Friday, October 7, 2011

The Typex Investigation – A WWII mystery

Shortly before the start of WWII the British government began using a machine cipher for high level communications.This equipment called Typex was a modification of the commercial Enigma but with one more rotor ( five instead of four in the Enigma) plus the rotors had  several notches that ensured irregular stepping of the wheels. 
 
It’s main advantage was that it was an ‘’online’’ machine and thus could be used in the DTN – Defence Teleprinter Network.Initially it was used mainly by the RAF but as the war went on more and more were built and used by the services.
 After the fall of Tynisia to the Allies in May ’43 two German prisoners of war made a startling statement to their interrogators.During the N.African campaign an officer named Wagner had in his possession a functioning Typex machine and with the use of a book containing indicator and real settings he decoded 8th Army messages.

Here is the actual report :
TOP SECRET U                                                                                             ZIP/SAC/G.34

       THE POSSIBLE EXPLOITATION OF TYPEX BY THE GERMAN SIGINT SERVICES
 The following is a summary of information so far received on German attempts to break into the British Typex machine, based on P/W interrogations carried out during and subsequent to the war. It is divided into (a) the North African interrogations, (b) information gathered after the end of the war, and (c) an attempt to sum up the evidence for and against the possibility of German successes.
Apart from an unconfirmed report from an agent in France on 19/7/42 to the effect that the' GAF were using two British machines captured at DUNKIRK for passing their own traffic between BERLIN and  GOLDAP, our evidence during the war was based on reports that OKH was exploiting Typex material left behind. in TOBRUK in 1942.
On 14/7/42, a special German mission consisting of Inspektor HARMS and Dr. VOEGELE was proceeding to Africa to inspect captured British material, "with a view to obtaining cypher documents one equipment such as keys, encoding machines and the like".
On 21/6/43, Lt. BODE of 3/N/26 (later NFAK 621) claimed that he had been engaged on translating and emending British machine messages from 1937 until 4/6/1940. The machine was "a sort of typewriter; a man just typed the nonsense stuff, and the English came out on a tape". He said that messages were of all types, army, air force and navy. "We were, officially congratulated  after the DUNKIRK business; we had told them some things that materially accelerated the campaign". Messages which did not decode were sent to BERLIN for special treatment. BODE stated that the man who knew most about it was a certain Sgt. WAGNER of NFAK 621 who used to work the machine. Later, however, he retracted his statement somewhat, and said that he wasn't so very sure about the machine. As this P/W seemed generally unreliable, no great importance was attibuted to his statements at the time.
On 23/8/43, however, the interrogations of Lt. HAUNHORST and Oberlt. POSSEL confirmed some of BODE's less wild statements. HAUNHORST was a divisional intelligence officer and POSSEL was O.C. of He Fu 7, the senior fixed army wireless station in Africa. Both stated that they wore on friendly terms with the officers of NFAK 621, and they received information from the latter in the form of BJs.
"All high grade traffic was handled by a certain ,warrant Officer WAGNER. This man had at his disposal one or more British Typex machines captured at TOBRUK, a machine resembling a German Enigma. machine, and a special type of typewriter which had an adjustable keyboard. In addition a number of reference books were employed, from which the Typex settings were taken. The messages were examined and the discriminant looked up in a book. This gave a reference to another reference book where the actual setting of the machine was found. In certain cases this procedure did not work, and in this case certain information was sent to OKH, BERLIN who were able to supply the necessary information for the , decyphering of the traffic. All high-grade traffic is passed to BERLIN for detailed  examination, and such traffic as cannot be read by any of the above means appears to be subjected  to an analyzing machine, which, providing some evidence regarding the traffic is available, gives  results in many cases".
In a later interrogation (20/9/43)  HAUNHORST said  "There  were several machines about the size of a typewriter …. It was from OKH that those funny type-written sheets came. They got some keys out themselves, somehow. I don't how they did it, and some things they got from OKH or OKW.
He had previously described the machines as having "a central typewriter keyboard flanked on either side by complicated hollow pots …A paper strip was fed out , from the inside of the two pots. The machine had  five drums.
The bulk of the traffic was either addressed to or signed by 8th Army. "Very little traffic of the 1st Army was read’’. The average time-lag between TOI  and  receipt of the BJs was 12 – 24 hours. Great secrecy was maintained, documents were always kept locked  away and it was almost impossible to obtain permission to visit the office. BJs were almsot invariably sent to Berlin by courier.
Asked when the Company had its first success, P/W replied "Some high-grade success was obtained early in 1941, but this was sporadic". He had heard from the 0.C. that high-grade traffic was being read from the first days of the war. It was only after the capture of TOBRUK that the Company started to produce high-grade British BJs on any scale.
The interrogation of various officers of NFAK 621 mentioned by HAUNHORST yielded no results at all, HABEL  and BREMER refused to give any information, and Sgt. SEVENHECK  at first thought to be Sgt Wagner , appeared to know nothing. The identity of Wagner  remained  a mystery, and further formation on this subject was discovered during the war.
That was a summary i decided to post the actual interrogation because it gives more details :

First interrogation report on two German Army officers captured in Tynisia

  ……………………………………………………………………………………………………….

4. Cryptographic methods employed

All high grade traffic was handled by a certain Warrant Officer WAGNER. This man has at his disposal one or more British Type-X machines captured at Tobruch, a machine resembling a German enigma machine, and a special type of typewriter which had an adjustable keyboard. In addition a number of reference books were employed from which the Type-X settings were taken. The procedure appears to have been as follows:

The message was examined and the "Kennsgruppe" (discriminant) looked up in a book. This gave a reference to another reference book where the actual setting of the machine was found. The machine (Type-X) was then set and the message decoded.  In certain cases this procedure did not work, and in this case, certain information was sent to O.K.H. Berlin who were able to supply the necessary information for the deciphering of the traffic.
The documents used appear to have been of two varieties - Captured British Cipher documents giving message settings and machine settings, which the interrogator has not yet been able to identify, and documents compiled by the "Haupt Chiffrier Stolle OKH", -which appear to be reconstructions of message settings and drum settings based on a detailed analysis of high grade traffic during the past 3-4 years. From these documents it would appear to be possible, once the message setting has been deciphered to determine the drum setting's and tyre positions by reference to the previous occurrences of the message setting.

All High-Grade traffic is passed to Berlin for detailed examination, and such traffic as cannot be read by any of the above means appears to be subjected to an analysing machine which, providing some evidence regarding the traffic in available, giving results in many cases.

M.I.8(a) 23rd August 1943


Answers to GCCS questionnaire

1) What did Wagner's Type X machines look like? Obtain as full a description as possible to ensure Type X and not American Ciphering machines involved.,

Three different types of machines were in use in the decoding section -a) What appeared to be the large German Enigma Machine, i.e. the automatic machine as opposed to the usual small manual field type.

b) An American machine dimensions, 25 cm. by 35 cm. by 15 cm. No details are available concerning this machine.
c) Two  English machines having the following description:- Central typewriter keyboard flanked on either side by two complicated, hollow pots, having the appearance of the outside housing of a small electric generator. A paper strip was fed out from the inside of the two pots. The lefthand ane appeared to be the one mostly in use. At the back of the keyboard was the main mechanism of the machine, P.O.W, only once saw the inside of this 'box", which he said contained 5 drums. In addition to the machine itself there were several very attractive brown polished boxes about 15 cm. long and. with square section of about 5 - 8 cm.

……………………………………..

4) What did the catalogues prepared by Berlin look like? How were the ones set out  which determined drum fittings and tyre positions? Were they in manuscript, typed. or printed?

The Documents prepared by Berlin were foolscap sheets covered with typing. Each sheet had a complicated reference number at the top and they were all kept in a large indexed folder. The ‘’Kenngruppe’’ of the message was looked up in several "Books’’ and after a rather complicated system of cross reference, one typescript sheet was selected and given to the operator. The latter took this sheet and with its aid, "fiddled" about with the machine and then began to decipher the message. The contents of the typescript were incomprehensible to P.O.W. and appeared to consist of a maze of letters and figures.

These typescript documents arrive by courier from O.B.S Rome who in turn received it from Berlin. They do not appear to arrive regularly, but at intervals and in packets of several sheets. Very considerable security measures are taken to assure the secrecy of the document.

7) What were the most common Allied addresses in the more secret BJ’s? Did American authorities appear frequently in the addresses?

The bulk of the traffic was either addressed to or signed by 8th Army. The traffic was almost entirely concerning the "G" brandies of A.F.H.Q., 8th Army and the Corps subordinate to 8th Army. The links were those between 8th Army and A.F.H.Q, and 8th Army and subordinate Formations. This traffic accounted for 80%  of the B.J. material.

The remainder was for the most part "Q" traffic of the American 7th Army and A,F.H.Q. Very little traffic of the 1st Army was read.Unfortunately  P.O.W cannot recall any verbatim addresses, and in any case, these had been translated into German equivalent, to, I (a), 0 (qu), I(b), etc.

…………………………………………………………………………………..

12). How often were the BJ’s sent to Berlin by wireless teleprinter ?

B.J.’s were hardly ever sent to Berlin by wireless T.P. . POW has only heard of two instances. As a general rule they went by courier every day. The intercept Coy NEVER  sent any themselves this was the duty of I(c) at Heeresgruppe.

…………………………………………………….

15) When did  they get into their stride?

It was only after the capture of Tobruk that the Coy. started to produce High Grade B.Js an any scale. It was at Tobruk that the two British machines together with documents wore captured.

1st September 1943

The Brits were obviously alarmed by those reports.Typex was their highest level cipher machine. Immediately they started an investigation on the fate of the Typex machine used in Tobruk.However since the personnel there had been taken prisoner when the city fell to the Germans in ’42 it was only possible to interview the persons responsible after the war.The testimony said that they had thoroughly destroyed the only Typex machine used there.

A high level meeting was also called to discuss appropriate measures.One of the people involved was the famous cryptologist Gordon Welchman of ''Hut 6''  fame.Some of his comments in the reports are interesting :

As regards breaking I have always felt that the Germans could not be breaking any of our Type X traffic because if they were they would take steps to prevent us breaking their enigma traffic.But this may be overestimating the efficiency of the Germans; after all we ourselves have made no serious attempt to use the experience of the experts on breaking the German enigma  to improve the security of our Type X. Alternatively  the Germans may have been breaking Type X before the introduction of pluggable heads by catalogue methods which would be knocked out by their stecker.

………………………………………………………………………………………

 
On the whole I feel a thorough investigation would be a good thing but I don’t see who could do it.However it may be possible to shoot down the Wagner story after further discussion here and further interviews with POW’s .It is quite possible that Haunhorst was merely shown how the Typex machine worked and it would be interesting to know whether he actually saw an  English message decoded.

W.G.Welchman

Despite their efforts the Brits were not able to identify the mysterious ‘’Wagner’’ during the war or after.Their inquiries however concentrated on the Signal Intelligence Agency of the Army High Command - OKH/GdnA .After the war they interrogated Mettig who was in charge of the  Army Agency from Nov '41 to Jun '43 :


Statements by PW Obstlt Mettig on Typex   
 
1. In homework completed about 24 Jul 45, he wrote (exact translation): ‘’The Referat Zilmann, despite great efforts, was unable, to break the English cipher machine. It is true that during the campaign in 1940 several English cipher machines were captured but in all cases the wheels were missing." It should be noted  that PW never mentioned TYPEX by name.

2. On Sunday, 29 Jul 45,  in consultation with major  Morgan, asked PW directly whether the Germans had worked on Typex. The reply was in the affirmative.PW  was then asked had the Germans been successful. He replied without any hesitation that they had been. Asked to a date for this success, he stated  spring 1942. The success had not been maintained owing to the capture of the Seebohm company in 1942 near ALEXANDRIA, which led to a tightening up of British signals security. No further inroad after that date is known to PW. IO then asked PW whether he was sure he was referring to TYPEX and not to War Office Cipher, which the Germans had cracked; PW replied categorically that he was referring to Typex. In reply to further questions he added that the breaking of Typex was done partly in BERLIN and partly in AFRICA; he assumed that the success was achieved, cryptoanallytically and not through compromise. Had the latter been the case, he would have known, as all compromised material secured from Abwehr or other special sources was passed to him as head of in 7/VI for distribution to particular sections .PW, was told to write a detailed memo on this subjects in particular he was told to name  all individuals concerned in breaking  Typex.

3. On 31 Jul 45 IO, after 'phone conversation with Cmdr DUDLEY-SMITH, visited PW to collect memorandum. PW declared that all his statements re Typex in para 2 were rubbish and that the machine was unbreakable. Although he noted the astonished expressions of the visiting officers (Major Morgan  and Lt Col Lewis) it only dawned on him when he returned to his room what was the  significance of what he had said. IO then asked why he had been so precise in his statements about the breaking and had given  dates etc.PW  replied that he did not know what he was saying and that he must have been thinking, of some minor procedure as double transposition or syllabic cipher. IO informed Cmdr Dudley-Smith immediately; it was agreed to postpone action in this matter until it had been exhaustively discussed.

CSDIC (UK)
Sept ‘45


So instead of resolving the Typex affair Mettig made things worse.Now the British had three people claiming Typex was read ,one of whom was the man in charge of the Army Signals Intelligence Agency.Further investigations singled out the top mathematicians of OKH/GdNA Doering and Pietsch plus Menzer of OKW/Chi:

From the information so far obtained it seems probable that if Typex was ever broken it was done exclusively at OKH and the information was not passed on to other departments .The men most likely to know about this subject are Pietsch , Doering and Menzer .Meanwhile it would be of the greatest interest to see how Haunhorst reacted to further interrogation.
G.C.C.S. (S.A.C.)
15th October  1945

................................................

‘’Strong efforts should be made to apprehend the men who were chiefly responsible for Typex work in OKH – Pietsch ,Doering ,Menzer , Luzius ,Schulz , Zillmann. Fregattenkapitaen Singer should be placed on the Brown list.’’

 G.C.C.S. (S.A.C.)
20th October  1945


I have been unable to find any proof that Pietsch and Doering were interrogated by the Allies after the war.It’s highly probable that they slipped the net.Inspector Menzer also disappeared after the war.According to an NSA article he lived for a time in East Germany but after being thrown in prison for 6 months by the Russians he surrendered to the American authorities in West Berlin in May 1949.No other information is available on that group.

So what are my thoughts on the possibility that the German actually exploited Typex ?

Possible explanations :

1).The Germans one way or another managed to get their hands on the ‘’black’’ typex rotors.Using them with their captured machines and figuring out the indicator process they decoded messages sent by 8th Army.The problem they had with 9th Army can be explained by the fact that a plugboard was used with their Typex machines.A fact not known to the Germans.

Of course this hypothesis would generate another question : If Typex was insecure that would mean that  so was   Enigma ,so why did they continue using it? Since they continued to use it it means that they never managed to read the Typex…
Again this question can’t be answered conclusively .

First of all for the Germans the Typex was not directly comparable to the Enigma I because it lacked a plugboard.Without this modification it was a simple commercial Enigma with one more rotor.So breaking the Typex would not mean that Enigma was compromised ,just that the commercial version could be solved.A fact that was not new to them since they themselves attacked and read the Swiss diplomatic Enigma (no plugboard).
The Enigma machine was also constantly upgraded in terms of security.For example in 1942 the Germans split their traffic in many different ‘’keys’’ and the U-boats got the 4-rotor Enigma.In late ’43 early ’44  they introduced  the Uhr device and the UKW-D rewirable reflector in Luftwaffe links and the Army Home administration network ( Source:UMKEHRWALZE D ,Cryptologia article ).Operators in ’44 had to change the plugboard settings thrice daily.From late ’44 each U boat used a separate Enigma ‘’key’’.Those are a lots of measures for a device that was considered secure…

Another problem if they had success is that  high level personnel all claimed Typex was secure. For example Huettenhain and Voegele ,the chief cryptanalysts of the OKW and OKL respectively ,stated several times that Typex was never read.Wouldn’t they know of such an accomplishment ? Considering that the people of OKH didn’t tell Huettenhain about their success with War Office Cypher it could be argued  that they’d also keep the Typex affair secret.

2). Mettig was probably confused by the similar names that the Brits gave their codes (Typex,Codex,Slidex, Morsex ,Sheetex).The code mentioned by Mettig was probably the War Office Cypher which the Germans read in N.Africa from March or August ’41 till January ’42.According to a POW assigned to KONA 4(signals regiment) in Athens from that time on it was not solved.This may be true for forward units but Mettig in a detailed report mentions reading it in summer ’42.So the main agency in Berlin had success with that code.

This solves one part of the equation but what of the Typex machine used by Wagner? It could be that ‘’Wagner’’ used a teleprinter and the POW’s mistook it for a Typex.However how could the Tynisia POW's know that the British machine had five rotors and a drum on each side? This was not the layout of German equipment.The only explanation that I can come up with is that the machine they saw was a captured Typex modified to work with German rotors.

The Luftwaffe actually used captured Typex machines with three German rotors for it’s communications.Those machines were modified by  Inspector Menzer of OKW/Chi .Still 3≠5 rotors….

All the information presented comes from two reports in the British archives :

HW 40/88 : Investigation into POW reports that German Sigint authorities (NFAK 621) were exploiting TYPEX (British cypher machine) in North Africa

HW 40/89: Investigation into POW reports that German Sigint authorities exploited TYPEX (British cypher machine)


Both files are available from my Google docs and Scribd account.

I suspect that there is a third part for the Typex investigation.Unless it’s released questions will remain.A small piece of information regarding the efforts of OKH codebreakers versus Typex is available however from Ticom D-83 which based on captured documents says :

  ‘’However , by about May ,1941, OKH were in full possession of all the theory necessary for solving the problem and there can be little doubt that ,had they been lucky enough to capture a set of drums at Dunkirk as well as the three machines the bulk of the Typex traffic up to July 1940 would have been read’’.
So IF the Germans had the rotors they would be able to decode Typex….

1 comment:

  1. enigma traffic was supposed to be tactical (and short). type-x traffic was supposed to be high-level (and long).

    The manual breaking of type-x 5 drums, with improved staggering, does not seem incredible. Without a plugboard in type-x, plenty of evidence is available for decoding analysis before the first stagger of the third wheel takes effect (adding complexity), etc.

    ReplyDelete