Frank’s RSMS History
My brief personal history of the Department of Commerce Radio Spectrum Measurement Systems
Go to the Using My Photos page on this site for instructions on how to obtain high-resolution versions of these images from me. I don’t charge money for my photos, but I do require a photo credit in your end-product in exchange for the use of my images.
Background on U.S. Federal Government Spectrum Management
Back in the 1920s the U.S. Department of Commerce developed a mobile radio-measurement system. The responsible agency was the National Bureau of Standards. Then as now, the Commerce Department needed to gather quantitative data on the strength, frequency, and amount of usage of radio signals across the entire radio spectrum. The early Commerce measurement vehicle had many features in common with later systems: a measurement system in a rear area that could be operated from either a generator or from shore power, a telescoping mast, and a variety of radio antennas.
During the tumultuous years of the Second World War and early part of the Cold War, the Department of Commerce did not have a mobile radio measurement system. This was too bad, inasmuch as the Federal Communications Commission (FCC) in the USA is only responsible for non-government radio assignments, and as such only monitors those assignments. The FCC answers to the U.S. Congress; it is a congressional commission. Examples of radio assignments for which the FCC is responsible include commercial broadcasting, local public safety, and amateurs (hams.
In contrast, all federal government radio systems (such as government radars and radio systems by all federal departments including Defense, Interior, Agriculture, etc.) in the USA are controlled and regulated through the executive branch of government, which is to say the office of the President of the United States. The President’s agency for this regulation, control and coordination of federal use of the radio spectrum is the National Telecommunications and Information Administration (NTIA), which in turn is an agency of the Department of Commerce. Why the Department of Commerce, you might well ask? The Department of Commerce got this job because it was the home of the original Bureau of Weights and Measures, later the National Bureau of Standards (NBS) and now called the National Institute of Science and Technology (NIST). Back when NIST was NBS, that agency spun off the organizations that became the predecessors to NTIA. The arrrangement may sound weirdly convoluted, but actually it works extremely well.
The Commerce Department’s NTIA chairs the Interdepartment Radio Advisory Committee (IRAC), which was founded in 1922 (when radio was called wireless) and is still the overall coordinating body for radio systems between all U.S. federal government agencies. Thus NTIA is also responsible for keeping track of the use of the spectrum by federal government agencies.
The RSMS-I
By the 1970s there was concern that Federal agencies either were not using many of their radio frequency assignments or else might be using them improperly. The only way to verify the proper operation of these assignments was to design and build a system that could independently monitor the assignments. Thus was born the Radio Spectrum Measurement System (RSMS). The RSMS was designed and built as a joint effort between the Department of Commerce’s Institute for Telecommunication Sciences (ITS) in Boulder, itself an offshoot of the wartime Central Radio Propagation Laboratory (CRPL) which had in turn branched out of the National Bureau of Standards. These activities were located in Boulder because of Congressional politicking in the 1950s, when Boulder was worried about being totally dependent upon the University of Colorado for the health of its local economy. (The Commerce lab where ITS is housed was opened by President Eisenhower in Boulder in 1954.)
By the mid-1970s a variety of swept-IF spectrum analyzers were available on the commercial market, notably produced by the Hewlett-Packard (HP) company. Early ideas for a mobile spectrum measurement system focused on the concept of tuning a spectrum analyzer to individual frequencies to check the usage of the those frequencies. But it was obvious that, to check any significant number of frequencies for any significant amount of time it would be necessary to do more than play with a manually controlled spectrum analyzer. Instead, a computer-controlled measurement system would be needed, one that could check hundreds of frequencies every minute, over and over again, twenty-four hours a day. For this purpose the Commerce Department’s ITS lab teamed with HP to develop a computer-controlled version of a HP spectrum analyzer. This system was the ARS-400.
The ARS-400 was integrated into an RF-shielded Dodge Travco motor home. The Travco was fitted with five full-height equipment racks to house the ARS-400, two on-board power generators, two air conditioners and a 30-foot-high rigid mast that was raised and lowered with a motor-driven pivoting screw shaft. (See more below on the ARS-400 and its significance in the history of spectrum measurements.)
When the van was delivered in 1973, it was a political hot potato. The Watergate scandal was building up steam. President Nixon would resign in August 1974 in the wake of the scandal, which had devolved upon issues of illegal monitoring of people whom Nixon had believed had opposed him politically. So it was with some trepidation that initial RSMS operations took place. The trepidation was exacerbated by the fact that the RSMS was being operated by an Executive branch, as opposed to Congressional branch, agency. We stayed out of trouble by adhering strictly to technical work that was itself constrained a set of rules, called the RSMS Program Instructions, that had been drafted by Commerce lawyers and engineers to keep us out of trouble.
Radio Enforcement or Radio Research?
When our operations began, the initial idea was to use the RSMS as an enforcement tool to check compliance of government agencies with applicable radio regulations and frequency assignments. But early in the RSMS history, by 1977, two people decided that it was mostly a waste of time to use the RSMS for such purposes. These two people in NTIA played an instrumental role in reshaping the remainder of the history of the RSMS. They were Bob Hinkle of the NTIA Office of Spectrum Management (OSM) in Washington and Bob Matheson, head of the Radio Spectrum Occupancy Group (RSOG) at the NTIA/ITS lab in Boulder. These two guys realized that the RSMS would be far more useful to the government as a basic research tool than as an enforcement tool. Accordingly, they worked to change the attitudes of other people in NTIA, the Commerce Department, and other federal agencies as to what the best and most proper use of the RSMS should be.
Bob Matheson in the late 1970s, with part of our first radar measurement system(specifically the pulse-sorting part of it) looming behind him.
Hinkle and Matheson prevailed. By 1979, when I joined the RSOG group (still myself technically in high school), the RSMS was already being used as a research tool, with just a bit of enforcement activity on the side. The enforcement activities were eventually dropped entirely, in the early 1980s. By that time, NTIA had discovered that enforcement activities paid very small dividends compared to acquiring data, developing new knowledge, and applying that knowledge to the management of radio and radar systems.
A particularly interesting research topic that Hinkle and Matheson undertook was to learn what radar emission spectra looked like, to determine what made the spectra look the way they did, and to determine ways to make radars use less spectrum. I came into the group just as this radar emission spectrum work was really getting started. It was the perfect time to join. Once again, as has so often been the case in my life, I totally lucked out. How does that work? I will say that I made the most of my luck by working really hard on the effort.
The radar emission spectrum project effort was a big success. We gathered a huge amount of data on emissions from nearly every type of radar in the U.S. inventory. For each radar, we gathered emission spectra and all other pertinent emission characteristics. We traveled for weeks at a time with the RSMS, moving from one radar location to the next like a band of radar gypsies. (Maybe radar Vikings, raiding one radar site after another, would be a better metaphor for us in those days.)
We became very efficient at arriving at a location in a stone-cold condition, getting the RSMS into a good position near the radar (usually about 0.3 to 0.7 miles away from the radar), firing up the generators and the air conditioners, booting the computer, setting up the measurement system, getting it calibrated, and then starting the radar measurement. We finally got to where we could do all that work in under an hour, starting at the moment that we threw wooden blocks under the RSMS vehicle jacks as we pulled into position near the radar.
As for the radars themselves, the operators were usually worried that they would have to do something special to their radars, to alter their operations for our measurements. But we always insisted that they just operate the radars normally--we did the rest. We shot the radar emission spectra like wildlife photographers shooting their quarry from a blind--no interaction with our quarry, and please, just act natural while we’re working.
We developed a special measurement technique that allowed us to quickly and accurately obtain all the emission characteristics of a radar, including its emission spectrum, pulse width or widths, pulse repetition patterns, pulse modulations, antenna patterns, and so forth, without needing any access to the radar itself--we worked out how to get everything from a distance. We had built our own radar electronic intelligence (ELINT) system, we had trained ourselves in radar ELINT techniques, and then we had become highly proficient at getting results from our new capabilities. And we did it all for better spectrum management for the taxpayers. Wow. Those were the days.
We measured emission spectra of airborne radars, naval radars, meteorological radars, air traffic control radars, little radars on boats and giant space-search radars--you name it, we shot it in those years. I’ve subsequently kept up with newer radar types. It has been a great adventure, and its still a lot of fun. I just regret that so many of the early guys have now retired. I miss them. Oh, well.
Development of the Famous HP-8568A and 8566B Spectrum Analyzers
Not only did we have a lot of fun, but Hewlett-Packard, the company that built the initial RSMS hardware and software, had a lot of fun with the system as well. Their engineers were brilliant. As noted above, their original computer-controlled radio measurement system was called the ARS-400. It occupied five full-height equipment racks, and is shown in the slide show that is attached to this page. It was fabulously complicated. It took a dedicated team of well-trained electronics engineers (highly competent engineers, who thoroughly understood both the theory of spectrum measurements and the application of state-of-the-art technology) to keep it running. We literally disassembled and reassembled various parts of the system on a routine basis just to keep it operational. But, oh boy oh boy, did we learn every little thing about how an automated spectrum measurement system needs to work. HP built very few of the ARS-400s, and so far as I know our group in NTIA was the only one that ever successfully kept the system running on a long-term basis. We were definitely the only group that not only used the hell out of the ARS-400 system, but who used it on the road, in extremely demanding field situations, as opposed to using it in a pristine laboratory environment.
On top of all the other weirdly complicated aspects of the system, our 8500-series HP controller computer was an early 16-bit digital machine. It had 16 big flashing lights on the front panel, numbered 0-15, which showed us the actual machine state at every clock-cycle interval. We could actually stop the machine on any given cycle and change its state manually--now those were the good and early days of digital computing! The computer used genuine non-volatile magnetic-core memory: every logic bit was stored as a hysteresis state in a single magnetic donut that was mounted on a circuit card at the junction of two wires that ran across the board at right angles to each other. Somebody told me once that HP paid relatively high wages to Navajo women who were especially skilled at building (weaving, actually) these fabulously intricate wires with these hundreds of thousands of magnetic donuts mounted on them. They claimed, in fact, that the wages paid to the Navajo women by HP were so high that they upset the economic balance between those women and their men. I can’t vouch for the truth of that statement.
Detail view of nearly microscopic core memory from the first RSMS computer. Every magnetic donut stored a single logic bit. The memory mechanism was magnetic hysteresis. Each donut was addressed, read and switched by a pair of wires that ran through the open center. Simple parity checking on rows and columns of donuts helped to keep the memory robust against random errors.
The computer was huge. It was nearly cubical, measuring about eighteen inches high, two feet deep, and nineteen inches wide. It took two strong men to lift it and put it into its nineteen-inch-wide equipment rack mount. The total amount of memory was 64 kilobytes (32 kilowords). That’s half-a-million magnetic donuts that were individually woven onto eight memory planes on four big circuit cards by those Navajo women. It was not just an astounding work of technology, but also of art. It was based on a DEC design. There was nothing else like it in the world at that time. I salvaged one of those memory planes for my private collection when we finally excessed the computer years later, and I’ve still got it. I photographed the memory-core images on this page from my salvaged memory card.
And what did all this lead to? Well, HP used their experience in designing, building, and fielding the ARS-400 to design their extremely famous, breakthrough-capability, computer-controlled spectrum analyzers, the HP-8568A and the slightly later and somewhat more capable HP-8566B. They were miracles, each one just a pair of electronic boxes that fit into a nineteen-inch rack mount. They were just light enough to be lifted by a single person! Those analyzers were the best machines that you could get (in my opinion) until the Agilent (the successor of the HP microwave electronics division) E-4440A analyzers that recently became available. I could (and still can) really make the 8566B sing, playing it like a musical instrument, and why not? I understood every part of it, after having run the ARS-400 all those years. (With that said, fundamentally all spectrum analyzers have to work the same way, even the new machines that directly digitize at the final-stage downconversion mixer output, when it comes to the theory of spectrum measurements as a convolution and detection problem. So you name the analyzer, I can use it to get the spectrum measurement done.)
What We Achieved with the First RSMS
As I mentioned above, we absolutely ran the hell out of that first-generation system. We took it all over the country in the RSMS-I and used it to measure radar emission spectra for several dozen radars. We used it to perform land mobile radio usage measurements. We used it to resolve numerous interference problems. We used it as a research tool to better understand the characteristics of spectrum occupancy in major metropolitan areas. We used it to better understand the potential and the limits of existing radio and radar transmitter technologies. We answered all sorts of questions that were critical to better spectrum management for the U.S. taxpayers. And relatively speaking, we did it all on a shoestring budget. It was a triumph for us and for the U.S. taxpayers who were paying us!
How to Manage Projects for Success and Becoming an Expert on a Technical Topic
Allow me to digress on that topic for a moment. I would rather, in any given circumstance, do a job with a small cadre of highly trained and motivated individuals who know exactly what they are doing, than to attack a problem with a vast armada of people and resources who are disorganized, poorly trained, and badly led. Our experience with the RSMS helped me to form that opinion. Time and again, I have seen failure occur when people were well-funded but too numerous and poorly led, as opposed to seeing success occur when the number of people and funding for a project were more limited but better directed, more knowledgeable and more focussed.
I know how to do a lot of things merely well, but I dare say that I’m either the top expert in the world on the topic of measuring radar emissions, or at least one of the top three. Admittedly it’s a niche market, but it’s important that we know how to measure these emissions. Using our special technique, we can measure a radar emission spectrum with at least 100 dB or dynamic range, and usually with 110-120 dB of dynamic range. That’s as good as or better than anyone anywhere can do. I can look at a radar spectrum and immediately tell what the transmitter output device probably is (magnetron, crossed-field amplifier, klystron, etc.). I can watch the radar emissions in the time domain for a few seconds, noting pulse width, pulse repetition rate, stagger (if any), frequency hopping and diversity behavior (if any), pulse repetition interval(s), and then tell you almost exactly how the radar works, what its functions are, and what its maximum range will be. It’s so much fun to do that, I’d rather --not-- know anything about a radar before I measure it. It’s more fun to work it all out for myself, without any briefings.
RSMS-II, III and IV, and the Compact Radio Measurement Suitcase Systems
By the mid-1980s we began work on a major upgrade of the RSMS capabilities. We retained the Dodge Travco motorhome but we completely rebuilt all of the custom boxes in the old ARS-400 measurement system. We migrated to a new computer (a Hewlett Packard A600 which replaced the original Hewlett-Packard computer with its 64 kilobytes of non-volatile magnetic-core memory for both the BASIC interpreter and the application program). We began to record our data to disks instead of magnetic tape. We copied all of our old data files to disks as well. The new system took a few years to build, as were doing all of the work on our own, but the RSMS-II worked well when it was finally finished.
By 1990 we needed a new vehicle and yet another new version of the RSMS. We built the measurement system into a box on the rear of a Chevrolet Cheyenne pickup truck chassis. It had two telescoping masts and an onboard generator, with air conditioners and so forth. The RSMS-III was PC-based and was able to take advantage of much higher-quality, commercial off-the-shelf spectrum analysis equipment than had ever been available before. We also began to do more and more suitcase-based measurements. I like suitcase-based system because they are light and fast for deployment and use. I can really fly with one of those babies, setting up in a location and getting the necessary information before anyone quite knows what’s going on. Suitcase systems allow us to get the job done with an even lower profile than before, and that’s often a good thing.
The latest RSMS is the fourth version, RSMS-IV. This one is based on a large diesel truck chassis. It likewise has two masts, on-board generator, etc. It is able to take advantage of the latest radio measurement equipment, and is very nice to use and to operate.
Finally, I have to say that the real key to successfully operating all of the RSMS and suitcase measurement systems is two-fold: The people and the software. The best hardware in the world is useless if you don’t have people who really understand it and know how to use it, and if you don’t have software that does what you need to do. We’ve worked pretty hard to cultivate a good set of people over the years for our spectrum measurement operations, and that investment has paid off handsomely. We’ve worked tirelessly on our software, to make it do what we need to do, and that’s had a big pay-off for us, too. I’m looking forward to many more years of great experiences in the field of spectrum measurements with the U.S. Department of Commerce.
Original RSMS 1927. They had the right idea and essential elements early on.
RSMS-I/II in Travco motor home
RSMS-III during a desert exercise with the U.S. Army.
Our most current system, the RSMS-IV.
RSMS-I staff, 1970s, with part of the Hewlett-Packard ARS-400 system.
RSMS original radar DF system
Low-profile suitcase measurement system
Early RSMS radar measurement in progress
My primary mentor John Smilley
An important mentor & my first Group Chief, Bob Matheson
Vince Lawrence, Group Technician
Rooftop spectrum survey in progress
RSMS-I ARS-400 land mobile radio (LMR) measurement system
RSMS-I ARS-400 radar measurement system using core-memory computer.
RSMS-I core memory detail. Each donut stored a single logic bit. 1973-1983.
Core memory. Two wires controlled addressing & switching on each donut.
RSMS-I core memory card, 1973-83. They were supposedly woven by Navajo women.
Part of the early RSMS-I ARS-400 development system
RSMS-III interior view. This system was based on HP-8566B spectrum analyzers.
RSMS on a landing craft. We go anywhere and everywhere.
RSMS-III on tactical desert ops
Desert chopper deployment
Air search radar
Interference troubleshooting in the desert
