The picture series I introduced in my last post contains a photograph of a 1930s electronic machine. It seems completely unknown how it was used and if it was a cipher machine at all.

In my last blog post, I reported on a series of photographs showing cipher machines built by US constructor and business man Edward Hebern (1869-1952). These pictures are part of a large collection the Smithsonian Institute recently put into the public domain. Hebern never sold his devices in larger amounts, so he had to give up his business after a few years. Perhaps, the following postcard Hebern had produced found more customers:

Source: Smithsonian Institution

The vacuum tube machine

One of the pictures I introduced in my post is the following:

Source: Smithsonian Institution

I originally thought that this is an encryption machine using lightbulbs to display the ciphertext letters, but my readers Richard SantaColoma and Kerberos corrected me. In fact, the glass objects on the right-hand side of the picture are not bulbs but vacuum tubes. This is interesting, as one wouldn’t expect vacuum tubes in a machine that was supposedly built in the early 1920s. In addition, Richard SantaColoma pointed out that the tubes we are dealing with here (Raytheon 6A5G) were introduced only in 1935. They were primarily used as amplifiers, but in this case they may have served as contactless switches.

Here’s the description of the machine on the Smithsonian website (thanks to Ralf Bülow for posting it). To my regret, this text is not very helpful, as it neither provides the year when the device was built nor the way it worked. A wiring diagram would certainly be useful, but is not available on the Smithsonian page, either.

At least, a few more pictures of the same machine can be found in the collection. Here’s a closer shot of the vacuum tubes:

Source: Smithsonian Institution

And here’s a label of a condenser:

Source: Smithsonian Institution

The following label belongs to the electric motor of the machine:

Source: Smithsonian Institution

No year is shown, but the patent that is referenced was issued in 1934 (thanks to Gerd for this information). The following picture shows the machine from a different angle:

Source: Smithsonian Institution

Conclusion

As it seems, we are dealing with a machine from around 1935, which means that it was built much later than all other Hebern cryptographs I am aware of. The device involves, among other things, a typewriter keyboard, a motor, and an array of vacuum tubes. So far, I don’t see a proof that this device was really constructed by Edward Hebern, and I’m not sure whether it is a an encryption machine at all. If this design was indeed meant for encryption, it is completely unclear to me how it worked.

The only source about Edward Hebern’s life after 1930 is the article by Glenn Zorpette I already referenced in my last post. Zorpette writes:

He [Hebern] remained active as an inventor, and not just in cryptography. He filed a patent application in November 1938 for a “printing telegraph system,” which converted electrical signals of different time duration to letters or numbers. The signals could be sent by either radio or wire.

To me, this vacuum-tube device doesn’t look like a printing telegraph system, but I may be wrong. Perhaps, a part of the machine is missing.

Can a reader say more about this unusual machine?


Further reading: Update: A complete (?) list of German cipher machines in World War 2

Linkedin: https://www.linkedin.com/groups/13501820
Facebook: https://www.facebook.com/groups/763282653806483/

Subscribe to Blog via Email

Enter your email address to subscribe to this blog and receive notifications of new posts by email.

Kommentare (16)

  1. #1 Richard SantaColoma
    https://proto57.wordpress.com/
    25. März 2020

    Thank you for the mention, Klaus.

    As I also pointed out your other post comments, if Hebern was using these triodes as switches, in some sort of logic array, then it might qualify as one of the earliest examples of this being done. Perhaps one of your readers would know more about the circuits of electronic cipher machines, and would have an idea how 30 switches in an array might make sense. For instance, could it be done to emulate, electronically, a series of rotors, for enciphering and deciphering?

    “Later, triodes were used in the first electronic computers. The earliest example was the Atanasoff-Berry computer:

    https://en.wikipedia.org/wiki/Atanasoff%E2%80%93Berry_computer

  2. #2 Detlev
    Leverkusen
    25. März 2020

    The Smithsonian Institute should be able to draw the circuit diagram so we can understand the function.

  3. #3 ShadowWolf
    triode logic
    25. März 2020

    Basically tube logic is the same as modern transistor logic. You just have a different active element and the voltages are different.

    The big problem is it requires two triodes to make just 1 basic gate in many cases. The other problem is some gates would require a dual control gate tube. If you look up Eniac and some pictures of those modules, you’ll find they are quite large and they were using dual triode tubes.

    The most logical thing is that it is just a driver for some sort of electronic printer. It would be easy to verify if there was a schematic available.

  4. #4 Detlev
    Leverkusen
    25. März 2020

    You have to understand the circuit. There is a connection from the keyboard to the tube array with brown wires, the red wires from the tube array end in a metal cylinder. One cut cable comes out to the cylinder. And then?

  5. #5 Peter
    25. März 2020

    Wenn man auf der Seite https://americanhistory.si.edu/collections/search/object/nmah_1214327
    Ins Suchfeld „Edward Hebern“ eingibt, kommt man auf die Dokumentation des ganzen Materials. Es handelt sich meines Erachtens um den Inhalt einer „Bastelkiste“, bestehend aus mechanischen und elektrischen Einzelteilen, Rotoren, fertigen Chiffriergeräten und anderen Baugruppen.

    Zu finden sind unter anderem

    27 Solenoide in einer Baugruppe, 1991.3033.120, Seite 5

    Ein Koffer mit einer Chiffriermaschine (4 Rotoren) und einer Schreibmaschine + verdeckten Elektronikteilen. 1991.0190.02, Seite 7. Dazu erklärender Begleittext, der Koffer misst 20.5 cm x 61 cm x 42 cm

    Ich glaube, Herr Hebern arbeitete an einer Chiffriermaschine, die den chiffrierten Text auf Papier drucken soll. Dabei sollten die Typenhebel über die Solenoide angesteuert werden. Da dies ziemlich hohe Ströme benötigt, wurden die Kontakte an den Rotoren durch die beim Ausschalten des Stromes entstehenden Funken rasch zerstört. Wenn diese Ströme durch Vakuumröhren geliefert werden, geschieht das nicht, solche Röhren sind gegen hohe Spannungen ziemlich unempfindlich. Man kann zudem das Ausschalten langsamer gestalten, was die entstehende Spannung vermindert.

    Dazu kommen drei Items mit Dokumenten:
    Legal documents 1991.3033.127, Seite 4
    31 Fotografien 1991.3033.125, Seite 4
    Blueprints, tracings and drawings, 1991.3033.128, Seite 6

    Wenn jemand gute Beziehungen zum Museum hat, kann er vielleicht die Digitalisierung der Dokumente erreichen, dort dürften dann die Schemata zu finden sein.

    Peter

  6. #6 Detlev
    Leverkusen
    25. März 2020

    Peter, Deine Erklärungen sind sehr plausibel. Wenn jeder Leser auf der von Dir angegebenen Seite bei den Blueprints “Nominate this object for photography” anklickt, wird das vielleicht noch was.
    Detlev

  7. #7 Klaus Schmeh
    25. März 2020

    David Tillemans via Linkedin:
    “This is really nice and impressive.”

  8. #8 Gerd
    25. März 2020

    Peters Erklärung, nachdem das Array aus Röhren eine Art Ausgangsverstärker mit 30 Kanälen für ein Druckwerk ist, klingt plausibel. Es gehen auch etwa so viele braune Drähte in das Gerät rein wie rote Drähte herauskommen. Was mich dabei stört ist: Ein solches Gerät mit 33 Röhren muss zur damaligen Zeit eine sehr teure Sache gewesen sein. Warum baut man so etwas aufwendiges für ein Chiffriergerät, das mit einer einzigen Walze eher den Sicherheitslevel “Micky-Maus” aufweist? Es scheint kein verkaufbares Produkt gewesen zu sein.

  9. #9 Detlev
    Leverkusen
    25. März 2020

    Vielleicht war das eine Studie für ein abgesetztes Druckwerk.

  10. #10 Peter
    26. März 2020

    Zu Gerd, #8

    Ich vermute, das Gerät ist für einen “proof of concept” gedacht, z.B., ob die Kontakte immer noch abbrennen. Dafür genügt ein einziger Rotor. Dedshalb auch der Motor, der ja für ein Chiffriergerät unsinnig ist, die Walzen werden üblicherweise durch den Tastendruck bewegt. Der Motor hat zudem ein Untersetzungsgetriebe, so dass eine einigermassen realistische Druckgeschwindigkeit möglich scheint.
    Beweisen könnte man das alles wohl nur mit den ebenfalls eingelagerten Dokumenten.

    Peter

  11. #11 Christian Berger
    26. März 2020

    @Peter: Was leider das Problem an der These ist, ist dass Röhren typischerweise keine hohen Ströme liefern können. Man müsste also mit Spulen mit sehr vielen Windungen und hohen Spannungen arbeiten. Das bringt ganz neue Probleme.

  12. #12 Peter
    26. März 2020

    zu Christian, #11

    um 1940 war das kein Problem, es gab nichts anderes als Röhren. Das Datenblatt gibt den Arbeitspunkt für A-Betrieb bei 60 mA Anodenstrom, die Röhre lässt sich also kurzzeitig bis ca 100 mA aussteuern. Ein Fernschreiber Siemens T37, den ich mit Röhren betrieben habe, benötigte 20 mA bei 150 V für das Solenoid.

    Peter

  13. #13 Ralf Buelow
    26. März 2020

    Here is a NSA machine with vacuum tubes from 1952, and the tubes were apparently used for timing signals, see https://users.telenet.be/d.rijmenants/en/kl-7.htm

  14. #14 Kerberos
    27. März 2020

    Hallo Ralf,
    a small correction, there is only one vacuum tube
    in this machine (12AX7), the other three are 2D21,
    Hydrogen Thyratrons.

  15. #15 Dirk Rijmenants
    27. März 2020

    Fascinating pictures, great find, Klaus! From what I can see from the picters, this machine cannot be used for encryption, as the only visible key variables seem to be the rotating contacts. This is clearly a proof of concept.

    To have an idea of the function of the tubes, we’d need a wiring diagram. The tubes could be used in many ways, not only for switching or logic, but also simply as diodes, to create a logic grid, f.i. a setup like the Hagelin B-21 ( https://www.cryptomuseum.com/crypto/hagelin/b21/index.htm ).

    Unfortunately, this machine doesn’t seem to have an output, which obviously also points to prototype. There are however enough (fixed) contacts and wires that later on could be modified into plugboards or outputs, once the prototype were to be developed further. Sometimes, digging up lost patents or other drawings can shed more light on a prototype than having acces to the hardware.

    I think Peter’s comment is the closest you can get. Many parts, many ways to use those tubes and many ways to arrange those wires or add plugs. You could call it Hebern’s Arduino precursor to fidle around with ideas. Fascinating mystery.

  16. #16 Detlev
    Leverkusen
    27. März 2020

    The contacts are labeled with A to Z and there are also some unnamed. The red output wires end in a small cylinder laying between the keyboard and the tube array.