Tuesday, March 8, 2011

An Unexpected Turn of Events

An gift from a friend turned up in my mailbox two weeks ago. It was an envelope containing some 1960's-era computer magnetic memory ferrite toroids. The cores measure 1.9mm in diameter with a thickness of 0.5mm. These are somewhat larger than other magnetic memory cores that I've seen from that time-period.

The B/H loops shown in the above photo were made at 100kHz. Both loops are identically scaled. The left-hand loop was taken from the best of my junkbox collection of ferrite cores. The right-hand loop was made from the computer memory core. Please notice the knee at the onset of saturation for this semi-hard ferrite material is considerably sharper than for the soft ferrite loop shown on the left. Given the areas inscribed by both loops are comparable, the sharp-kneed computer memory core ought to be a more efficient harmonic generator.

I wound 98 turns of 0.09mm diameter enameled wire on one of the computer memory cores. The inductance measured 134uH. Swapping this for my previously used "junk-box" toroidal-core inductor in the 700kHz to 3.5MHz frequency multiplier, the RF output power jumped from 250mW to nearly 350mW! The total, two-stage (35x) frequency multiplier efficiency had increased to 10.8% (9.8dB conversion loss). 


My original intention had been to move forward on the project by including a 9x multiplier stage in front of the 100kHz -> 700kHz frequency multiplier. However, I now began to wonder if it might be possible to raise the final working frequency to 7MHz. With this in mind I began working on a plan for a frequency multiplier to follow the 3.5MHz stage.

A frequency doubler using a pair of DC-biased magnetic saturable cores was invented by J. Epstein in 1902 and refined by Maurice Joly in 1910. Shortly thereafter the great radio pioneer, Georg Graf von Arco, employed a cascade of these multipliers for use in Telefunken's enormous station (callsign, POZ) located in a swamp outside the village of Nauen, Germany. 

The first frequency-multiplied RF alternator of this type was installed there in September of 1912. In June of 1913 the engineers at Telefunken installed a similar transmitter at a sister station (callsign, WSL) on Long Island, at Sayville.

Nearly the entire August, 1919 edition of the Telefunken Zeitung was devoted to the history of the Nauen station. Dr. Walter Dornig (who went on to co-develop the Schmidt-Dornig shock-multiplier transmitter that I'm presently using on 80m) explained on page 65    

"Die Versuche mit statischen Frequenz-Transformatoren hatten äußerst günstige Resultate ergeben. Erreichte Wirkungsgrade von ca. 90 Prozent veranlaßten Telefunken im Jahre 1912 einen 150 kW Hochfrequenz-Generator für 8000 Perioden in Auftrag zu geben und
gleichzeitig eine Versuchsanordnung für 100 kW Antennen-Energie mit vier Frequenz-Verviel-fachungs-Transformatoren zu bauen.  Diese Anlage ergab bei den ersten Versuchen mit zwei Verdopplungs-Stufen  (8000 auf 16000 auf 32000 Perioden = 9400 m Wellenlänge). Hiermit wurde an die Telefunkenstation Sayville bei New York am 18. Oktober 1913  3°° nachmittags — also bei Tageslicht auf der ganzen Strecke — das erste Telegramm gesendet und dort fehlerfrei aufge- nommen.  Damit war der erste drahtlose Tagesverkehr mit Amerika eröffnet."

Dr. Dornig is reporting favorable results with the 
Epstein-Joly-Arco (EJA) frequency multiplier. The stage efficiency had reached 90% in the first transmitter to be installed at Nauen in 1912. They used a pair of these passive multipliers, doubling first to 16kHz and then again to 32kHz. By the use of these frequency multipliers, on October 18, 1913 the first error-free, daytime radio message between Germany and the United States was exchanged; thus opening the link between Germany and the US to commercial radio-telegraph traffic.

The British cut Germany's transatlantic cables
only hours after the start of WW1 in Europe. In her book, The Guns of August,  Barbara Tuchman wrote

"... from that moment on, Germany was sealed off from direct cable communication with the overseas world, and the burden of communication fell on Nauen, the powerful German wireless station a few miles outside of Berlin."

The radio link between Nauen and Sayville was also to be of immense historical significance in the lead-up to the United State's entry into the war. For example, the press speculated that secret messages, such as the order to sink the Lusitania and the infamous, Zimmermann Telegram, were transmitted from the station at Sayville.

In hopes of uncovering these secret messages, a Sayville radio amateur, Mr. Charles Apgar, made the very first off-the-air sound recordings and delivered them to the U.S. Secret Service.

Following the war, Hiram Percy Maxim lobbied the U.S. Congress using Mr. Apgar's exploits as a prime example of why radio amateurs should be allowed back on the air.

Grosssender Nauen was invaluable to the German war efforts. For example, the transmitter at Nauen played a part in the fascinating and true story of the SMS Wolf.

British commandos captured or destroyed all of the German colonial wireless stations soon after the war began. Strange as it seems, the takeover of the Sayville, New York station represented the first U.S. military action of the First World War, when a contingent of armed U.S. Marines marched from the train station to commandeer the radio station.

On March 2, I wrote to some friends

"Today's tests went extraordinarily well. Using my Karl Schmidt 80m transmitter as the exciter, I measured the 7MHz output RF power just shy of 150mW!  During this test the 80m input power was just over 300mW. Thus, the EJA frequency doubler stage-efficiency is approaching 50% (3dB conversion loss). These numbers are astonishing. In fact, these would be considered quite good results for a varactor-based frequency doubler. I was just saying a few days ago that I would have been happy to make as little as 10 to 40mW of 40m power. Needless to say, I've been kicking up my heels here! Please see my photo of the 3.5MHz input and 7MHz (unfiltered) output waveforms."

"Aside from the unexpectedly high efficiency, the circuit is stable and the unfiltered output waveform is remarkably clean right out out of doubler. I had it running into a quarter watt, 50 Ohm load resistor all afternoon. I couldn't resist touching this resistor from time to time, just to feel the warmth. Again, the last semiconductor used in this transmitter is working at 100kHz. A total frequency multiplication of 70x is being produced by three historical circuits employing passive, saturable core inductors.

My EJA frequency doubler is nearly an identical copy of the one used in the big Telefunken transmitters. A schematic of the Nauen transmitter may be seen here on page 67 (Seite 67), figure 61 (Bild 61).

On the evening of March 3rd, I used my Schmidt-EJA transceiver (mit gain-less receiver) to answer a 40m CQ from VE1BHH. Peter replied from Wolfville, Nova Scotia (657km) with a 569 report. The historical circuitry last used at Nauen on 32kHz was now back on the air!

Please click here to hear my 40m signal (over some fairly high QRN) at W1TXT's on-line Connecticut receiver. 

Der Geist der Nauen lebt auf!


  1. Thanks Hamilton!

    I plan to add a second frequency doubler tomorrow. 20m is going to be a real stretch. Then again, that's what I thought about 40m! :o)


  2. This comment has been removed by a blog administrator.

  3. This comment has been removed by the author.

  4. Hi Michael,
    Can you provide more information and pictures of how you made the multiplier?
    I would like to show you my results on square loop ferrites, it may be of use to you, please contact me at sv3ora AT qrp.gr