Before the advent of high power vacuum tubes, one of the only practical methods of generating large amounts of RF energy was through the use of a motor-driven, high frequency alternator. Depending on the power level and frequency, some of these alternators were of massive weights and dimensions. As technology advanced, these old, electro-mechanical workhorses of the airwaves were scrapped one by one. The Alexanderson alternator was a popular model in its time, and today, only a single perational example survives in the whole world. That machine, whose power output is 200 kw, is located at the Grimeton Radiostation, 10 kilometers (6 miles) east of Varberg on the Swedish west coast. While no longer in commercial service, station SAQ is still in an operational state and usually makes annual transmissions on 17.2 KHz.
During WWII, Alexanderson alternators located in Hawaii were the only reliable long-haul method of communicating with the Pacific submarine fleet 24 hours per day, seven days a week. In the 1950's, the USAF used an Alexanderson machine located at Marion Massachusetts, near Cape Cod . The last one in the United States was still in service until 1960 when it was donated to the Smithsonian Institute.
Edwin (Ed) N. Myers, Lt. Col., USAF (Ret'd) helped to commission the unit installed at Marion Massachusetts, call sign AFA2. The frequency of operation is not verified at this time but it was either 25.82 or 28.52. KHz. He provides some background information about it. "The US Air Force Communications Service bought or perhaps leased that one from RCA. The purpose was to send Weather Sequences via RTTY to weather stations in the Arctic - especially to the new base at Thule, Greenland. This VLF transmitter was the only way to punch through the Northern Lights. There were no satellites or ionospheric scatter techniques around in 1951. I was the Air Weather Service individual in the group and based in the 8th Weather Squadron in Newfoundland at the time. The main Air Force Communications Service rep was an RCA Tech Rep (H. John Heffernan, now deceased) also stationed in Newfoundland.
We were both based at Ernest Harmon AFB in Newfoundland. We both spent time at Marion helping to get the station on the air. There were quite a few problems that we had to solve but when it was finally debugged, we could sent RTTY at 60 wpm. Problem was, the Comm Service Headquarters thought the alternator it couldn't be keyed faster than 20 wpm so they ordered 20 wpm gears for all the teletype machines in the Arctic!
Instead of assigning any electronic technicians to Marion, the Comm Service assigned mechanical techies. The Comm Service had also hired a retired RCA engineer named Mr. Eschelman who had worked there when RCA ran the station. There were 100% spares available at the site and none were ever used - that was one reliable machine".
|1951: This was the alternator installed at Marion. The machine was actually on the air when this photo was taken. All Alexanderson alternators were manufactured by The General Electric Company since its developer was a GE employee. (Photo by Ed Myers)|
|The contacts on the keying relays were the size of a half dollar coin and there was a blast of air directed at them to blow out the arcs formed during keying. Sometime after 1952, the keying relays were replaced with thyratrons. This must have greatly reduced the noise level in the building. (Photo by Ed Myers)|
|1951: Mr. Eschelman (left) is meeting the Air Force team at Marion. (Photo by Ed Myers)|
|1951: Aerial view of the Marion site showing the array of antenna towers. At each pair of masts supporting the flat top array there was a reactor connected from the flat top to ground. The reactor at the building end was a saturable reactor. This reactor was keyed with a large DC current via the keying relays inside the building. With "key down" this reactor and the others resonated at the alternator operating frequency. During "key up", the alternator output was diverted to a dummy load. This kept the load on the alternator fairly constant so the speed wouldn't vary. There was no signal on the antenna during key up. (Photo taken by the USAF Air Weather Service. From the collection of Ed Myers)|
On October 17 1996, between 0900 and 0930 hours UTC, it was planned that the Grimeton alternator would be used to send test transmissions on 17.2 kHz (17,442 meters) in conjunction with a ceremony declaring the Grimeton Radiostation a historic, memorial site. The call sign of SAQ was assigned and a CW message would be sent twice in English at a speed of 12 wpm. The keying principle of the machine is interesting. With the key up, power is applied across a dummy load; with key down, power is sent to the antenna. This is accomplished with a magnetic amplifier circuit (also an Alexanderson invention) that changes a coupled circuit from being non-resonant in a key down state to being resonant at key up state in order to absorb the radiated load. If this was not done, there would be speed variations between the 'load' and the 'no load states' resulting in frequency excursions. The antenna at Grimeton is 2,280 meters (1.42 mi) long.
The alternator itself runs between 2,500 to 3,000 rpm, and needs to be set precisely to the correct speed in order that it's 300 poles will generate the required frequency of 17.2 kHz. In order to spin up such a monster, a motor in the range of hundreds or horsepower is required. As one can imagine, the noise level would be very high when the alternator was running at full speed. In their heyday, most alternators ran on frequencies between 10 to 30 kHz, although a few ran as high as 100 kHz. Some of these machines could be up to 50 feet long and weigh many scores of tons. Even the smallest alternator rated at 2 kw at 100 kHz, weighed in at several tons and ran at a speed of 20,000 rpm. It would generate over 15 amperes of RF current into the antenna at 130 volts. Imagine what a higher power unit would do! Antennas for alternators were usually in the order of 1 to 3 miles long, depending on the operating frequency. Radio Central, which used to operate on Long Island, NY had its antennas arranged like the spokes of a giant wheel in order to reach anywhere.
Except for the northeastern seaboard of North America, the signals were essentially unreadable on the day of the test. Some copy was possible around Boston, but it was marginal at best. The transmission started a bit earlier than planned at 0852 UTC with a message in English that was later repeated four times by different operators. The final transmission ended at 0909 hours UTC.
This is a copy of the message that was sent:
CQ DE SAQ CQ DE SAQ CQ DE SAQ.
THE RADIO STATION GRIMETON IS NOW DECLARED A LISTED
HISTORIC BUILDING BY THE COUNTY COUNCIL OF HALLAND,
SWEDEN AND GOVERNOR BJOERN MOLIN. THIS MESSAGE IS
TRANSMITTED BY A 200 KW ALTERNATOR, CONSTRUCTED AT GE
BY ERNST ALEXANDERSON 80 YEARS AGO. BY SM0AGP.
73 TO THE RADIO CLUB OF AMERICA FROM SETH MYRBY + SK
Seth Myrby SM0AGP, is head of the government agency for radio communications
in Sweden. The main part of the message was then repeated by three other
operators SM6DM, SM5BF, with the last one unknown. Bengt Wilander SM7BKH,
made a successful recording of the transmission although his location was
300 km south of Grimeton and the receiving equipment not very suited for
17.2 kHz. An old DRAKE low frequency converter feeding a Collins 51S-1
receiver and an 80 meter dipole made it possible to copy the signals. In
the VLF bands, noise is 'king', so it takes some pretty sophisticated techniques
to receive weak signals. Alan Douglas, a well known radio historian and
author of the three-volume "Radio Manufacturers of the 1920's", was one
of the very few individuals who heard SAQ's transmissions in the United
1) Rick Ferranti WA6NCX, Alan Douglas, Bob Keys NA4G, Sandy Blaize W5TVW and Bengt Wilander SM7BKH who provided the source material for this article.
2) Ed Myers <emyers2(at)cox.net> Lt. Col., USAF (Ret'd) Alexandria, VA 22308