[1000mp] ARRL Lab transmit IMD test-result reporting

Tue, 9 Apr 2002 12:07:30 -0400

Jeff Pawlan writes:

> One of my serious pet peeves about their reporting procedure is that
completely
> contrary to both industry standard and FCC part 97 rules intent, they
report the
> 2 tone intermods levels relative to an IMAGINARY carrier level by
combining the
> power of the two tones. 

Hi, Jeff,

There is nothing imaginary about combining the two tones -- the result, 6 dB
above the level of either of the two steady-state tones, is called "peak
envelope power." :-)   

There are a number of reasons that ARRL reports the data this way.  FCC
rules on amateur output power, and the specifications of all amateur SSB
equipment, is specified as peak-envelope power. Thus the IMD on a 100-watt
PEP amateur transmitter is reported relative to that 100 W PEP output power,
not to the 25 watts of either tone.   And as to the "intent" of the FCC
rules, I don't know of any reasonable way to fathom meaning from the rules
past the language of the rules.  Amateurs do NOT have the same stringent
requirements as the commercial rigs.  Do you want them?  Amateurs do NOT
have specific limits on the close-in IMD that exist for the commercial
transmitters.  And even the transmitted IMD that is greater than +-250% of
the carrier, and thus is spurious emissions, does not have to meet the same
requirements as the commercial transmitters.  And, as spurious emissions,
the FCC regulations do not mandate that they be N dB down from a single
tone, but that the level of the emissions be N dB down from the average
power of the transmission.  For a two-tone test, that would be 3 dB higher
than a single tone , if you want to insist that reporting be done by the
standard defined in the FCC regulations.

The ARRL Lab has been using the transmitter PEP output as the reference
level on its IMD measurements for as long as I have worked here.  We
engineer types can easily interpret results presented either way, so saying
that a radio has two-tone IMD that is 25 dB below PEP says exactly the same
thing as saying the IMD product is 19 dB below a single tone -- just like
saying that a sine wave has an RMS voltage of 1.0 volts is the same as
saying it has a peak voltage of 1.414 volts.  However, changing the
reporting convention at this stage would mean that the non-engineers could
no longer easily compare the results of a rig tested and reported in 1985
with one tested and reported in 2002.  The ARRL Lab is always open to
changing the way that it tests products and reports on the result, but I do
not think we should change things that break the continuity of being able to
easily compare past and present results without some pretty powerful
reasons.  I do not think that those reasons exist in this case.

The ~-25 dB PEP for third-order IMD that has been cited in some of this
discussion is really not that atypical. Since the ARRL Lab started testing
and reporting the IMD on ALL bands, instead of just 80 and 20 meters, as it
did when I first started, many of today's rigs have IMD that is
approximately at that level.  This is not a good thing, mind you, but
indicates that this is really the typical performance one might expect,
worst case, from 12-volt powered finals.  I do agree that the Product
Reviews should do a better job at noting whether IMD performance is
excellent or marginal.

For reference, I plotted the 10-meter IMD performance of the IC-746PRO vis a
vis the commercial requirements for out-of-band and spurious emissions at:

http://www.arrl.org/~ehare/temp/imd.doc

> My background and work every day is as a Consulting
> engineer in RF/microwave/wireless communications equipment design and I
would
> never report to a customer that the intermod products are n dB down from
an
> imaginary power level; they are always relative to the actual tone levels.
In
> fact, several radios tested by the ARRL if reported this correct way would
NOT
> pass the FCC equipment requirements.

Almost all of the rigs made today would not pass the FCC equipment
requirements, especially for spurious emissions. The limits for spurious in
Part 97 are a LOT less stringent than the commercial limits.

> Another major deficit of the ARRL testing is the apparent lack of concern
about
> the phase noise (which is never shown wider than 22KHz) and also the
obvious
> spurs in the phase noise sidebands.

The limits of our test set-up are around -150 dBc/Hz, if memory serves.  By
the time one gets to 22 kHz away from the carrier, most rigs are pretty
close to the lower limit of what we can do.  The League instituted
phase-noise testing of ham equipment, in the mid 1980s.  The publishing of
that data has, IMHO, lead to a significant improvement in the transmitted
noise component of modern ham rigs.  They have published useful phase-noise
plots of every rig since they started testing. Mike Tracy, our test
engineer, agrees with me that the Lab needs to do a better job at flagging
to the Product Review editor phase-noise performance that is marginal or
excellent so that such things can be discussed in the running text.  As to
the spurs, keep in mind that the reference level of the phase noise plot is
-60 dBc/Hz, with the bandwidth of the analyzer set to 100 Hz, so the
reference level is -40 dBc. Those spurs are way down -- about -60 to -90
dBc, in most cases. I do not agree with you that they should be given any
more significance than what is shown in the phase-noise plot.

> I am not certain, but there may be a correlation between the reviews and
the
> booking of classified ads upon which the budget is highly dependent.

I think that you mean the display ads, not the classified ads.  As I am sure
that any manufacturers who post here (W8JI?) can attest, the independence of
the Product Review and Advertising departments is pretty evident.  Although
advertising revenue IS important to ARRL, all ARRL owes manufacturers,
irrespective of their advertising, is to treat their products fairly. To
that end, the League obtains a product on the open market, tests it and
reports on the technical test results and the impressions of the reviewer on
the equipment. The only major concession to manufacturers (advertisers or
no) is that just before press time, we give them a copy of the Review so
they can correct any factual errors.  Their response ranges from correcting
errors (as in "that feature does exist, but wasn't documented in the manual"
to "can you tone down the statement where you said 'a pox on the
manufacturer' because that will not translate very well into Japanese and
when our parent company reads it, they will wonder why you want them to
develop sores on their bodies'".  We also now share the technical test
results early on, to good effect, because the manufacturers have helped find
units with component or design defects (each treated differently in the
Review), and even a few out-and-out ARRL Lab testing errors that, due to our
willingness to look closely at our results and work with the manufacturers,
did not make it to print.

All in all, I think that the technical part of the process is pretty good.
What do *I* think could be improved?  I think that some of the issues
brought up here are valid.  I don't think that the Reviews always pick up on
a rig that is overly clicky or has marginal IMD, for example.  I, too, offer
my feedback to the Product Review editor from time to time.  In response to
some of the discussions between ARRL's test engineer, Mike Tracy, and Tom
Rauch, Mike has been putting actual measurements of keying sidebands into
the expanded reports. Mike reports that most rigs he has tested are NOT
excessively clicky.  Ditto on some of the discussions that they have had wrt
IMD -- the Lab now adds close-in test results to the test table.

As an aside to that, some of the worst clickers had rise and fall times that
were >1 millisecond. A careful look at the keying envelope shows that the
worst rigs often have some small portion of the envelope that has sudden
transistions -- an almost vertical section of the waveform envelope for 10%
of the rise time, for example, that generates most of the noise. I think
that this is an area where the expanded test reports can continue to make a
positive contribution.

But over the years, the Lab has added a LOT to the battery of tests -- look
at the 40+ page expanded test result reports, available on the Product
Review page, just for an example/

Are there areas for improvement? Sure, and both HQ staff AND member input
have resulted in many changes to the way that ARRL takes and reports data
over the years.  I hope that we can continue to do so.  Some of those
improvements are taking place this week, with Mike Tracy working on better
ways to test tuners, based on input from W8JI. (The jury is still out on how
the improved methods will stack up against the methods the ARRL Lab used the
last time we tested tuners.)

73,
Ed Hare, W1RFI
ARRL Lab
225 Main St
Newington, CT 06111
Tel: 860-594-0318
Internet: [email protected]
Web: http://www.arrl.org/tis

ARRL is the National Association for Amateur Radio.  It is supported by
membership dues, individual contributions and the sale of  publications and
advertising. For more information about ARRL, go to
http://www.arrl.org/news/features/inside-your-league.html. For more
information about membership, go to http://www.arrl.org/join.html.  Your
contribution can also help support ARRL's ongoing efforts to protect Amateur
spectrum. Go to https://www.arrl.org/forms/development/donations/basic/ to
learn more about the ways you can support the ARRL programs and activities
of most importance to you. You can help ARRL protect Amateur Radio for you
and future generations to enjoy.