Ray White, Chief Petty Officer 2nd Class RCN (Retired)

Dedicated to my messmates throughout the SUPRAD System

The Canada/United States Direction Finding Networks, both Atlantic and Pacific, performed their assigned functions under cover of great security and provided an indispensable adjunct to the National Defence of both countries during the cold war of the 1950s and 1960s. The work of these networks was done by Communications Technicians (CTs) in the USN and Radiomen Special (RSs) in the RCN.

Operation of the Atlantic HF/DF Net, in which HMCS Coverdale was an important player as Alternate Net Control and as a participating DF Station, involved cryptographic cover in real time. The following are the steps taken to inform the net of a target, the frequency being used by the target, identify the target by call sign if possible, and provide other details to assist the HF/DF stations to obtain bearings.

Net Control in Cheltenham, Maryland, transmitted on “flash” frequencies in the 4, 8, 12 and 16 MHz bands. The basic transmission was a series of the letter F at about 12 wpm in Morse code. On those occasions when Coverdale assumed the duties of Net Control, all of the following would apply except the flash identifier would be a series of the letter M.

When it was desired to transmit a flash, the letter F would pause for a few seconds and then a long break BT would be sent in manual morse followed by the letter F with the next line in the one-time crypto pad, eg, FC FC FC followed by a five digit group repeated several times. This group, when decoded, provided the frequency to which the DF stations would tune. A few seconds later the identifying characteristics of the target would be transmitted in encrypted form, and operators would be able to distinguish between stations transmitting on or near the target frequency.

Let's say the target is a station sending on 8367 KHz, and its call sign is GBSS, the British ocean liner RMS Queen Elizabeth calling WCC, the US Coastal Station at Cape Cod. (In actual fact the net would quite often use moving targets from friendly nations, especially ships who had a reputation for accuracy in reporting their position. In addition to providing a means to exercise operation of the net, it also would provide a live “check-bearing” function.)

The one time pad cryptosystem was provided by the United States Armed Forces Security Agency and was in the AFSAV series. It was a highly classified cryptographic publication with a unique registration number. The system used on the HF/DF nets was called the JOVE Cryptosystem and consisted of lines of five-digit groups identified alphabetically from A to Z. In our example, Line C of the JOVE one-time pad would resemble the following:

C 87319 70143 82290 54036 77290 52317

Net Control would write in the plaintext for the target. Frequency in Kilohertz was always in five digits, Call signs used two digits for the numbers 1 to 0 to 9 (01 to 10) and the letters A to Z (11 to 36). Amplifying information would follow. For purpose of this example we will use the hypothetical two digit group 42 to signify “calling”. The Flash would then look like this before encryption:

08367 17122 92942 33131 39999

8367 KHz – Target frequency.
17 for the letter G, 12 for the letter B, 29 for each of the two letter S, 42 to indicate that GBSS was calling
33 for the letter W, 13 for each of the two letter C.
The block of four figure 9’s was known as padding to ensure the final group consisted of five digits -- a group of less than five digits was forbidden for security reasons.

To encrypt the flash, a process known as “blind subtraction” would be employed: the subtraction was done from left to right without carrying over. Thus if a seven is under a zero, the zero is considered to be ten and the subtracted value is 3. If a digit is under a number the same or larger than itself, then the true subtracted value is used:

The encrypted flash would then be:

89052 63021 90358 21905 48301, followed by the flash identifier, say, F105, numbered one-up daily.

The first group, the encrypted frequency, is transmitted and then, after several seconds, the identifying information. Then Net Control would continue transmitting F F F etc., until the next flash, F106, in which case line D of the crypto pad would be used.

Each outstation in the net was equipped with two communication receivers in addition to the CRT DF set. In the case of the RCN stations,  the receivers were usually SP600s and the DF set was the CNF4. One of the SP600s would be tuned to the Flash broadcast frequency (this would vary through the day depending on the ionosphere and the state of HF propagation). The second receiver would be tuned to the reporting frequency, which would be in the 5, 9, or 13 MHz bands. (The author apologizes for lapses in memory here…many years have passed since actual exposure to the net.)

On hearing the pause in the F F F stream, outstation operators would prepare to copy the first group and subtract it from the printed key group on line C. In actual practice, operators quickly became adept at doing the subtraction in their head and writing only the result on the one-time pad.  The CNF4 would then be tuned rapidly to the frequency, and the operator would prepare to decode the remainder of the flash message. It became an automatic procedure and could be done as a reflexive act. An average day would often have more than 200 flashes.

After a suitable interval, usually about five or ten minutes, Net Control would call the net on the reporting frequency and ask for bearings:


Stations would reply in predetermined, usually alphabetical, order, but only after being directed individually to report. Circuit discipline was extremely strict and deviations from established procedures were not tolerated.

The encrypted report consisted of a four-letter group, the first three letters of which indicated the numerical bearing from 000 to 359 degrees, and the final letter indicating the reliability or spread of the bearing. A would indicate a spread of +/- 2 degrees, B would indicate +/- 5 degrees, C would indicate +/- 10 and D would indicate a very unstable bearing. There was also provision for Q, a quick bearing, one which was not fully identified or classified.  There were groups for NH, nil heard, OS for Out of Service etc.

Thus, Coverdale’s report would be coded from that day’s page in the bearing report book and transmitted as follows. On receipt of the CKT K instruction, Coverdale would reply:


Other stations in the net would report when called. If a station was slow in reporting, Net Control would move on to the next in line and pick up the stragglers at the end. On receipt of all the bearing reports, both Net Control and Alternate Net Control (Coverdale), who maintained landline communication with each other, would plot the bearings to try to determine a “fix”, the spot where, in theory all the bearings would cross. A rough fix would be made by using weighted string markers centred on compass roses at the location of each HF/DF station.  If an important target was being processed, the bearings would be plotted on a chart to determine a precise position, if possible.  These fix charts, or maps, were drawn on gnomonic projections of the area of the target and could be quite accurate. On a gnomonic projection, all great circles, i.e., the direct path taken by radio waves, can be drawn as a straight line.  This is quite different from the Mercator chart used in marine navigation, where great circles are plotted as a curved line but rhumb lines are plotted as straight lines.

Computers were also used to plot the fixes obtained by the net, especially in the late 1950s and early 1960s. The manual drawing of fixes was limited to those instances where a human interpretation was required to help in the automated processing of the information.

Another indispensable aspect of HF/DF Net operations was its role during emergencies.  Merchant Marine and Aircraft distress calls, both CW and voice, were not uncommon in the 1950s and 60s, and the net would suspend normal operational duties to concentrate on the emergency at hand. It was a matter of pride throughout the system to participate in a successful search and rescue operation, especially as no credit was given or sought for the help provided.

Back to HMCS Coverdale

March 30/04