Talk about the Lorentz cipher
Tony Sale, May 17th, 1998
Teleprinter non Morse signals with multiple tone modulation on a carrier with the normal 5 bit Baudot code
The Germans produced the Lorentz cipher machine specifically to encipher teleprinter traffic. It was important as it was used between elements of the high command and Hitler. It was identified fairly early on in 1940 by a group of policeman on the south coast who were looking for German agent transmissions in the UK. They heard these transmissions which were originally sent by a method called Hellschreiber, an early type of fax system which was difficult to synchronise and to read. The first link was from Athens to Vienna. The generic name given to all this type was the Fish traffic. The Lorentz one was called Tunny and the Geheimschreiber (the Siemens machine) was Sturgeon. All the various links were given fish names. Traffic analysis soon revealed that this was in fact the German high command. A section here was set up called the Testery under Major Tester who died about three weeks ago at the age of 97. The Testery tried to devise various hand methods to break this cipher and they didn't get very far until the Germans made one horrendous mistake. A German operator had a very long message of nearly 4,000 characters to be sent from one part of the German high command to another. He set up his Lorentz cipher machine and sent a 12 letter indicator by radio to the operator at the other end to tell him the exact settings he was about to use on the Lorentz system. This was a real time link whereas Enigma was an off-line cipher. When a key was pressed on a teleprinter at one end a few milliseconds later it struck the platen the other end. So both machines had to be exactly synchronised for it to work. The machines were set up and the 4,000 character message sent. When complete the receiving operator sent back a message in German "Didn't get that. Send it again".
Like idiots they both put their cipher machines back to the same start position; strictly verboten but they did it. The sending operator began to key his message again. Before he did so he sent exactly the same 12 letter indicator to tell the world that he was about to do the stupid thing of setting the machine back to the same start position. He started to key the message in by hand. Had he used exactly the same key strokes, as a robot would, the two messages would have been identical. But being thoroughly p****d off at having to send the message all over again he started to make differences in his key strokes. The messages begins with that well known German phrase spruch nummer (message number) and then the rest of the text. The next time being thoroughly miffed with all this he keys in spruch NR and then the text. Although NR means the same as Nummer from then on the relative character entry of the two texts was different, slipped by about 4 characters. He began to make similar changes later in the message so that the second message ended up about 500 characters shorter than the first but it was basically the same text. These messages were received in Knockholt, Kent the main intercept station for the non-Morse or FISH traffic. They sent these messages post haste to Bletchley Park because they recognised that the indicators were the same and there could be a major mistake made by the Germans. They arrived in the lap of John Tiltman, the top code breaker here, and he realised that because the indicator was the same it probably meant that the machines were set to the same start positions and he was then able to manipulate both of these cipher texts and by the combined fact that both machines were set to the same start positions and it was very nearly the same text, he was able to recover both of those texts after about ten days of hot brain cells. He got both texts out completely and because the second one was 500 characters shorter than the first one he could use those texts to reveal the key stream, or the obscuring character set, generated by the cipher machine. We, the British, knew nothing about the Lorentz cipher machine and never actually saw one until the end of the war but we now had in our hands a long stretch of pure key generated by this machine. This was given to a gentleman called Bill Tutt, a young chemistry graduate from Cambridge and a brilliant mathematician, who was asked to find out what this machine might be. Bill Tutt began to write the text out at various repetition rates because that is what you did to try and find the structure. This was B.C. (before computers) when you had to write everything out by hand. When he got to a repetition rate of 41, various patterns began to emerge in the structure of the repetitions. So 41 had some significance. Bill Tutt and the research section worked out, over the next two to three months, the complete logical structure of the Lorentz cipher machine without ever having seen it, from that one mistake the Germans made. So by the beginning of 1942 having the logical structure they got the Post Office at Dollis Hill to build them a rack of equipment called Tunny and this was the exact logical equivalent of the Lorentz machine. It used selectors and relays instead of wheels but was otherwise exactly the same. When the code breakers in the Testery had laboriously worked out the settings for each message, because they changed for every message, then they could put the settings on to this machine called Tunny and could then read in the cipher text and as if by magic out came plain German on the teleprinter; if they got it right. It was taking 6 to 8 weeks to work out by hand the settings for each message so by the time they had decrypted the messages the information was quite stale and too late to be operationally usable. Then Max Newman came on the scene and he had the idea of automating the most difficult part of finding the settings which defined the wheel start positions on the Lorentz machine. He built a first machine, called Heath Robinson, which worked well and long enough to prove that Max Newman's mathematical theories were correct. Max Newman went to Dollis Hill where he was put in touch with Tommy Flowers, who was a brilliant Post Office electronic engineer. Tommy had the idea of designing and building Colossus to meet Max Newman's requirements for a machine to speed up the breaking of Lorentz.
Tommy Flowers started in March 1943 with a blank sheet of paper on something which had never been done before. He was proposing 1,500 valves in this machine. The biggest machine that existed at that time had 150 valves in it, so it was an enormous leap into the dark into the dark.
But Tommy Flowers had the confidence to do it and it was constructed by December 1943. It was then dismantled and brought here to Bletchley Park and assembled on the piece of grass out there where F block was. It was operational in January 1944 and successful in the very first tape it ran against. The tape was loaded onto wheels and was being read at 5,000 characters per second, which is not hanging about even by modern standards using photo electric readers. It reduced the time taken in breaking the Lorentz cipher from weeks to hours. So suddenly, just in time before D day, information was decrypted in time to give Eisenhower and Montgomery the very vital information about Hitler's intentions. Hitler was convinced that the attack was coming across the Par de Calais and he was telling his generals to keep panzer divisions in Belgium. So those fragments of information gave Eisenhower and Montgomery the confidence to go ahead with D-Day. After D-Day the French Resistance and the British and American air forces bombed and strafed all the land lines in northern France forcing the Germans to use radio communication. Suddenly the volume of Fish traffic went up enormously and we rapidly built ten of these Colossus computers here at Bletchley Park; four in F block and the remainder in H block which is still standing. It is in the room where number nine stood that I have actually rebuilt the Colossus computer. So we continued to decrypt an enormous amount of messages between the German generals and Hitler. The Germans were so convinced that the Lorentz links were completely secure that on at least two occasions they sent the next month's setting sheets via the radio link and they were broken. Such gifts were gratefully received. By the end of the war, in the Newmanry Testery they had decrypted 63 million characters of high grade German text so it was a very important contribution to the war effort. Hence on D-Day, with the combined breaking of Lorentz and Enigma, of the 60 German divisions we knew the precise location of 58 of them.
At the end of the war eight of the Colossi were dismantled here, two went to Eastcote and then to Cheltenham; finally being dismantled in 1960 when all the drawings of Colossus were burnt. I started this campaign, with two of my colleagues, to save Bletchley Park in 1991. I was working at the Science Museum restoring old computers and I thought it was possible to rebuild Colossus but nobody else believed me. So I gathered all the information together which consisted of eight war time photographs and a few fragments of circuit diagrams which engineers had kept quite illegally, as engineers always do and that was it. It has taken three and half years but it is there for you to see as a fully working machine now and it breaks codes exactly as it did during the war. It is very very fast. There were two reasons for getting it rebuilt. First of all to demonstrate the speed because nobody believed it was that good and secondly; the Americans had got away with it for far too long with the myth that their Eniac computer was the first in the world, but that wasn't working until 1946. Because Colossus was kept secret they got away with it until 1970, so I made certain that when 1996 came around, the fiftieth anniversary of Eniac, Colossus was rebuilt and working here in Bletchley Park and there was a stunned silence from across the water.
The only reason I have been able to do it at all is because it was designed and built at Dollis Hill using standard Post Office components. The Post Office being what it is, nothing has changed much in fifty years, I was able to go out and raid exchanges of a vast amount of equipment. There are only eight original components but the rest although manufactured a few years after the war are still exactly the same as the original. It is a 90% authentic replica of the original machine, using war- time valves which I have collected from various places. It is a very fast machine indeed. The simulation of the same task which Colossus did, and still does, on a modern Pentium PC takes twice as long to break a cipher as Colossus did in 1944. It is great tribute to the design and skills of Tommy Flowers and his team and the code breakers here in Bletchley Park.