From msuinfo!agate!ames!sgi!wdl1!nebula!koontzd Fri Aug 6 12:37:34 1993
Newsgroups: sci.crypt
Path: msuinfo!agate!ames!sgi!wdl1!nebula!koontzd
From: koontzd@nebula.lrmsc.loral.com (David Koontz )
Subject: Re: KY-28
Message-ID: <1993Aug3.024829.21493@wdl.loral.com>
Sender: news@wdl.loral.com
Organization: Loral Rolm Computer Systems
References: <CB27K3.Fn2@sybase.com> <1993Aug2.195750.2076@schbbs.mot.com>
Date: Tue, 3 Aug 1993 02:48:29 GMT
Lines: 96

>The KY-28 was a "black box" used for air-air and air-to ground Comsec
>during the Vietnam War. I doubt if they are still in use, and doubt if
>you could get a schematic or system specification.
>
>During the Vietnam war, fighter pilots complained the KY-28 voice
>encryptor took "too long to synchronize". It was "at least a nuisance and
>at worst a hazard." As a result, the equipment was often switched off and
>transmissions were made in the clear.

The KY-28/TSEC was part of a family (KY-8 family) of secure voice
equipment. Crypto (COMSEC) Custodians had lots of them in their vaults in
various locations in the late '70s. They were used in Viet Nam, more
commonly by the Navy. It was keyed through a small coverplate, that
when opened exposed a number of holes. A key charging device was used
that was comprised of slide stops on a series of rods and a cam
mechanisim to force the rods out to the distance set by the slide stops
against spring pressure. The rods would penetrate the holes and set permuter
contacts according to the depth of penetration. A zeroize pin was
activate by closing the cover, such that opening the cover again would
zeroize the permuter settings, which would also render the equipment
incapable of passing an alarm check and passing to operating mode. It took
a lot of pressure to load a key set. This was all in a 1/4 ATR package in
infamous NSA grey. The charging device was set according to a distributed
and classified key list. The KY-28 was classified Confidential, or held the
classification of the keys when loaded. The equipment was owned by individual
services (who were stung by the 100 million plus price - procurement during
the Viet Nam war included retrofitting aircraft inventory. This resulted in
a slow down in procurement for AUTOSEVOCOM II, as the customers (services)
didn't want to pay the cost and congress did not want another lemon.
This eventually lead to STU-IIIs and would make a good 'Connections' episode.
Some of this was covered in AVIATION AND AEROSPACE WEEKY before the advent
of CCEP.

There were several connectors below the charging port, for audio, digital,
power and remote control. There was a green painted version in a different
case for tactical land vehicle use. The voice encoding method was CVSD
and ran at 20 some odd Kbaud (to match the bandwidth of a particular radio).
I don't recall if I ever heard who manufactured KY-28s, old issues of AIR
FORCE TIMEs would be a good place to check. AFT used to list contracts
although crypto stuff fell off the public lists in the late 70s. Harris
used to get the lions share for secure voice.

The initialization delays were caused by two factors: crypto synchronization
delays and digital transmission synchronization delays.

The crypto sync delays could have been alleviated with cheaper digital
storage (to MILSPEC) and/or by using other design principals for coming
up with message indicators (initial vectors), the limiting factor being
digital storage. During the era, fast training MODEMs were a bit
of a rarity, covered I believe by a patent held by Bell Labs. Some of
the delay might have been due to digital synchronization which could be
overcome more readily today. The problem in its use was as alluded,
you had to PTT and wait out synchronization tones. Slow synching equipment
is better suited to full duplex installations not requiring frequent
resync. The implemented message indicators were slow, long and relatively
error immune. Trying to speed up the message indicator would call for
error correction, which was rare and expensive during the era. Once
synchronized an isosynchronous system using CVSD is not particularly
susceptible to transmission errors, which causes only momemtary distortion
as long as no clock event mistakes occur. Speeding up the message indicator
would have required faster hardware and might have been technically
unfeasible or prohibitively expensive.

The KY-28 was/is handled through COMSEC distribution channels, in essence
controlled by the NSA. If you have found access to one or more, the chances
are they came from Viet Nam. Their introduction to service was predicated
on the fact that the North Vietnamese were intercepting plaintext radio
transmissions and exploiting the resulting gathered intelligence, as was
done against the U.S./Allies by the Germans and Japanese in WWII. The KY-8
family has been replaced in U.S. inventories. The above reference to
"black box" is from the fact that field level maintenance for the KY-28 was
restricted to remove-and-replace. Troubleshooting the internals of one
would consist of remove and replacing subassemblies with known good items
according to a troubleshooting chart analysis of symptoms.

All retired crypto gear is destroyed. The Navy used to dump emptied
chassis (fixed plant crypto) off the Farallons, next to all that nuclear
waste (heresay). I used to be an Electronic Communications and Cryptographic
Equipment Systems Technician (30670) with SEIs for AUTOSEVOCOM and AUTODIN
among others. I remember one of the old hands,an instructor in school
describing a KY-1 (retired in the early 60's, replaced by the KY-3). I always
thought there should have be a museum for this stuff. Some of the stuff
used tubes and magnetic logic circuitry and should be preserved for its
uniqueness. You want art deco, look at the color selection chart for KY-3s
installed in quarters. KY-3s were built into safes.

I had a change to buy an M-209 Hagelin mechanical cipher machine at the
flea market in the early 80s, but didn't have the $450. I wouldn't know
what a KY-28 would be worth to a collector (if there were a market).
The government, if aware, would almost be sure to intercede. If you were to
shoot lots of pictures first, they would have some value.

One wonders about the availability of Soviet communications security equipment
since the fall? Perhaps someone could recover gear from the Pueblo or Viet
Nam? I also wonder about knock off copies as was done with the B-29 and
space shuttle.