The Orestes cryptosystem employed the KW-7. It was an on-line, send/receive crypto unit installed in shore stations and aboard ships. In one application, it was used for ship to shore and for intership radioteletype communications. To send messages over a secure UHF teletype circuit, a model 28 Teletype or reader (T-D) sent the prepared message to the KW-7 which in turn keyed a UHF transmitter in the AM mode. The KW-7 was also used aboard aircraft such as EC/RC-135's.
KW-7's saw service in three variants. The original version used wire cords to set the daily key. Next, the plugblock version was introduced while the last variant used a paper card reader. Externally, one could tell the machines apart by the front covers. The original version of the machine had a flat, front cover. When the plugblock version was introduced the cover had a small square bulge. When card readers were introduced, the cover shape changed to a large rectangular bulge.
|Above: This is the card reader version of the KW-7. . Both examples are on display at the Communications and Electronics Museum in Kingston, Ontario. (Photos by Jerry Proc)|
|Plugblock version. The KW-7 was not a totally
synchronous unit, therefore, it required a phasing signal to be sent in
order to attain a lock on the received signal. top. The circuitry
was all solid-state and fit in a fairly rugged housing that featured a
Tempest-sealed lid on top, removable for servicing. (Photo by Jerry
Examples of the plugblock and card reader versions of the KW-7 can be found at the Communications and Electronics Museum in Kingston, Ont.
SETTING THE DAILY KEY
THE WIRE CORD VERSION
The original KW-7 had 30 fixed wires coming out of the bottom half of the key area. Using a daily key list, the operator plugged the wires into the top half a flat block. Many operators joked that part of the KW-7 course consisted of basket weaving since that seemed to be a prerequisite skill for this task. The task of coding the plugblocks was sometimes referred to as knitting. The wires were about 14 to 18 inches long and it took forever to get them all plugged in and still be able to close the door over top of them. There were metal loops provided to help shorten the slack but it was still difficult to get them packed in.
Many a time, a KW-7 used in flying command posts for TTY traffic would be pulled in for maintenance only to discover there were little black loops of wire hanging outside of the edge of the locked door because it was such a tight fit inside. This was frowned on since the purpose of the gasket was to reduce Tempest.
PLUGBOARD VERSION (?)
A former KW-7 operator, indicates that there was a variant of the plugboard circa 1969. It was likely a predecessor to the plugblock version and does not appear to be commonly known. He recalls " the black cables were entirely separate from the box both ends. The bottom end of one cable is shown in the illustration below. The numbers were aligned perpendicular to the cable so that when properly plugged in, they were easy to read. As with all of the plugs, it was a "bear" to get all of them in place and still be able to lock the door without pinching the cables". Any collaboration on this cabling scheme would be appreciated. Contact: Jerry.Proc@sympatico.ca . Did this system have any official name?
He also indicates that his KW-7 came with a special accessory. "The only other “attachment” I had was a thermite grenade, which sat atop the box the whole time I was with it. I’m pretty certain that this fitting was Standard Operating Procedure after the USS Pueblo was captured by the North Koreans in 1968.
|This image depicts the connector type used at one end of the patch cable. The other end would have been identical.|
THE PLUGBLOCK VERSION
Plugblocks had to be set up by hand and were not considered to be a fun job. However, the introduction of the plugblock offered two important features:
First, it reduced the damage to the cords. Second and more importantly, it allowed the key to be set up in advance on a spare block thus allowing the key to be changed in moments instead of the 10 to 15 minutes it took to manually patch and route the cables in the original variant. This allowed the operator to have the blocks "pre-corded" before the start of the next crypto cycle (HJ). A locking lever at the right side of the plugblock was lifted and swung to the left. This permitted the entire plugboard block to be removed and another one inserted.
KWK-7 plugblock assembly. Once wired up, this mated with the plugboard in the machine and was locked in place. The three silver coloured bars were used as retainers for the plug wire connectors. (Photo by Jerry Proc) Contact side of plugblock assembly. (Photo by Jerry Proc) This image illustrates the markings on the plugboatd. (E-bay photo)
In the photo titled "A close-up view of the mating plugblockmodule", the little white handled tool (clipped to the left of the plugblock) was used to push the little white square jacks out of the block. Flat face and block modified KW-7’s were totally compatible with each other and used the same key lists.
A KW-7 with fitted with a card reader could not communicate with a machine fitted with a plugblock because there were many more permutations available with the card reader than with the plugblock or the 30 cables.
|A close-up view of the mating plugblock module. The locking handle is easily seen. (Photo courtesy of the Communications and Electronics Museum, Kingston, Ontario).|
For maintenance, there was a back-to-back reader that would accept two plug blocks and it would test to see if both blocks were sequenced identically. If not, an alarm would sound and the tech had proceeded to search for a broken wire. Crypto techs considered this to be a real pain when troubleshooting. There were clumps of cord plugged all over the board and each cord would have to be wiggled until the intermittent short was located. Each cord had two conductors with a numbered pin at the end. If one broke, you could not just get a spare cord, you had to get one with the correct number at the end of it. To further add to frustration, a punch card machine could not communicate with a plug board machine!
|This KW-7 test bed was made by SUB NDA West, Denmark around 1987-88 in order to test the KWK plug board. (Photo by Henrik Teller )|
Steve Gardner, who formerly maintained the KW-7 offers these comments. "You seldom got the plugblock correctly routed the first time. It would work - BUT you couldn't get the cover closed! Rerouting some of the patch cords would cure the problem, but it required a new BI when you did it. The KW-7 does not provide "Traffic Flow Security." If it is not sending a message it spits out a "one mark" Baudot pattern, the same as BI pattern".
A KW-7 maintainer in Europe relates his maintenance experiences with the machine. "We had bi-stable relays in the KW-7 which switched the loop current between the mark and space state. Rarely did we have problems with the KW-7, but if I had to repair one, it was most likely one of those two relays that had a dirty contact and produced gash on the Telex line. We opened the top cover of the KW-7, pulled them out and unscrewed the aluminum cylindrical cover. They were then attached to a small tester which had a little 5 cm cathode ray tube. It was solely designed for testing the KW-7 relays. When a test signal was applied, lots of spikes and noise were observed on bad relays instead of nice rectangular blocks on the CRT. Next, we took a piece of non-perforated, 5 level paper tape, sprayed it with contact spray and moved it up and down between the contacts. Bit by bit, the rectangles on the CRT had fewer spikes. The process was sustained until the relay was free from noise. That KW-7 would then work (around 16 hours per day) perfectly for another 3 or 4 months. I can't remember actually replacing a relay. We always used all possible tricks to get those things running again".
THE CARD READER VERSION
Introduced in 1977-1979, the daily key was changed by installing a fresh key card - a vast improvement over the plugboard method. The card reader modified KW-7 was not only easier to operate in regards to key changes but it was slightly more secure as a result of the additional permutations it provided.
|The card reader door in the open position.|
|Above and below: Close up views of the card reader contact assembly. Apparently, the contacts on the door (bottom photo) are "conductive rubber".|
|All photos in this table by Jerry Proc|
|Rear view: Unfortunately the power supply is missing. It connects to the KW-7 via a 'D' shell connector.|
|Rear view showing the placement of the clock module and two E-AJJ boards.|
|A closer view of the E-AJJ modules which employed "flyball" technology.|
|Close up view of the clock module compartment. Note the switch at the left top which could select 20 or 60 milliamps (input) current loops. The step/cont switch is used during maintenance.|
|Top view looking towards rear.|
When new, these machines cost around $4,500 a piece and required 6 to 8 weeks of training in order to maintain them. KW-7's were also used in U.S. military radio shelters like the TRC-136, the GRC-122 and the GRC-142. The GRC-142 or GRC-122 used one or two KW-7s in a simplex or full-duplex HF radio link In either case they would have been attached to a GRC-106 radio set. Early shelter versions use Kleinschmidt teletypes and later versions use UGC-74 teleprinters. The KW-7 was self-synchronizing to the extent that whenever the other fellow stops sending, you push a button which sends out an initialization sequence to the other end to put it in sync.
On close inspection, the red and yellow pushbutton switches on the base of the machine are surrounded by large protective hex nuts. That was a modification that was added shortly after the machines were placed in service aboard the various airborne command post aircraft (EC-135 variants). Seems that the equipment was mounted close to the floor and it was easier for an operator to press those buttons with the toe of his/her boot than it was to bend over and press them with a fingertip. This repeated action resulted in the mashing of the button faces and that would cause the indicator lamp to short out which put the machine out of commission. To prevent the problem, a modification was issued to secure a collar to both switches so that the boot toe could no longer mash it.
Jim Dell, who served in the Royal New Zealand says that "the BID/660/1 was used in lieu of the KW-7 in our major units. The advantage of the BID/660/1 was that you could also employ Traffic FlowSecurity (TFS) the same as the KW-37T. This meant that you didn't have to rephase to resync and also you didn't need to keep a rubber band on the Transmit button as you did with the KW-7 between traffic! TFS was only ever used in overt operations".
Jerry Kemp used to maintain the KW-7. He provides some insight on the use of the SEND button. "Aircraft personnel apparently had a hell of a time keeping KW-7's synced. So much so, that they would take the cap from a Bic ball point pen, and jam it into the crevice of the SEND button, keeping the unit(s) syncing all the time it wasn't in actual use. This activity caused all kinds of havoc on the mechanical relays, located on the bottom side of the chassis. I can honestly state that there wasn't a KW-7 that ever came in for maintenance and/or PMI, that I didn't spend more time than anything getting those mechanical relays back in operation again.
The security lock on the permuter cover is unusual. Most common locks have one row of pins, corresponding to one row of bumps/serrations on one side of the key.
This photo illustrates the unusual keys used on the KW-7 lock. (Keys donated by Tom Chirhart, Stafford, VA. Photo by Jerry Proc)
The KW-7 lock has three rows of pins, two rows on one side of the key and one row on the other side of the key. The key is next to impossible to duplicate by the average locksmith because cutting that third row of bumps/serrations can't be done on a regular key cutting machine. By 1988, KW7's were being replaced by KG-84's.
Keying method: Cables, keylist.
Service life: 1960's - 1990's. It is confirmed that the KW-7 was still in use as late as February 1993 at HQ PACAF
Output: On-line encrypted teletype.
Speed: 50+ wpm.
Weight: 77 pounds with the permuter block and 82 pounds with the card reader.
Applications: Strategic and tactical environments.
Note: Typically used to send encrypted teletype from ship-to-shore.
Dick Morris comments on the weight. "The weight of the KW-7 was significant to those of us who had to carry them into the EC-135 command post aircraft through the crew hatch. The crew hatch door was at the left side of the aircraft under the cockpit. It hinged downward and had a ladder with about four steps. The way to get into the plane with a KW-7 was to hang onto one handle of the KW-7 with one hand and grab a rung of the ladder with the other.
When you were standing on the first step of the ladder, you leaned against the back wall of the crew hatch which let you release your hand from the ladder and grab the next rung. When your shoulders were a bit above the floor of the plane, you leaned against the wall behind you and lifted the KW-7 onto the floor in front of you. Not a task for the faint of heart".
There were at least two known variants of the clock module whose designator is E-AJY. It produced a precise 1 MHz signal.
|KW-7 clock module. These failed quite frequently. The internal temperature was monitored by a thermostat consisting of a mercury thermometer with contacts. It was calibrated to shut off the oven at around 65 degrees C (?). Sometimes the thermostat would malfunction, thus overheating and melting the Styrofoam covering the 1 MHz crystal. Some identification has been removed. (All photos in this table by Jerry Myers)|
|Detail of E1/E2 power input connection.||When the screw is removed, it allows access to a glass piston trimmer inside thus allowing the module to be calibrated to the precise frequency.|
|The leftmost compartment contains input power and output RF circuit board.|
|Above and below: Two views of the oven assembly which resides in the middle compartment.|
|KW-7 clock module internal views.|
|The clock module had at least one known variant which is shown here. What appears as a vacuum tube is actually a crystal in a glass enclosure fitted with a 7 pin tube-style plug. Nylon lacing cord is looped around the crystal to assist in its removal. This oven had a short in the mercury thermostat thus melting some of the styrofoam at the bottom of the compartment.|
|The crystal plugged into a 7 pin tube socket mounted at the bottom of the oven compartment.|
|All photos in this this table by Jerry Myers|
Jerry Myers explains the mercury thermostat version of the clock module. "The thermostat was very short and had a bulb just like any standard mercury thermometer except there were two wires imbedded into the glass side at two places. These wires would make contact with the mercury column when it rose due to the heating of the oven. The mercury would make contact with the first wire then, after additional heating, it would make contact with the second wire. This shorted the connection to a transistor and killed the power to the oven. After prolonged cycles, the mercury would sometimes fail to make contact with one of the wires thus allowing the oven to overheat until the thermometer broke due to the excessive pressure. At this point you the technician would not be able to adjust the module for a 1 MHz output".
The KW-7 used circuit boards that employed potted modules. One of the transistor types found in the potted modules was the 2N404. Introduced by RCA in 1957, it was a Germanium PNP alloy transistor in a TO-5 case. Soon adopted as an “industry standard” and sold in the millions. Multiple companies started to produce equivalent devices (ie Raytheon, Sylvania, Tungsol, TI, GE, General Instruments, and others).
|Photo courtesy Transistor Museum|
The KW-7 contained an integral power supply capable of running on 120 VAC 60 Hz or 24 VDC. It was mounted on the chassis of the KW-7, but shrouded in its own enclosure and removable from the KW-7 without disturbing anything else. Former technicians who maintained the KW-7 speak highly of the internal power supply reliability since they never had to fix one.
|In Europe, the KW-7 was operated with an external 220 to 110 VAC step down transformer. It's dimensions are approximately 15 x 15 x 15cm (Photo by Henrik Teller )|
In Belgium, the Comm shelter was fitted with both 24 VDC and 220 VAC. The 220v was provided by a small generator on a trailer A power supply connected to the 220v mains would charge the batteries. Normally the equipment always ran on 220v, but when the generator went down, an inverter transformed the 24v from the batteries into 220v for the equipment, the KW-7 power adapter and some other gear. Although a few things ran directly on the 24 VDC mains, the KW-7 was always used on 220v. (Our teleprinters needed 220v, so we had to have 220 anyway).
The KW-7 provided either a 20 or 60 ma loop current.
KAM-143B/TSEC Description, Installation, & Operation of TSEC/KW-7 (a 1967 manual) KAM-144B/TSEC Preventive Maintenance & Troubleshooting of TSEC/KW-7
KAM-145C/TSEC Operating Instructions for TSEC/KW-7
REMOTE CONTROL PANEL
Photo by <firstname.lastname@example.org> <Photo by email@example.com>
This is the cockpit remote control unit which could operate the KW-7. It was made by Collins. The Orestes system, used in strategic and tactical environments, was not limited to ships and land based systems but also used in aircraft. The panel has 16 Clare Pendar switches: they are a combination of DPST, 4PST, latching, momentary and all are illuminated. Face plate is backlit as evidenced in the right photo. Switches have the following labels, TTY / KEY, KW-7, MODEM TEST, POWER, KW-7, 700B-2 1, KW-7 2, 700B-2 2, 700B-4, TE-204, KW-7 1, 700B-2 1, 700B-4, TE-204, KW-7 2 .
Dick Morris adds. "The Collins remote unit depicted above was actually part of the USAF ARA-60 system along with the TE-204A-2 synchronous modem. This system was on the airborne command post aircraft. The code name for the planes that I worked on at RAF Mildenhall in the U.K. was Silk Purse. Similar planes at Offut AFB were code named Looking Glass."The KWX-7 was another type of remote control for the KW-7.
KWX-7As a result of the capture of the USS Pueblo in 1968 by the North Koreans, the following cryptographic systems were compromised: KL-47, KW-7, KWR-37 and KG-14. An NSA report dated July 1969, details the cryptographic losses. A copy of this report can be found here.
KWX-7 Front View. (Photo via E-bay) KWX-7 Rear View. (Photo via E-bay)
KW-7 IN SERVICE
This is a photo of an New Zealand Army FFR (Fitted For Radio) vehicle showing the Land Rover backed into an 11 foot by 11 foot tent. On the ground, immediately below the tailgate, are two carrying cases for the KW-7. The KW-7 is sitting on the table with the dust cover in place. To the left of the KW-7 is a Kleinschmidt teleprinter. This FFR belonged to 2 Sig Sqn. (Photo courtesy Ian Fraser, RNZ Signals Ret'd. E-mail: <ifra1616(at)bigpond.net.au>)
A Landrover 109 SIII was used as a teletype van by 2 (DA) Signal Battalion of the Danish Army. Above the KW-7 stack and over to the left are two Siemens T-56 line converters. To the left of the operator is one of two Siemens T100Z teletypes. (Photo courtesy Royal Signals Historical Collection, Fredericia, Denmark) This teletype van was used by 2 (DA) Signal Battalion of the Danish Army. It was a Danish produced shelter (known as the M/83) and mounted on a Magirus-Deutz 168M11FAL chassis.
At the left, and only partially visible behind the blue chair was a Siemens T100- STV20 teletype and a Siemens M190 crypto equipment which used keytape. It was only used for preparing messages (poking). Behind that,
are four operator positions each with a Lorenz LO-133 teletype and a KW-7. To the right (and not visible) were two positions with Lorenz LO-133 teletypes and KW-7. (Photo courtesy Royal Signals Historical Collection, Fredericia, Denmark)
For additional information on the vehicles mentioned in the photo captions, please visit the Danish Army Vehicles Homepage of Henrik Teller.
When the U.S. Embassy in Tehran was stormed in 1979, the crypto equipment or key pieces of it were destroyed by the communicators. That's a KW-7 in the lower left portion of the picture. The front panel has been removed along with the critical key generator boards. It would appear that the personnel that ran the commo site knew what to do when overrun. (Photo courtesy Memory Hole web site).
For additional information on the KW-7. please visit the Cryptomuseum web page. http://www.cryptomuseum.com/crypto/usa/kw7/index.htm
Credits and References:
1) Lee K. Shuster - Salt Lake City Utah.
2) David Ross, N7EPI
3) Tom Chirhart - Stafford, Virginia
4) Steve Gardner. E-mail: steveg(at)dfa-inc.com
5) Jim Dell, ZL4JAD <dinger69(at)xtra.co.nz>
6) Memory Hole web site: http://www.thememoryhole.org/espionage_den/embassy1.html
7) Henrik Teller <henrik_teller(at)vip.cybercity.dk>
8) Gregory W. Moore <gwmoore(at)moorefelines.com>
9) Dick Morris <rmorris(at)alaska.net>
10) Jerry Myers <jerry1joan(at)aol.com>
11) 2N404 photo courtesy http://semiconductormuseum.com/MuseumStore/MuseumStore_2N404_Index.htm
12) John Dill < jdill1927(at)bellsouth.net>
13) James Gabrisch <cpgarbanzo(at)sbcglobal.net>
14) Communications and Electronics Museum, Kingston Ontario.
15) KW-7 manual info William J. Neill <wjneill(at)consolidated.net>
16) Jerry Kemp <cryptosys(at)oryx.us>