KL-29= CSP-2900 = BACCUS = GORGON = DEMEDOR. The KL-29 was a Navy modified ECM.
In the USAF, the KL-29 system was code named Gorgon. The code names Bacchus and Gorgon were also used during the 1950s as designators for CSP 2900 based systems. It is not known which users referred to the machine as DEMEDOR.

Don Watson used this machine and he recalls the following. " The KL-29 was an on-line machine  which used 5 rotors about the size of a donut. It was mounted in  a 7 foot high cabinet and connected to a Teletype machine. The rotor drawer rolled out for easy access. We reset them out every 12 hours --- if we went over, the system was assumed to be compromised. To change the rotor settings, we went to condition "Lo".

Counters for sender and receiver had to be set the same before transmitting again. We were USAFSS, 1956, got this system from the US Navy cryptology division, battleship grey, of course.The ones we had before were maybe 3' tall, looked like a small desk. There was a toggle switch on the front to "Go Lo" and back up. Rotors were the same for both systems, but on the smaller system we opened a door on the top to access them. We used erasers to clean up the rotors a couple of times in an 8-hour shift.  I remember the code name BACCUS -- I think that is what we called the small desk-like assembly before the Navy's advanced system arrived.

"Go lo" (go low) was what we typed when we wanted a distant Gorgon operator to shut down on his end --- ie stop transmitting. He turned off the TD that was transmitting a teletype message to us, in code to anyone who might be monitoring us, but coming out in "plain text" on our end. He then flipped a switch on Gorgon. That's how the crypto part was shut down. Now we typed on the KL-29 keyboard "in the clear" over telephone wires or over radio-teletype, using typical codes to explain why we wanted him to shut down (go lo).

On many an occasion we typed ZBK2 (garbled transmission). Something had happened to cause the far end  system and ours to get out of sync. We'd clean the rotors on both ends, then we would tell the far end what setting our counter was on. He would tell us what his was on. Still "in the clear" (anyone monitoring us could read what we were typing back and forth) we would tell him "Go hi" or simply "GA" (go ahead). He and us would switch the Gorgon switch to the right, almost simultaneously. Then after 5 or 10 seconds he would start transmitting again. If the system was in sync and was operating properly, his coded transmission would begin again, and we would read it plain text. It worked the same if we were transmitting to him and he asked us to "go lo". During transmission sometimes a tape would get stuck or tangled up. When that happened we stopped the TD, cleared the obstruction, and, without any need to "go lo", we simply typed this:

---H/W IMI---

That meant we were backing the tape up a little, repeating a section. The garbled portion would be repeated, but accurately, if we had fixed the problem. If we couldn't do it accurately, we had to have at least that portion of the tape retyped. Once the transmission had reached 5400 characters, we had to end it with the following on the tape:

Two carriage returns
8 line feeds
4 "n's"  (nnnn)
Twelve letters   (nothing on the screen but part of the signal to shut down the transmission).

We then picked up with Part Two of the same message, with a new "header". The thought was that messages in excess of 5400 might cause Gorgon to start repeating itself, and, more practically, the tape became too bulky to roll up in typical "butterfly" fashion.

In some cases our "landline", usually telephone wires, would start "running open". Something had caused a break in the line. This could have been due to the failure of a relay, a line accidentally cut, or a line purposely cut in sabotage. I was the only one on our shift who could speak the foreign language necessary to call the governmentally controlled telephone company to get us back in operation. When they didn't do that quickly enough, I'd strap on a .45, grab an M1 Carbine, hustle up a jeep, load it with teletype tapes in duffel bags, dust off the jeep seat for the Captain who rode with me, also armed, and away we would go to another facility 20 miles away, enter the classified area, and begin transmitting the tapes from there. I don't know what we would have done if we had been ambushed by unfriendlies.
Thank God, never had to find out.

Lots of us on the shift could type faster than the off-line teletypes. We'd start "poking" a tape, insert it in the off-line TD, then type as fast as possible while the hard copy was printing. Early machines transmitted at 60 wpm --- piece 'a cake, right? --- then the 100 wpm teletypes came in. Now some of us had to type, accurately, at 110 wpm. Believe it or not, I've done it, and I've watched others do it, one guy in particular, who had been in USAFSS for 11 years prior to arriving at our location. He was amazing. By the time I left the AF, transmissions of entire messages were just a "beep" over secure pathways --- no more 100 wpm. In civilian life folks were amazed at how fast I could type on a simple typewriter. What memories!! I'd do it all again in a heartbeat. IF ONLY......."


1) Don Watson <dwats(at)cox.net>
2) KL-29 designators  http://www.maritime.org/csp-designators.htm

Back To Menu Page
Aug 26/07