Long Range Submarine Detection
By 1952 there appeared to be two promising techniques for detecting submarines at long ranges (up to 100 miles) by novel processing of the signals from a long array of hydrophones sited on the seabed. As well as employing novel processing techniques, the postulated systems were equally novel in that they would exploit the (fairly recently observed) low audio frequencies emanating from submarines i.e. they were listening, commonly known as passive, sonars.
Candidate Detection Techniques
The US system employed a narrowband technique using frequency analysers capitalising on the discrete frequencies radiated, whereas the UK system grew from an idea of applying a technique developed by Sir Martin Ryle (a radio astronomer at Cambridge University) for detecting weak radio emissions by correlation of the signals from widely separated antennae, over a wide frequency band.
The US concept development went under the project name of JEZEBEL, which evolved into an airborne sonobuoy system and also the SOSUS – the extensive underwater surveillance system. The UK initiative became known as
CORSAIR, and actually reached the stage of having a RN operational sonar (number) allocation of type 191.
In all events things moved very rapidly. The impression is that senior naval staff viewed a seabed-sited long range detection and location system to be a force-multiplier, relieving naval and air units from patrols, and so allowing them to concentrate on the prosecution of targets; thereby compensating for the lack of (suitable/capable) deployable units.
The US JEZEBEL
system developed quickly with the Bell Telephone
Laboratories, Murray Hill, New Jersey, under contract to the USN via ONR, to
produce the acoustic signal processing portion based on their research of frequency analysers for telephonic work.
The Admiralty Research Laboratory (ARL) assessed the UK CORSAIR technique in 1952 at the (ARL run) Admiralty Experimental Station (AES) at Perranporth on the north Cornish coast. The assessment was highly successful and demonstrated a spectacular improvement over previous methods of passive listening. The passive detection of submarines by cross-correlation of a pair of widely spaced hydrophones, to ranges of the order of 100 miles was indicated. This was a factor of ten improvement over existing sonars, with the enormous tactical advantage that the target had no clue that it was detected.
In some haste, two inter-linked UK initiatives were launched circa 1953. The first was the establishment of a
CORSAIR listening shore-base on the coast of Northern Ireland at Portballintrae, in a strategically important area, with the hydrophone array looking north across shipping lanes.
From the outset, the intention of the underwater system was to provide a
nominal course and speed of detected submarines such that other units
could be directed to intercept the target. Unfortunately the sound propagation conditions were so poor that only ranges of a few miles were obtained. Therefore, it was impossible to carry out adequately the sort of operational investigations for which the station was envisaged. The title of Admiralty Marine Physical Station signifies the specific purpose of the short-lived Portballintrae
installation to intercept Soviet-operated submarines, which was shutdown for
experimental work and the site cleared by May 1955.
The second initiative involved staff at Her Majesty’s Underwater Detection Establishment at Portland (HMUDE - the equipment development arm of naval R & D) where it was assigned the sonar outfit type number 191. The project included the design and construction of hydrophone arrays to be sited on the seabed.
UK Technique Development
ARL propagation studies continued at AES Perranporth with special purpose arrays being installed in the summer of 1954 for trials to assess different methods of siting the hydrophones, in the wake of the disappointing performance from the installation in Northern Ireland. During the trials a detection range of 70 miles was achieved against a snorting submarine, and the first test of
CORSAIR detection was made against a UK silenced submarine, HMS/M SCOTSMAN. By this time, two major reductions in radiated noise from engines and propellers had been made and assessed by ARL, and
were being carried forward into the design of the PORPOISE class of
Experimental Establishment on Unst
When it became clear that detection ranges of the order of a hundred miles or more were conceivable in deep water, interest turned to siting arrays off the northern tip of the Shetlands, looking north into deeper water.
ARL's over-arching task to determine the system design parameters for
the operational system continued. Several propagation trials were carried out (1954-5) in order to assess
the likely system performance in this area, and to choose the most suitable site for the
Initial Equipment on Unst
Finally in the summer of 1955 the ARL experimental array 1200 feet long consisting of hydrophones spliced into a cable was laid 40 miles out to sea and connected back to a new shore station, AES Unst. A further
ARL array (hydrophone pair) was laid 7 miles to the West of the main array. These arrays posed many problems both in design and laying. The US were invited to participate in the forthcoming trials by providing
prototype JEZEBEL array and narrowband processing equipment, for comparison purposes. Their array
of eight hydrophones was temporarily laid in shallow water on account of defects. In July 1956 the US array was re-laid in its intended position in 350 fathoms but slowly became virtually
unusable (5 of the 8 hydrophones were unserviceable), and in August the main UK array was raised and repaired, as the watertight integrity of the junction box containing control logic had failed.
Most of the shore-based equipment (array signal processing and display) had been installed in the main laboratory of AES Unst by May 1956; such was the haste that the decoration of the
(new) laboratory was still incomplete the following year. In July 1956 staff from the Bell Telephone Laboratories in America (under contract to USN-ONR for Project
JEZEBEL) installed recording equipment and a LOFAR
Tightening of Timescales
Due to problems with the arrays and repeated delays onshore the main ARL trials were concentrated into the period from September 1956 to February 1957. The senior naval staff’s demand to telescope the R & D for a
CORSAIR system made it necessary to direct research work so that all information that might be of value in the design and siting of the development equipment (Sonar type 191) might be passed to CUDE (Captain in charge of HMUDE, Portland).
The list of ships assisting in the ARL work at AES Unst from May 1956 to May 1957 is a long one. HMS FETLAR and HMS TIREE continued to help in everyway in a number of trials to test and check the arrays. Use was made of RRS Discovery II whose experience in ARL work was most valuable. The cable ship ACS St Margarets carried out three intricate relaying operations with eminent success. The ships acting as targets during the main trials were HMS CUMBERLAND, HMS OCEAN and HMS/Ms ARTFUL, SCOTSMAN, TOKEN and TRUMP. Aircraft from Coastal Command, RAF Kinloss provided surveys of shipping in the area during the trials.
Trials at AES Unst
Many trials were carried out using this installation (in the above period), the most important being:-
A series of exercises (called THERMOSTAT
I to III) to determine the detection performance of the system against a conventional snorting submarine;
An investigation of the effect of spacing between receivers on “correlation loss”;
A semi-operational exercise called NIGHTSHADE in which an aircraft was vectored along the bearing of any targets detected.
The dates for the ARL led trials were: Thermostat I: 25 Sept to 3 Oct 1956; Thermostat II: 30 Oct 7 Nov 1956; Thermostat III: 6 – 14 Feb 1957; NIGHTSHADE:
early March 1957.
The trials were carried through with success, due to the co-operation of many authorities, and with considerable strain on ARL’s staff and resources. The Superintendent of ARL reported in May 1957
that much of value had been learned; but it is clear that to work in such a hurry is expensive, inefficient and
It became apparent from these trials that the high hopes for such a system in this area were not to be realised for two reasons:-
Detection ranges were only the order of 50 miles instead of some hundreds of miles as had been hoped. This dramatically reduced the expected area of coverage and consequently the time taken by a submarine to transit through it. Thereby reducing the likelihood of the submarine snorting during its transit and being detected and
The amount of interference from fishing vessels was far greater than considered likely.
Operational Sonar Type 191
Nevertheless preparations were already well underway for the prototype operational (Sonar 191) array to be laid by HMUDE. As a result of the propagation and correlation loss measurement both by HMUDE and ARL, the 191 array was positioned some 22 miles further north (70 miles from shore). Signals from all 12 hydrophones were brought back individually to the shore, so that the array could be electronically steered to form 40 beams covering a 180 deg arc,
centered on a north-south line. The ensuing CUDE led trial (THERMOSTAT IV) of the system was undertaken
from 27th November to 6th December 1957 and showed that a range of 70 miles could be expected against a conventional snorting submarine to the north of the array, and about 43 miles inshore of the array. Interference from fishing vessels was (once again) extremely heavy, which greatly exacerbated the interpretation of the trials records.
During the THERMOSTAT IV, ARL staff made tape recordings and the output
from their DICE 1 detection processor was of assistance to HMUDE staff.
Following the trial, the Sonar 191 installation at Unst was
re-designated Sonar 191X (experimental and unique) and handed-over to
ARL. Further, HMS TIREE, which ARL had used since 1953 was loaned to
HMUDE, although it was expected that it would soon be reduced to
reserve, the same fate as HMS FETLAR that also had assisted in many ARL
In the wake of the Sonar 191 trial undertaken by HMUDE staff, it was concluded that this
(CORSAIR) concept of fixed shore-based passive listening systems would be ineffectual against the evolving quieter submarines as typified by the UK PORPOISE class; mainly due to resilient mounting of the engines, together with improved propeller design. These submarines while snorting radiated about 3% of the noise power radiated by earlier submarines. Consequently the whole of the advantage offered by passive long array techniques using broadband or narrowband processing
was nullified, and detection ranges were once again back to a few tens of miles, at best.
Project CORSAIR's Legacy
The equipment at AES Unst was put on a “care and maintenance” basis
(see note 1, below) as the principal HMUDE staff involved were reassigned to (more pressing) active sonar work for the UK’s nuclear-powered submarines, at the behest of senior naval staff. Likewise most of the ARL staff moved onto active sonar studies under projects
LIONESS and VERONICA. Fortuitously ARL’s parallel work that commenced in 1954 (under the project KNOUT- a
topside array experimentally fitted to HMS/Ms TIRELESS and THULE) on submarine-borne, long range passive sonar was continued. The first materialisation of this type of system was engineered by HMUDE as Sonar type 186. By 1957 ways were being considered for the inclusion, in a submarine, of one of the signal processing techniques used at AES Unst (produced under project DICE) for the Sonar 186 array. The cross-correlating equipment known as DICE 1 was used successfully at AES Unst in
the key exercise NIGHTSHADE – the operational assessment of prosecuting land-based
CORSAIR detections using aircraft.
The legacy of the work that culminated in the systems and styles of
signal processing evaluated at AES Unst was principally twofold.
Processing for the type 186
array was developed under Project SOAP, the follow-on to Project
DICE. This together with ARL’s further self-noise reduction of the P-boats fed into the O-boats (and retrofitted in the Ps) proved highly successful and led to Sonar outfit type 2007 produced by HMUDE for the submarine fleet.
ARL work continued to investigate off-board long arrays leading to fleet-wide
passive towed array systems. Early towed array trials utilised the
prototype Sonar 2007 processing produced under ARL's Project SOAP.
Whilst the end of Project COSAIR had been reached, this
was by no means the end of AES Unst. For reasons yet to be established,
there was renewed interest over the period 1960-63 followed by another
quiet spell until 1968. Finally, all interest in ARLE Unst (its name
from 1969 - standing for Admiralty Research Laboratory Extension Unst)
ceased in the 1970s.
Today hints of the installations are
to be found, curiously enigmatic!
Land cables exposed on the side of Saxa Vord - photo
courtesy of Gordon Carle
Sea cables terminal building at Buddabrake, Burrafirth - photo
courtesy of Gordon Carle
NOTE 1: It was hoped to make
routine tape recordings of background noise for analysis at ARL and at
Bell Telephone Laboratories.