> Thanks every so much for you very informative post to
> me and to the Ten Tec reflector. That, coupled with your
> willingness to discuss your techniques with Jim, the techniques
> and the process/procedures you and your staff use for generating
> your results, certainly produces a heretofore nonexistent, positive
> impression of what really is going on within the ARRL.
Thanks, Tom. That is why I join in these discussions. I agree that ARRL can and
should
do as thorough a job as possible telling members how we do things. Of course,
in one
sense, we did, because the test procedures manual and several articles
available at:
http://www.arrl.org/tis/info/rigbuy.html
Explain a lot of it. Unfortunately, the world has become diverse and something
put in
one venue may not be seen by someone who looks in another venue. The flip side,
of course, is that assumptions are even more dangerous than the lack of
information.
Hopefully, the three posts I will make, and the expected followup discussion,
will
help on all counts.
The ARRL Lab has a reasonable suite of test instrumentation. The newer stuff is
not quite top of the line, but is right up there, chosen to help ARRL extend
its test
capabilities into the future. The older equipment is quite servicable and is
very much
up to the task at hand, with better specifications than needed for the
performance of
the equipment being tested. The Lab's ability to easily test microwave
equipment is
somewhat limited, although Zack Lau is quite skilled at microwave design and
has been
able to design nearly any test fixturing needed. Most of ARRL's Product Review
testing
has been for equipment operating at 1.2 GHz and below.
All equipment used for major Lab Product Review testing is calibrated per the
cycle
recommended by the equipment manufacturer. We use an external calibration
service,
Essco, near Boston, MA. Every November, Mike Tracy drives the equipment to cal,
and picks it up a week later, with the cal usually sandwiched in between Product
Review testing. Other, less-critical, Lab-owned equipment is self calibrated,
using the
calibrated equipment and the good judgement of the Lab engineering staff. (By
"less
critical" I generally mean equipment that will not be used for Product Review,
publication
or advertising acceptance.)
The majority of the Product Review testing takes place in ARRL's shielded
screen room.
The room is well grounded and the power-lines going into the room are voltage
conditioned with a Stabiline BCR3205U voltage regulator, 5.5 kVa. The output
voltage
will stay at 234 V +- a couple of volts over a wide range of input voltages.
The power
is then run through an EMI filter, to help keep the RF gremlins out. With the
door closed,
the room has about 100 dB of isolation.
The room is not temperature or humidity controlled, other than by the building
heating
and air conditioning. Normally, about 75 F or so, and the building humidity
ranges from
about 10% in the winter to up to about 40% in the summer. For some of our
testing,
we use a small Tenney, Jr temperature chamber, typically testing that equipment
still
functions at the low and high temperatures we expect the human users could
tolerate.
I won't go over every single piece of equipment the ARRL Lab owns, but will
give
the highlights of the more important and heavily used equipment. It should
give a good
overview of the test capabilities of the Lab. In addition to the equipment I
will describe,
the Lab has the usual collection of cables, connectors, adapters, attenuators
and the like.
Only cables that have been tested by the test engineer can be used for Product
Review
testing, and at the least sign of trouble, they are discarded, or relegated to
the other Lab
uses.
For receiver testing, the centerpiece are the two IFR 2041 signal generators,
+13 dBm,
LF-2.4 GHz. These are used in conjunction with calibrated step attenuators and
a few
hand selected MiniCircuits two-port combiners. This fixture has been tested to
at least
+40 dBm IP2 and +85 dBm IP2, with an independent verification using the Lab's
2nd
receive test fixture, which uses 1 W ultra-linear amplifiers donated by Ulrich
Rohde
and attenuators and the like. For miscellaneous bench use, we also have a pair
of
HP8640B generators. They are currently "self-cal" but in good working order,
used
for miscellaneous tasks around the ARRL Lab.
The receiver output is measured with an HP-339A distortion test set. This is an
older
instrument, but in excellent working order and perfectly suited to RMS readings
of the
receiver output, distortion measurements for SINAD sensitivity and the like.
Its analog
meter actually makes it more useful in many cases than would a more modern,
digital
readout. If a measurement is not affected by noise, it will give the same
results as one
might get making the receiver output measurements on a spectrum analyzer, but,
as
I had said in an earlier post, with a much better absolute accuracy.
If needed, our HP-8563E spectrum analyzer (DC-26.5 GHz) does handle audio
analysis
well. We also own a couple of National Instrument A/D cards and can digitize
any signal
at up to a 500 kHz rate and use the accompanying Lab Windows CVI software. We
use
CVI to control all of our test equipment and have written routines to program
the spectrum
analyzer for various measurements. This package is the top-of-the line industry
standard
test software, one step up from LabView.
We also occasionally use the HP8970A noise-figure meter to measure
preamplifiers and
the like.
For transmitter testing, the HP8563E spectrum analyzer is the pride and joy of
the ARRL
Lab! It is good from DC to 26.5 GHz. We ordered the high-stability option and
its
minimum resolution bandwidth is 10 Hz. It can be used, with care, to measure
harmonics
and intermod down to about 90 dB below the reference level. If needed, we do
have some
VHF and UHF notch filters that could be used to attenuate the fundamental, to
make
measurements well below that. The analyzer is used primarily for transmit
spurious
emission and transmitter IMD measurements. We kept our old HP-141T analyzer
frame
and plug ins and sometimes use it for various Lab tasks.
For "gross" power measurements, we have a number of Bird wattmeters, with
elements,
all in current calibration. We self-cal the elements because the times we have
sent them out,
even to Bird, we were not happy with the results. For accurate power
measurements, we use
an HP-437B microwattmeter with various power sensors (all calibrated by HP). To
make
power measurements, the Lab connects a transmitter to a power attenuator whose
exact
attenuation has been previously measured. We measure the power at the output of
the
attenuator, then add back the value of the attenuator. This gets us power
readings to within
about 4% or so, better than the Bird in most cases.
For frequency measurements, we use a portable Optoelectronics counter, with the
high-
stability TCXO option. We have recently acquired a pair of HPZ3801
GPS-conditioned
10 MHz frequency references that are now being integrated into the frequency
counter,
to get accuracy down to a fraction of a Hz. We can also measure frequency into
the
microwave range with the HP5351B counter, with 1 Hz resolution.
This year, the Lab purchased a long-overdue digital storage scope, a Tektronix
TDS3052B.
This will be used for our keying and T/R turnaround meaurements, replacing our
older Tek
scopes.
The Lab has some excellent "homebrew" equipment, too. For phase noise, we use a
test
fixture designed and built in the mid 80s by Zack Lau. A few years back, we
tested a rig and
then sent that rig to Ulrich Rohde, for testing on his $100K+ phase-noise set
up. We found
correlation, +- about a couple of dB, across the entire range.
A number of years back, Jon Bloom designed a 9600-baud bit-error-rate test
fixture that is
used for some of the digital radios.
Perhaps the most impressive homebrew fixture we have it the one that we use for
amplifier
IMD testing. It uses a 100-w coaxial power combiner whose terminations have
been carefully
controlled with matching networks to give at least 60 dB of port-to-port
isolation on 14 MHz
for up to a 3:1 input SWR on an amplifier under test. With this fixture, we can
measure
amplifier performance down to -60 dB PEP on third order and about -70 dB on the
higher-order products. This is much better than amplifiers we have tested, and
we have on at least one
occasion measured an amplifier 10 dB better than the manufacturer, who had
tested the same
amp, indicating that the amplifier manufacturer had measured his test setup,
not his amp!
While not all of this equipment is brand new, it is all quite suitable for the
task at hand. ARRL
has carefully characterized its test setup and knows its capabilities and
always ensures that
its published data are that of the equipment under test, not the test
equipment. The only
exception I know to that occured before my time, where the Lab used the
spectrum analyzer
to measure a transmitter's phase noise, only to realize later that the Lab had
measured the
phase noise of the analyzer local oscillator.
Although newer equipment would improve the usability for the test engineer, and
may gain a
dB or two of accuracy in some measurements, the differences would not be night
and day.
Those reading our results can feel comfortable that they are accurate, human
foibles
notwithstanding, within a reasonable and acceptable accuracy.
I hope the Ten Tec list readers will find this description useful. My wish
list for new equipment
is appended, if anyone wants a nice tax writeoff! :-)
73,
Ed Hare, W1RFI
ARRL Lab
225 Main St
Newington, CT 06111
Tel: 860-594-0318
Internet: w1rfi@arrl.org
Web: http://www.arrl.org/tis
1. Agilent 8753ES Network Analyzer
2. Agilent 35670A FFT Dynamic Signal Analyzer, DC-102.4 kHz
3. Agilent 5504B Phase Noise Test Set
4. Agilent 6542A power supply
5. Agilent 85640A Tracking Generator
6. Tektronix TDS-5052B (with Opt. "PS" - multiple option package)
7. Tektronix TDS-3052B (with TDS3AAN, TDS3BAT and TDS3GV options) - qty 2
8. Agilent E4407B Portable Spectrum Analyzer with E1779A battery pack
9. Rohde & Schwarz FSH3 Portable Spectrum Analyzer
10. Agilent 34401A Digital Multimeter - qty 5
11. Radar Engineers Model 240 HF-UHF RFI Locator
12. Radar Engineers Model 250 Parabolic Ultrasonic Pinpointer
13. Radar Engineers Model 247B Hotstick Line Sniffer
14. Simplex Inc. Watt-Muncher Jr. Portable Load Bank
15. Rohde & Schwarz ESIB26 EMI Receiver
|