ATSORS, the Precursor
It was as early as 1983 that they brought that system, and I say early because the MRL-5 radar in Camagüey had been delivered on December 16, 1981 (a gift for my first birthday, just starting out working in Meteorology). In 1982 the radar barely worked, due to air conditioning problems—
It was as early as 1983 that they brought in that system, and I say early because the MRL-5 radar in Camagüey had been delivered on December 16, 1981 (a gift for my first birthday, just starting out in Meteorology). In 1982 the radar hardly worked at all, due to air-conditioning problems and the conversion needed to run on Cuba's 60 Hz industrial power grid. So, under that enigmatic name, ATSORS (transliterated from the Russian АЦСОРС, standing for Автоматизированная Цифровая Система Обработки Радиолокационных Сигналов, which translates as Automated Digital System for Radar Signal Processing), ATSORS was the first system in Cuba capable of recording the digitized video signal data coming out of the MRL-5 radar's receiver and later processing it at a Computing Center. If you keep reading, I'll tell you how it went.
ATSORS consisted of two things: hardware and software. The hardware was meant for acquiring the digitized signals, and the software handled processing them afterward. The hardware was based on a Magnetic Tape Data Preparation Unit — that is, equipment originally meant for an operator to key in data that would then be recorded onto a magnetic tape. In COMECON (the Council for Mutual Economic Assistance of the socialist bloc countries) coding it was the ES-9002. In the photo at the top of this post you can see the equipment as it looked in 1984, in a feature from the magazine Juventud Técnica, which, incidentally, is the only photo I ever found of that equipment (back then we didn't have cameras). Both the hardware and the software were built by Yuri Ivánovich Rudenko. Rudenko was basically a programmer and systems analyst — more of a software man, really, with some basic hardware skills thrown in.
The ES-9002 unit had an analog-to-digital converter added to it, and its operating modes had been modified so it would automatically record the signals at the digitizer's output. Several antenna rotations were recorded, starting at an initial elevation angle of 1.5 degrees and then increasing gradually by 1.5 degrees according to the radar's original programmed control sequence. Recording began automatically once the antenna crossed north (zero degrees azimuth), after manual setup to start recording. The digitizer was very slow: in each Pulse Repetition Period it could only process a SINGLE sample 2 km (13.33 microseconds) long. The radar operated in Long Pulse mode, in which the Period lasted 2000 microseconds. The first sample covered the range from 0 to 2 km; in the next Repetition Period the sample shifted to cover 2 to 4 km, and so on until the fortieth Repetition Period, where the sample covered 78 to 80 km. On the next one, it started over from 0 to 2 km. The antenna's azimuthal rotation speed was set at 5 rpm. By the time the fortieth sample finished, the antenna had moved 2.4 degrees in azimuth. There was no angle-time integration. This way, 40 polar cells were formed, each 2 km long and 2.4 degrees wide. In total, 150 angular sectors were produced, each 2.4 degrees wide, with 40 cells of 2 km each, totaling 80 km — that was the maximum range. Every time the radar completed a scan (one antenna rotation), 40 cells x 150 sectors were recorded on the tape, equal to 6000 bytes of information per rotation, and so on, raising the antenna's tilt angle each time, with the cycle repeating every five minutes.
The software was in charge of building the CAPPIs (Constant Altitude Plan Position Indicator) from 0 to 1 km, 1 to 2 km, and so on up to 16 km of altitude. That is, horizontal slices were built where all the data shared the same altitude. The algorithm for this polar-data interpolation used to make cartesian slices was inherited from the Canadians (via PEP in Spain), as I learned from Alamelu Kilambi at McGill many years later. Each CAPPI was printed out so it could be viewed and processed manually. In the printout, the Reflectivity value at each point was represented by a hexadecimal code (0 through 9, then A, B, C, D, E, F). The Russians called this cifrokartas, meaning digital maps, and Cubans, out of habit, called them cifrokartas too.
The system had very serious problems: 1) You never knew whether the data was being recorded correctly or not — only when the tapes reached the Computing Center and were processed (several days later) did you find out that on a given work day nothing had been recorded at all (and the planes had been flying, doing their job of seeding the clouds with silver iodide). 2) The system depended on the stability of the rotation speed: if the antenna wasn't turning at 5 r.p.m., then the angular sectors weren't 2.4 degrees each, and the last sector either failed to complete or completed before reaching 360 degrees. Really, this form of digitization was pretty crude. That's why we developed the HERMES System and RADEX.