RFP (Radio Finger Printing) and TINA

During WWII, the most common method of locating enemy radio transmissions was through the use of radio direction finding whereby two or more monitoring stations would obtain bearings on enemy transmitters then triangulate the results to obtain a fix.  There were at least two other techniques which could provide data about the characteristics of the transmitter making the emission and the Morse code sending style  of the operator, These techniques were known as RFP and TINA and are discussed in more detail below, When used jointly, these techniques were known as 'Z' intelligence. As with many other WWII activities, RFP and TINA were shrouded in a cloak of secrecy.

RFP

Radio Finger Printing (RFP) was the process used to catalogue a specific transmitter through its distinct characteristics with the aim of locating it at a future date.  The idea was to identify individual transmitters by their emitted waveform. The rectified signals were applied to a cathode ray tube, (without a time base), and photographed onto moving film.

The German naval transmitters were remarkably uniform and showed little more than an initial damped oscillation produced, perhaps, by a carbon-pile voltage regulator. The Italian mobile naval transmitters were remarkably erratic, and never looked the same twice; while their shore stations could often be identified by the ripple frequency of anything but 50 Hz.

TINA in the UK

TINA was the method used to recognize specific radio operators by their Morse code "fist" and habits.  The code name of TINA was derived from the Latin word "tinea" which meant "worm".  Extracts from a TINA article written by John Roscoe G4QK provide a more detailed picture of TINA.

"The early TINA recordings of Morse were made with a siphon-pen recording on paper tape .For copying high speed Morse,  a device known as an undulator was also used.  At a later date, Morse transmissions were recorded on 35 mm film with a slow running RFP machine. This allowed measurements to be made through considerable QRN.  This permitted the desired waveform to be distinguishable from the ambient noise.
 

rfp_tina_undulator.jpg
This is an example of the undulator. It is not known if this exact model was used in TINA operations but it serves as a good example. (Courtesy of High Speed Recording of Radiotelegraph Signals)
In the UK, the standard method of measuring Morse at that time involved several WRENS. Two WRENS at a time would start the process. One would read off the measurements from the film using an enlarger and the other transferred them to graph paper. The dashes and dots were then marked with red and blue circles. Attempts at identification with previously captured records were made by direct comparison.  This procedure inevitably relied on the acuteness of observation  and perhaps the memory of the operator on duty. This was not a fast method so two ideas were pursued to help automate the process.

(1) A method of recording that would permit simple and more rapid measurement;
(2) A machine-compatible system for classifying the records.

To summarize:

On the issue of measurement, the TINA device would make a vertical deflection on  the 35 mm film instead of instead of a horizontal one.  The vertical deflection would correspond to each dot or dash in the Morse character.

The idea behind the mechanical classification scheme was to extract parameters from the Morse that could be coded  numerically and searched mechanically. In those pre­computer days, the best machine that was available was the Hollerith 80-column punched card sorter with a built-in 8-column group selector which worked at the astonishing rate of 400 cards per minute!  However, there were issues with the generation of Morse code before a solution could be implemented.

In the WWII era, Morse was sent mostly by straight key.  Semi automatic keys did exist in 1942 but they sent relatively clean code so it would be challenging to make an identification to a specific operator using  TINA.  So TINA was limited to the analysis of Morse generated by straight keys.  In addition, sidetone oscillators were uncommon, so practically all the operators wouId have been sending by "feel" alone.

rrp/rfp_tina_undulator_output.jpg
At the top is the output tape of the undulator with the equivalent Morse characters shown below . From this hard copy the TINA operator could study the length of the dots and dashes in the character stream. (Courtesy of High Speed Recording of Radiotelegraph Signals)
The most consistent characteristic in Morse code was the ratio between an adjacent dot and dash in the longer symbols. For example, the dash might be markedly longer than the following dot in D, B, and 6. This would almost certainly be matched by a similar disparity in U, V, and 4. Invariably, A and N did not fit into this pattern. The fate of the other dots, in 4 and 6 for example, 'was a matter of individual taste. This simplified measurements, as there was no need to establish a notional average length for the dot. The ratio between adjacent elements could be immediately coded on a scale of 1 to 10, with adequate provision for the spread of results.

The most interesting application of this technique was to the U-boats in the North Atlantic. Although they carried more than one radio operator, only one appeared to do the transmitting, thus reducing the number of our records that had to be maintained by TINA personnel.  Mass training for the German U-boat radio operators  was so standardized that their style of sending made TINA analysis impractical.

The U-boats were at sea, generally, for a maximum of 6 weeks, so a preliminary scrutiny could be made,  and had to be made, in a few minutes of all records obtained within this period. More leisurely checks could then be made for long term files. Of course this was only part of the picture, as direction-finding, RFP, decoding. etc., would also have contributed their intelligence to the records base. Of course,  the "victims" were well aware that they were being monitored but since the U-boats invariably sent short messages (quite unlike the Italian submarines in the South Atlantic). they presumably felt the risk was small ".

TINA and RFP in New Zealand

In the early 1940s, at least seven New Zealand signals intelligence stations were constructed and operated as part of the British-American
intelligence system. Threeof these stations were set up in Awarua (ZLB) ,  Auckland (ZLF) and Waipapakauri (WPP)  which became part of the US  Navy's Pacific-wide DF net.  The purpose of the net was to locate Japanese ships and other sources of enemy radio transmissions.

In 1942, a new station  (RNW) joined the net. RNW, was the Post Office call-sign for Renwicktown, near Blenheim NZ. In the direction finding stations,  little was told about RNW's function except that it was a naval W/T station and that all personnel were to fully co-operate with it. The new station would be staffed by the Womens Royal New Zealand Naval Service (WRENS). These WRENS were all competent operators, and  fluent in the Japanese Katakana code. Many years after the war, it was revealed that RNW was set up as a Radio Finger Printing station.

The job of the New Zealand D/F net, was to take radio bearings on enemy signals that were radioed from the US Pacific Fleet’s Headquarters (NIT) in Hawaii, and later Guam. RNW tracked around with these frequencies and call-signs received from NIT.

At RNW, the identification process was started by photographing signals which were displayed on an oscilloscope and examining the developed print in minute detail for any peculiarities in the received waveform. These would then be compared with previously recorded photographic strips to see if there were any similarities which would enable the identity of the station or vessel.

The RNW station itself was a two-storey farmhouse taken over for the duration of the war by the  New Zealand Ministry of Defence. It was at the end of a long no-exit road at Rapaura, terminating on the banks of the Wairau River.

The house was altered to allow an operations room downstairs but the radio equipment and sleeping quarters for the eight Wrens who would be staffing the station were on the second floor. The area surrounding the house was enclosed with a six-foot high, barbed wire fence, with a locked gate, and security was provided by a detachment from the Guards-Vital Points which camped on the site.

It was an ideal radio receiving situation, away from all interference. There was a clump of tall poplar trees close to the house and these were used to provide masts for the inconspicuous aerial, which was erected by using a bow and arrow! The equipment was based around a special receiver incorporating an oscilloscope and movie camera, known as REB 2, which had been sent out by the  British Admiralty from London. Also provided were two Collier and Beale HRO type receivers, fitted with oscilloscopes.

This equipment was installed by an ex-Post Office telegraphist who was a Leading Telegraphist in the RNZVR. He remained at the station to instruct the staff in the Japanese Katakana code and left when they were proficient.

The operating procedure was to tune into the frequency and identify the callsign which had been signalled from Awarua on the landline. The signals were checked on the oscilloscope attached to the standard receiver for photographic suitability. If suitable, a button was pressed which brought into operation the REB 2 equipment which then started photographing the waveform of the oscilloscope.

Later, when the film was developed, the Classifier examined the films in detail looking for imperfections such as harmonics, damped waved, key-relay effects and anything else peculiar to that transmitter. In many cases the operator’s style of sending could be recognized and this would be a further clue to the station’s identity.

The photographic strip details were recorded, given a file classification and stored for future reference. Once a file had been built up the station could be identified by the signal it emitted and the style of sending. This meant that no matter how often the callsigns were changed – and the Japanese changed theirs daily for a while – the station or vessel could be identified irrespective of call-sign.

The operators worked during the night, 6 pm to 6 am, on four-hour watches, and the classifiers during the day, examining the photographic strips that had been received overnight. If there was incoming traffic of importance during the night or if it was busy, the operators would thump on the ceiling with a broom handle to wake up the sleeping classifiers to bring them down to attend to the urgent traffic and clear the filled cassettes. The results of the classifiers’ analysis would be suitably coded up and telephoned to the Navy Office in Wellington over a Scrambler (inverted speech) telephone.

There were other Radio Finger Printing stations set up by the Admiralty, The one in the UK was responsible for identifying the German pocket battleship Bismark when it broke out into the North Sea and was ultimately sunk by the Royal Navy. There was another RFP station at Sri Lanka (formerly Ceylon) and possibly other locations.

The Japanese advance southwards was blocked by the Battle of the Coral Sea in May 1942, and their offensive power effectively broken four weeks later at the Battle of Midway. Japan was on the defensive and their large, ocean-going submarines were no longer able to roam at will. A considerable number were used to re-supply their isolated garrisons under Allied blockade in the Pacific.

Enemy radio activity steadily dwindled and in May 1944 the Rapaura Naval W/T station (RNW)  was closed, with the staff being transferred to the New Zealand Navy Office in Wellington, to work in the Intelligence and Communications sections.

During its operation, the Naval W/T station Rapaura (RNW) provided a valuable insight into the characteristics of enemy signals from Japanese forces that were menacing the islands of the Pacific. It is to the credit of the Wrens that this particular specialized service was performed with discipline, diligence and in total secrecy.  All the personnel involved at RNW were held under an Oath of Silence from the NZ  Ministry of Defence until 1982.

(The above are extracts from an article titled " NZ Navy Wrens In Secret Ops During WW2" written by Frank Barlow ZL2NB with contributing WRENS Bunty Longuet and Philippa Corkill, both Leading WRENS at the Rapaura Station.  The article was originally published in Break-In, December 1996.)

TINA and RFP in Canada

TINA/RFP equipment was given to the RCN by the British Admiralty in December 1941 and by January 1942, operations on an experimental basis had begun at a Department of Transport station near Ottawa. In May 1943, "Z" operations moved to the RCN station at Gloucester Ontario but due to the drop in U-boat W/T traffic, the RCN chose to install RFP/TINA at Harbour Grace, Newfoundland.

During a visit to England by Lt. Low(RCN) , he observed that all "Z" operations and classification work was performed by British WRENS who were most capable for this job.  The RCN based their decision on this observation and decided that Canadian "Z" operations should be operated by WRCNS personnel, wherever this could be done considering the locations and amenities of the isolated Canadian stations. Two RN WRENS were requested and ultimately sent to Canada to train Canadian WREN personnel for these duties. They arrived November 1943 and spent three months between NSHQ and Gloucester and Harbour Grace stations.

REB and REC EQUIPMENT

The first photographic recording equipment was designated REB and initially saw service in 1938. An improved model, the REB 2 was released in 1940. ANother model, the REB 3 was specifically used for noise investigation.

There are no known photographs of REB and REC TINA/RFP equipment however Clive Kidd of the Collingwood Heritage collection provides some elementary descriptions.

"The B40 receiver (a 1950's design), also known as outfit CDW (from handbook BR222)  has  two sockets -  SK202 and SK203.  These are for REB and REC equipment . REC ( SK202) is the output from the detector and is an audio output  (described as DC) . It was designed to record audio to a disc, i.e a shelac disc as per the old 78 records on a piece of kit known as REC. No further info was found on this equipment.  It could also be connected to an audio tape recorder, wire or tape.

SK 203 is connected to REB and is the IF output  for "photographic analysis of transmitter characteristics using equipment with the designation REB. Manual BR 1433 dated 1945 describes REB as being used to "...investigate the faults and characteristics of any radio transmitter..." The signals were fed from any  receiver (with an IF frequency between "...100 kc/s and 1 Mc/s..." ) into the apparatus and examined visually and also the signal could be photographed. The apparatus had two CRTs - one for the operator to sight on and the other covered by a camera. The Y axis of the CRT was fed with the signal from the receiver, via an adjustable gain amplifier and the X axis of the operators CRT was fed with a time base - as per a normal oscilloscope. The only difference between REB and a scope was REB had two CRTS. The camera could be set to drive the film past the second CRT at "...1.5/3.8/9.6/24/60/150 cm [ yes cm in 1945 in the Royal Navy per sec....] . Time markers could be produced internally to aid analysis.

The BR on REB does not contain any photos of the kit, only some line drawings of the chassis and a drawing of an amplitude modulated carrier signal, in the time domain, to show what a signal might look like - standard picture as per any text book. There is no description in the BR to aid in the determination of faults or special characteristics. Consensus of the volunteers is that the latter might be used to identify individual transmitters and thus fingerprint the sending station, potentially useful for tracking a ship by its radio transmissions.

I have also found out that there were a number of different marks of REB equipment -  mainly changes to recording speeds and the way the CRTs were fitted into the case along with other mods to the operator interface".

RANGE ESTIMATION EQUIPMENT (REA)

This was developed during the war as an ancillary to H/F D/F fixing. It consisted of taking measurements of the path differences between various incoming rays appertaining to the same signal, and, by correlating these differences with current ionospheric data, estimating the distance of the transmitter.

In 1944, after a great deal of experimental work, a method of analyzing results was evolved which enabled some 80 per cent of experimental intercepts to be assessed for range with a 10 per cent accuracy. The scheme never became a practical success as it was complicated and suffered from many difficulties.



Contributors and Credits:

1) "A Few Measurements" by John Roscoe,  G4QK. Published in Morsum Magnifact, Autumn 1987 issue
2)  Aryeh. Ben-Ami  <dufs44(at)bezeqint.net>
3) Frank Barlow ZL2NB Break-In, publication, December 1996.)
4) Clive Kidd <cjckidd(at)waitrose.com>  Collingwood Collection
5) http://www.disarmsecure.org/The%20Origins%20of%20Signals%20Intelligence%20in%20New%20Zealand.pdf
6) http://www.rcsigs.ca/index.php/History_of_Canadian_CESM
7) High Speed Recording of Radiotelegraph Signals by  R. B. Armstrong B.Sc.  and J. A. Smale, B.Sc,
8) Radio Warefare document 1949. Ref: SD 1080/47
 
 

Back to Table of Contents

Aug 21/17