article is an excellent treatment of how to do this
using 21st century measuring/generating equipment.
I also am sure, that in the Ten Tec engineering department,
that Doug Smith's hybrid coupler is used for combining the
test signals to the test apparatus; they ARRL labs do not,
per their own procedure (they use a simple two port
combiner (MCL ZSFC 2-6). Rohde also seems to imply,
in his referenced article, disagreement with the ARRL
test procedure/method (see bottom paragraph, page
30, Jan/Feb. '03, QEX).
Well, I have let off my steam! Won't change the ARRL lab
procedures, for sure. I believe the ARRL test results
may have validity only for comparative relevance between
rigs they have tested, but I seriously doubt the absolute
accuracy, or even typical accuracy of the reported results
to the procedure being used there.
73, Jim KH7M
|