Signalspaning: Link

Introduction - The Norrköping Enigma messages

By Dan Girard

Bild "German ENIGMA cipher system examples"
Example Page 40
While searching for information about the Allies’ code-names for different Enigma key-nets, I found an interesting PDF document. It contains a number of radio messages intercepted by the Swedish security police in December 1939 and in February, March and April 1940, including more than twenty intercepts that, judging from their formats, are almost certainly naval Enigma messages; and also a few that look like either German Army (Heer) or Air Force (Luftwaffe) messages. The naval Enigma intercepts are on pages 22, 23, 24, 25, 37, 39, 40, 47, 49, 69, 70, 71 and 81 of the PDF; and the five Heer/Luftwaffe intercepts are on pages 40 and 47. Most of these were intercepted by the police radio station at Norrköping. The pages of the PDF seem to run in more or less reverse chronological order — for example, the intercepts on page 40 are from the 7th of March 1940, and those on page 47 are from the 20th of February 1940.

I've tried my hand at breaking some of these, both with a hillclimber and with my bombe simulator, but so far without success in the case of the Naval messages. However, I have broken the first two Luftwaffe messages from page 40, and also the one from page 47, with my hillclibmer. The first two are definitely Luftwaffe, probably in the key called "Red" by the Allies. I think the one from page 47 is probably in the Luftwaffe key called "Blue" by the Allies; this key was used only for practice messages, and the plaintext of this message appears to be just a passage copied from a radio manual. The other two messages from page 40 do not break on either key; they must be either Army or a different Luftwaffe key. I haven't been able to break them yet.


I've solved the other two Luftwaffe messages of 7 March 1940, from the PDF file of radio messages intercepted by the Swedish police. Like the first two that I had broken before, these turned out to be just routine activity reports (Tätigkeitsberichte); but evidently these are on a different key-net. I had tried my bombe-simulator on both of these messages, with the word "Taetigkeitsberiqt" as a crib for the message beginnings; but without success. I then tried them with my hillclimber, and it eventually reached a break on the second of the two messages. Although the Grundstellung was missing from that intercept, I was able to use the Enigma settings from that break to decipher the first message and use its Grundstellung and enciphered message key to work out the full ring setting. Once I had the full ring setting, I was able to reconstruct the Grundstellung of the second message.

Once the messages were broken, it became clear why my bombe-simulator had failed. In one of the messages, the enciphering operator had forgotten to replace the "ch" with "q", and there was also another incorrect letter, so that the word came out as "Taetigkeitsbiricht"; and in the other, the operator had spelled the word with an extra 's', "Taetigskeitsberiqt".

Although the contents of these messages are simply routine and of no particular interest, there is something interesting about their keys: the Grund settings, "VLG" for the first message and "VJF" for the second, are both very close to the daily key's ring setting, "XLG". These are good examples of the "Herivel Tip" , or "Herivelismus", as it is sometimes called, wherein a careless or hurried operator would sometimes forget to randomize the wheel positions after setting the rings during the daily key change, and would then use as the Grund setting of the first message of the day whatever letters happened to be showing through the windows of the Enigma machine. These would likely be the letters of the ring setting, or very close — within one or two letters.

I wish I had thought to try the "Herivel Tip" on the first message. Since the ring settings of the middle and fast wheels are the same as the positions of these wheels in the Grund setting, which is given in the clear in the message heading, trying just those ring settings with my hillclimber would probably have broken the message more quickly!