> > It's peculiar that what must be essentially the same final comes in
> > with much worse numbers when tested as the OMNI VI +.
The final amplifier is a very small part of the system. In my IC-
751A's, most of the distortion comes from the driver transistors. In
my FT1000D, most of the transmitter composite noise and a noticeable
amount of IM distortion comes from an FET very early in the
transmitter chain!
> That is why I take the tested and published IMD figures with a grain
> of salt. The Omni Six displayed spectacular IMD figures at 14 MHz and
> the Omni Six Plus displayed IM distortion some 20 dB worse. I suspect
> that a faulty Omni Six Plus was secured for testing. Performance as
> measured on one radio sample cannot be extrapolated into the entire
> life cycle of the product.
Taking measurements with a "grain of salt" is a good idea.
In my experience, products stay very close to the measured values in
early tests all though the life of a product run if they are designed
and built correctly. For example, I can grab an AL800H (or any other
amplifier) prototype from many years ago, and it tests within
instrument tolerances like early unit that was set aside.
The only time retesting is necessary is when a critical component is
changed.
The real catch is proper IMD tests are critical for testing setup and
procedure. The tester must be VERY familiar with testing procedures
and potential problems. Because of that, someone should verify every
test with multiple tests. It is more likely someone would test the
gear incorrectly than have the gear itself be a flyer, assuming the
factory didn't change parts and set the bias correctly.
The best idea is to seek out multiple test reports from independent
sources and see if they agree, or make sure the test was done
correctly. When I test, I repeat the test a few days or weeks later
after the test setup has been disassembled and reassembled. (I
actually use two different test setups for amplifiers, with different
measurement devices. One is a selective level meter with two
generators mixed and then amplifier in a class A amplifier, the other
uses a spectrum analyzer and mixes two transmitters. Both agree with
a dB or two, unless there is an obvious problem with one setup. And
yes, I catch mistakes on occasion. Usually those mistakes are caused
by flaws like termination SWR on a combiner, or a level adjustment
error. Sometimes it can be a test lead that has gone bad.)
I would hope the ARRL does something similar, or at least has two
different people run each test from **scratch** (connecting the
equipment from scratch).
I watch to see if data from the RSGB or other sources and the ARRL
agree. I also listen on the air, to see if problems are evident.
Since I have never seen second-source verification of an Omni VI, I
have no idea if the -25dB is correct or not...so it would be unfair
to comment on it. Has anyone seen second source verification for that
rig, or does the ARRL do verifying tests?
> Does anyone know if IMD testing is conducted by most manufacturers on
> every radio that leaves the factory? Seems like this, and several
I doubt that happens, because of the time and equipment required to
do the test.
> other tests should be measured in QC prior to a radio getting into
> that hands of the customer. Kachina used to send a computer-generated
> BITE report with every transceiver purchased and matched against the
> serial number. That was a great idea.
I'm not speaking of Kachina, or implying anything about what they
might have done........but there is something we should consider.
We should always know how the tests are actually made.
For example, Amp Supply had a word file in the computer that printed
out test data and they included that data (with serial number of the
shipped unit added to the page) with each amplifier. Yet production
amplifiers were not actually tested for the printed data in a
calibrated test. It was a "nice" marketing touch to send out computer
printed sheets, but testing was no different than anyone else did. It
just gave the illusion of better QC.73, Tom W8JI
W8JI@contesting.com
|