6 August 2010
From: B. Rule, 3931 Brookfield Ave, Louisville, KY 40207-2001
To: VADM David J. Dorsett, Director of Naval Intelligence, Office of Naval
Intelligence, 4251 Suitland Road, Washington, DC 20395-5720
Subj: Why the USS SCORPION (SSN-589) Was Lost on 22 May 1968
Ref: (a) Originator’s ltr of 14 Mar 2009
(b) SCORPION SAG Report: “EVALUATION OF DATA AND ARTIFACTS
RELATED TO THE USS SCORPION (SSN-589) (U)” of 29 June 1970,
prepared for presentation to the CNO SCORPION Technical Advisory Group by
the Structural Analysis Group: Peter Palermo, CAPT Harry Jackson, Robert
Price, et al.
(c) Originator’s ltr of 28 Oct 2009
Encl: (1) Enclosure (1) to Originator’s ltr of 14 March 2009
The USS SCORPION was lost because hydrogen produced by the 65-ton, 126-cell TLX-53-A main storage battery exploded in two-stages one-half second apart at 18:20:44Z on 22 May 1968. These events, which did not breach the pressure-hull, prevented the crew from maintaining depth-control. As discussed by reference (a), the SCORPION pressure-hull collapsed at 18:42:34Z at a depth of 1530-feet. Noted times are actual event times on board SCORPION.
This assessment is NOT the generic attribution of the loss of a submarine to a battery-explosion advanced as a default explanation in the absence of any more likely construct. This assessment is based on (1), the results of examination and microscopic, spectrographic and X-ray diffraction analyses of recovered SCORPION battery material that confirm an explosion occurred, and (2), the July 2008 reanalysis of the SCORPION “precursor” acoustic signals that identified these signals as explosions contained within the SCORPION pressure-hull. Collectively, these findings indicate battery explosions were the initiating events responsible for the loss of SCORPION on 22 May 1968.
DISCUSSIONS: EXAMINATION AND METALLURGICAL ANALYSIS OF A RECOVERED SCORPION BATTERY COMPONENT
Section 7.1.3, page 7.2 of reference (b) states: (quote) ….the general battery damage is violent. The high velocity intrusion of pieces of the flash arrestor into both inside and outside surfaces of the retrieved plastisol cover attest to violence in the battery well. The damage to the terminal battery post coupled with the violent tearing of the plastisol covers indicates the possibility of a battery explosion. While it is possible that this damage could have been an after-effect of hull implosion, the SAG (Structural Analysis Group) feels that the intrusion of particles into the plastisol cover would have been much less severe had water been in the battery well at the time. (end quote)
Section 5.3.6, page 5.17 of reference (b) states: (quote) The battery installed in SCORPION was a TLX-53-A, manufactured by Gould-National Battery, Inc. Battery cell debris is in evidence over the entire debris field. Table 5-2, page 5.38 provides a list of the battery debris identified by the Portsmouth Naval Shipyard analysis team. (end quote) Comment: Table 5-2 notes damage from heat and melting. The presence of melting eliminates the possibility that such damage occurred as a result of pressure-hull collapse (implosion) because analysis of acoustic data discussed by Section IV of reference (c), confirms SCORPION was fully-flooded within 0.112-seconds of pressure-hull and bulkhead collapse; hence, the melting damage (and the battery explosion) had to have occurred within the still-intact SCORPION pressure-hull.
In consonance with this conclusion, Section 5.3.6, page 5.17 of reference (b) also states: (quote) the Portsmouth Naval Shipyard Analysis Group reports that the available evidence indicates the battery probably exploded at some time before flooding of the battery well occurred. Review of Figure 5-13 indicates that the threads on the terminal posts were sheared off and there are no cover seal nuts remaining. This indicates that an explosion took place on the inside of the cells. The covers were completely blown off. Had the pressure been applied on the outside of the covers, the cover support flange on the terminal posts would have held pieces of the covers and it is expected that the cover seal nuts would have remained in place in at least some instances. ( end quote)
Further, Section 5.3.6c, page 5.18 of reference (b) states: (quote) The (battery cover) sample from SCORPION had been violently, but locally, torn, particularly at the location of the bus connection bolts and nuts. The deformation in this region appears to have started on the inside, or battery side of the cover. (end quote)
And finally, Section 5.3.6e, page 5.18 of reference (b) states: (quote) Some 20 equally small (nearly sub-visible) fragments of material were imbedded at high velocity in both the inside and outside of the sample. The trajectories of the fragments were essentially random, ranging from grazing to vertical incidence. Microscopic, spectrographic and X-ray diffraction analyses reveal that these fragments are identical in composition and structure to the alumina flash arrestors used on the batteries in SCORPION. (end quote)
DISCUSSIONS: SCORPION ACOUSTIC DATA
Enclosure (1) to reference (a), forwarded as enclosure (1) to this letter, provides detailed discussions of four independent lines of evidence that, collectively, established, for the first time, that the two “precursor” acoustic events that occurred at 18:20:44Z, 21-minutes and 50-seconds before hull-collapse, were explosions from then unidentified sources that were contained within the SCORPION pressure-hull. The energy yield of these explosive events, now assessed to have been battery-associated, is estimated to have been no more than about 20-lbs of TNT each.
The July 2008 identification of the precursor acoustic events as explosions contained within the SCORPION pressure-hull strongly supports the battery explosion conclusion advanced by reference (b), i.e., the acoustic data identifies the actual explosive events previously assumed by the authors of reference (b), the SAG Report, to have occurred based on the observed damage to a recovered battery component discussed above.
Collectively, the above information indicates the two acoustic events that occurred 0.5-seconds apart at 18:20:44Z were produced by explosions associated with the SCORPION TLX-53-A battery, and were the initiating events responsible for the loss of SCORPION on 22 May 1968. Additional information will be provided as developed.
Copy to (w/ encl):
The RV Erline in transit between Argus Island and Tudor Hill Laboratories.
A view of Tudor Hill Laboratories’ listening post at the crown of the hill (a very common siting arrangement).
One of the cables in a SOFAR installation.
This shows the tapering of the line between the sea deployment of SOSUS hydrophones the Terminal Building where the leads are terminated into audio filters, storage, time shifters (delay lines built into the storage), and finally verniers (chart recorders) that make LOFARgrams.
And, of course, the cable has to emerge somewhere along a beach.
At this point I need to point out that the beach picture with cable, and earlier pictures are presenting the common design considerations for nearly any SOSUS station, but not for NavFac Bermuda which received data through RF transmissions from Argus Island.
Still, audio is audio, and the many Verniers (chart recorders) at Tudor Hill Laboratories operated the same irrespective of this link.
Aerial view of NavFac Bermuda, Tudor Hill Laboratories.
The speed of sound in the ocean varies with depth, but with a twist. And in that twist the United States Navy held a three decade technical lead in submarine detection.
The scenario, below, presents a case where a submarine is emitting a sound. It is immersed in a liquid that is layered with denser water above and below a region known as the SOFAR Channel. The submarine is in the top, dense layer. Within the graphic, the SOFAR Channel is the light blue layer with the converging strong lines of sound from the submarine. Note the shape of the lines are not straight lines. They bend due to the variation in water density like a lens. Within this channel, sound will propagate with far less attenuation (fading), and travel many hundreds of miles to remote detectors.
There are a number of dashed rays of sound that encounter loss in reflection from the bottom, or are situated along a path that does not lead into the interior of the SOFAR Channel.
For those who want actual numbers, the speed of sound varies in the following manner:
This chart presents the electrical signal from the underwater SOSUS array.
It is listening to ship traffic (and presumably with more interest in submarines). The scale across the bottom reveals a span of 8 or minutes of recording in hardcopy, colored format (in the old days, it was all black and white). Above the minute marks the scale at the left reveals the frequency of detection, within the field of color, the brighter the color, the louder the source.
As such, the span from very low frequencies up to 50 Hz (a power line kind of hum) has a bright patch across all time with very white (thus very loud) source noise from propellers that scar the field with horizontal lines showing at 30 Hz and 40 Hz (and other frequencies).
The picture below is more representational of the LOFARgrams observed during the 50s, 60s, 70s …. The aspect is as viewed on the Vernier (chart recorder) with the paper being advanced up as new data was burnt (no ink, the paper is thermally sensitive) into a line at the bottom. In this case, louder is equal to darker (more burning). With many Verniers running plots along bearing angles, there was a lot of burning going on, continuously.
Not labeled are the propeller shaft noise profiles. The shaft’s fundamental frequency is at the far left, about an eighth of an inch from the edge. Its first harmonic is an eighth of an inch to its right (and many more barely visible harmonics across the page).
As of this time, the picture is incorrectly labeling the propeller blade fundamental as an harmonic. Given its harmonics fall within groups of four shaft harmonics, it would seem to me that there were four blades on the shaft.
source: United States Navy
The floor’s stuffing tube contains a slanted, vertical cable tray. Significant are the two 100mm cables that go to the two arrays.
There were two SOSUS arrays of hydrophones reaching out from Argus Island, one towards the North-West and the other to the South-West. From the general specification:
source: United States Navy
This map displays SOSUS Atlantic listening stations lining the continental shelf of the United States.
Their mission was to detect and track submarines in the Atlantic ocean.
This was accomplished through a distributed underwater network of angle sensitive audio sensors. These directional microphones are connected at the end of a many miles long submarine cable out to the continental shelf that drops thousands of feet to the ocean floor. (The white coloring in the relief map reveals the shelf face feature.)
The SOSUS listening stations are located on coastal land. Their underwater hydrophone sensors are many miles out at sea (80 miles or more), and often half a mile deep in the ocean’s bottom or on the edge of the shelf (or island’s seamount in the case of Bermuda).
Robert A. Frosch, in his capacity as Technical Director of Hudson Laboratories (Columbia University), was technical director of Project Artemis.
Argus Bank (more properly known as Plantagenet Bank) is a submerged sea mountain that supports the man-made Argus Island (offshore platform, also called a Texas Tower). Argus Island is a SOSUS transmission line transfer point between the underwater sensors along Plantagenet Bank and the listening post located in Tudor Hill Laboratories (at the south edge of Bermuda island). Transportation of men and supplies was done between Tudor Hill Laboratories and Argus Island via R/V Erline.
Argus Bank had hosted the earlier Project Argus which employed a 1MW 400Hz SONAR. Aside from the active transmitter, much of that technology is similar.
The three seamounts are volcanic remnants from continental drift over a magmatic hot-spot.
source: United States Navy
These are views of Argus Island taken from the official report of its destruction.
Argus Island was an outpost SOSUS station between two arrays of underwater sensors and the Tudor Hill Laboratory located on the southern edge of Bermuda. It was manned by:
One supervisor , electronic technician.
Two senior electronic technicians (both with broad qualifications including commications and digital circuits).
Two diesel mechanics/diesel-electrical , welding, plumbing and crane operator.
Two cook-baker stewards.
One general helper/maintenance Janitor.
Argus Island also hosted non-SOSUS research (to aid in its cover story of being the equivalent of a benign weather station). Among those activities was SeaLab in 1964.
Quoting from a letter of Munk’s: …we were able to discover in the general wave record a very weak low-frequency peak which would surely have escaped our attention without spectral analysis. This peak, it turns out, is almost certainly due to a swell from the Indian Ocean, 10,000 miles distant. Physical dimensions are: 1 mm high, a kilometer long
This presents how the calls of the Humback whale would appear in modern SOSUS technology displays.
The top reveals a wider frequency range than is normally observed in SOSUS operation, but that is because it has been sped up by a factor of ten to allow the calls (normally sub-sonic) to be heard in sound recordings. This, then, suggests that the LOFARgram is not of live data, but of a sped-up recording.
If you apply a divide by ten to the left scale, we are back to the conventional LOFARgram display covering from 0 Hz to 50 Hz.
SOSUS technologists saw these waveforms for years, and were unable to determine their origin. Further complicating matters were that these noises moved. Because the early stages of SOSUS was called Project Jezebel, these noises were tongue-in-cheek attributed to the Jezebel monster.
The Argus Island Project Artemis array has two specifications as the project moved from active to passive SOFAR application. source wikipedia:
shipboard active array:
50 ft (15 m) by 33 ft (10 m) and
weighed 400 tons. It consisted of
1440 individual transducers (200 pounds apiece) in a
48 by 30 configuration.
fixed passive receive array:
10 Mile long
ten strings of hydrophones mounted on
200 eighty-foot towers laid down the side of Plantagenet Bank in Bermuda.
The strings were laid on the side of the bank using the U.S. Navy large covered lighter YFNB-12, reconfigured with a long overhead boom to handle the towers. Each cable had special takeouts built into it at intervals from which wires to the hydrophones were connected. Each tower was clamped onto the special cable with takeouts. At the upper end of the approximately 4-inch (100 mm) cable a wire rope was attached and led to an explosively embedded anchor shot into the flat coral top of Plantagenet Bank. Tension of more than 40,000 lb was applied to the wire rope and cable to lay it down the side of the bank in the straightest line possible. At one point all further construction ceased while a stopper was placed on the special cable because most of the connection to the wire rope had broken and the string was being held by a few strands of wire on the double drum winch on YFNB-12. The YFNB-12 was held in place with four Murray and Tregurtha Diesel outboard engines placed on the corners and capable of 360 degree rotation, developing tremendous thrust in any direction. The cables led to Argus Island tower, from which the signal was conducted to the United States listening post located at Tudor Hill, Southampton, Bermuda ( ) that had opened on June 1, 1955. At the time the post was classified top secret. Tudor Hill Naval Facility Bermuda was closed in 1995. The facility shares a short road with the Pompano Beach Club.
Picture of the probable interim design.
Picture of the probable final design.
These two views may not be with the same hydrophone array. The second is obviously larger than the first. Given the resolution’s rough ability to count the elements, this makes for difficulty in determining their construction details. Focusing on the second, my first impression was to call each white spot in the field of frames an element. However, there are white square covers over some of what should be elements (I would speculate they were there to mark what element had not been inspected yet—or had failed a test), and these covers center on the black areas surrounding the white areas. Thus the white areas are part of the frame.
The first picture’s array is in 3×5 panels, each panel is of 5×12 elements (900 total elements), but this does not conform to the specification above in any way—even if additional panels were added. However, as this is a difficult interpretation given both the poor resolution and the camera angle, each panel could be composed of 6×12 elements, and the entire assembly is only short a row of panels. I would like to think this first picture is displaying the array under partial construction.
The second picture’s array shows 4×5 panels, each panel is of 6×12 elements (1440 total elements) which does conform in layout. This also conforms to the extract from Project Artemis Acoustic Source Performance Characteristics:
The Artemis acoustic source was designed to meet the requirements for an ocean surveillance study program. These requirements included a source level of 152 dB in a 100-Hz band centered at 400 Hz with a transducer operating depth of 1200 feet. The transducer, which was completed in June 1964, is a rectangular planar array 33 feet wide and 50 feet high.
The ocean surveillance study program, Project Artemis, initiated by the Office of Naval Research in 1958, required a very high power, deep acoustic source. The acoustic requirements established. for the projector called for a capability of radiating 1000 kW of acoustic power in a 100-Hz band centered at 400 Hz with pulse lengths of 10 to 60 seconds at a 10% duty cycle. Beamwidth to the half-power points was to be 20 degrees in the horizontal plane and 12.5 degrees in the vertical plane. This combination of power and beamwidth would result in a source level of approximately 152 decibels relative to 1 microbar at 1 yard (dB//1 gbar at 1 yd). The acoustic axis was to have a fixed orientation of 11 degrees above the horizontal plane.
Several proposals for the projector were considered, the final selection being a rectangular plane array 33 feet wide by 50 feet high composed of 1440 variable-reluctance transducer elements. Each element is nearly cubical, being 11-1/8 inches square on the radiating face and 11-3/4 inches deep. They are assembled in frames in which 72 elements are closely packed in 12 rows by six columns. Each assembly, referred to as a module, is approximately 6 feet wide. by 12 feet high. The completed array consists of far rows of modules with five modules in each row. The modules are mounted on a suitable array frame which provides the proper tilt angle of the radiating face and supports auxiliary components associated with the electrical input to the elements. Radiation to the rear of the array is suppressed by a system of pressure compensated, flattened, gas-filled tubes which serve as acoustic pressure releases on the rear faces of the elements.
Extracts from Project Artemis Acoustic Source Characteristics Of The Type TR-11F Transducer Element (which concerns itself with an upgrade in impedance to help with transmitter loading, and with diaphragm spring displacement):
The present configuration of the ARTEMIS source employs TR-11C elements with a parallel electrical connection of groups of six series connected elements.
(source unknown) The SOSUS application of these elements in an array suggest a receive sensitivity of 1W audio power at 1,000km.