A review of Office of Naval Intelligence summaries from early 1963 shows a persistent concentration of Soviet-supplied assets under Cuban control. The public withdrawal of Soviet SS-4 missiles and Il-28 bombers after the October 1962 crisis did little to temper the conventional threat to American sea control. Cuban forces operated Komar-class missile boats and a large number of torpedo boats, all capable of swarm tactics in the Florida Straits. The primary underwater threat was assessed to come from Foxtrot-class diesel-electric attack submarines on extended patrols from the Soviet Northern Fleet. These were long-range boats, capable of threatening shipping lanes deep into the Atlantic. For the crews of American destroyers screening sea lines of communication, the operational area was a dense matrix of overlapping threats. A surface contact could materialize from behind a cay while a submarine lurked beneath a thermal layer.
The end of the missile crisis did not bring calm. It brought an undeclared war of nerves fought in the sonar shacks and combat information centers of destroyers. Archival evidence shows that U.S. Atlantic Fleet destroyer patrols in the Yucatan Channel and the Windward Passage were maintained at the heightened levels established during the quarantine. Gearing-class destroyers, many having undergone the Fleet Rehabilitation and Modernization (FRAM I) program, formed the core of this effort. These ships were rebuilt specifically for anti-submarine warfare, equipped with the SQS-23 sonar, ASROC anti-submarine rockets, and DASH drone helicopters. The geopolitical situation was volatile. Every Soviet-bloc freighter bound for Havana was scrutinized. Every submerged sonar contact was treated as hostile until proven otherwise. This constant pressure created an environment of extreme tension for the men on the front lines, operating on a war footing long after politicians declared the crisis resolved.
Fragmentary intelligence reports during the winter of 1962-1963 pointed toward the development of a new class of Soviet midget submarine. The vessels were unconfirmed, lacking a formal NATO reporting name or a clear project number, but the consistency of details from defector debriefings was alarming. Technical specifications, pieced together from multiple sources, described a small diesel-electric boat displacing less than 150 tons, crewed by a handful of men. Its design was believed to be optimized not for long-range patrols, but for clandestine, short-range operations from a host port like Cienfuegos or Mariel. The operational danger of such a craft was its ability to negate the technological advantage of US Navy anti-submarine platforms. The SQS-23 sonar, while powerful, was a large, hull-mounted active system designed to detect full-sized submarines in the deep ocean. Its effectiveness degraded in the acoustically complex shallow waters of the Caribbean, with their variable thermal layers and high ambient biological noise. A small, quiet-running submarine with a minimal acoustic and magnetic signature could slip through an ASW screen. The rumored design was said to carry two to four externally mounted torpedoes, a small but lethal payload capable of sinking a destroyer.
An analysis of declassified intelligence documents from early 1963 reveals a failure to anticipate the specific nature of the underwater threat in the Caribbean. The breakdown occurred within the Defense Intelligence Agency’s process for cataloging Soviet-bloc naval assets. A key document, the Federal Maritime Registry Supplement 7B (FMRS-7B), was intended to provide a comprehensive list of all known Warsaw Pact naval platforms. Archival copies of this supplement show that analysts misinterpreted human intelligence from sources inside Polish shipyards. Reports detailing the fabrication of small, pressure-resistant hull sections and compact diesel-electric propulsion systems were incorrectly categorized. They were logged as materials for deep-sea salvage submersibles, not as components for a new class of offensive midget submarine. This misidentification was compounded by an institutional disconnect. The raw data was never cross-referenced with acoustic intelligence profiles being developed by the Office of Naval Intelligence (ONI), which had noted unusual, low-frequency signatures during Soviet naval exercises in the Barents Sea. The information was effectively stovepiped.
The first hard evidence of this new danger emerged from a U-2 spy plane. On a photographic mission designated 3109, a Strategic Air Command U-2 was tasked with monitoring shipping activity in the Cuban port of Nuevitas. A secondary pass over the coastline of Cayo Romano to the northwest produced a significant finding. Tucked into a small, unnamed inlet on the cay’s northern shore, facing the deep water of the Old Bahama Channel, photo interpreters at the National Photographic Interpretation Center identified a series of suspicious structures. They were camouflaged with netting and local foliage, but the shapes were unmistakable: three concrete pens, partially submerged at high tide, just large enough to house a vessel approximately 75 feet in length. Further analysis of the imagery revealed a large, collapsible diesel fuel bladder hidden in a nearby mangrove cluster, torpedo-handling dollies, and a small barracks. The location was remote yet strategically perfect, offering direct, deep-water access for clandestine operations.
The discovery sent a shockwave through the U.S. Navy command structure. The existence of the Cayo Romano base was a complete tactical surprise. Nothing in existing threat assessments, including the flawed FMRS-7B, indicated the Soviets or their Cuban allies possessed the capability to forward-deploy a submarine of this type. Operational logs from CINCLANTFLT headquarters reveal a scramble to understand the new weapon. Its small size, quiet diesel-electric drive, and presumed operational profile made it a nightmare for the ASW tactics of the day. The SQS-23 sonar systems on the FRAM destroyers were optimized for detecting large submarines in the open ocean, not for picking a tiny target out of the acoustic clutter of coastal waters. Within hours of the U-2 photo analysis, emergency directives were issued to the Naval War College and the ONI to draft new tactical bulletins and re-evaluate all existing anti-submarine warfare doctrine for the Caribbean theater.
Deck logs and action reports from February 1963 place the destroyer escort USS Cromwell (DE-1015) at the center of the crisis. The Cromwell, a Dealey-class escort, was executing a standard barrier patrol within a designated 80-by-80 nautical mile sector designated Patrol Box Sierra. This area was situated directly astride the Old Bahama Channel, northeast of the newly identified submarine pens on Cayo Romano. Her mission directive was simple: steam a repeating picket pattern at 13 knots and report any submerged contacts. The ship’s Plan of the Day for 12 February 1963 shows the vessel at Condition of Readiness III, a standard wartime steaming posture. For the sailors in the sonar control room, the mission was a grueling exercise in auditory focus. They were tasked with interpreting returns from the ship’s SQS-23 sonar, a system retrofitted onto the Cromwell as a testbed for shallow-water operations. The acoustic environment of the channel was a known problem, a mix of thermal layers, unpredictable currents, and the constant crackle of snapping shrimp, all of which created a wash of false returns.
The ship was alone.
At approximately 15:02 local time, the Cromwell’s primary link to fleet command went silent. A technical examination of the ship's radio logs indicates a sudden failure in communications. The secure teletype link, transmitted via a high-frequency AN/URC-32 transceiver system aimed at a relay station in Key West, ceased its chatter. The radioman on watch attempted to re-establish the link using the backup transceiver. It also failed to acquire a signal. Electronics technicians diagnosed the issue not as an equipment malfunction aboard the Cromwell, but as a complete loss of the incoming signal. The ship was receiving nothing. The commanding officer was now bound by strict Emission Control (EMCON) doctrine, which forbade any transmissions that could betray his ship's position. Actively trying to hail the fleet would light up his ship on any enemy's electronic support measures receiver. His last valid order was to patrol Box Sierra. Without new instructions, he was to continue, steaming in a predictable pattern, deaf to the outside world.
With all external input severed, the ship’s Combat Information Center (CIC) became the vessel’s entire universe. The pre-engagement readiness of the Cromwell’s anti-submarine systems was at its peak. The ship’s sonar officer had the SQS-23 sonar set to an aggressive active search pattern. Every ninety seconds, the system sent a powerful 5-kilowatt pulse of sonic energy into the channel. On the main sonar console, the operator had the system configured for the difficult environment, with the range scale shortened to 8,000 yards and the Doppler filters adjusted to highlight any object moving at less than 5 knots. The ship’s primary anti-submarine armament, the RUR-5 ASROC launcher forward of the bridge, held eight rocket-thrown torpedoes ready for immediate firing. On the fantail, the experimental DASH (Drone Anti-Submarine Helicopter) system was brought to a heightened state of alert. A single QH-50C drone sat on the flight deck, fueled and armed with a lightweight Mark 44 acoustic homing torpedo. Every system was energized.
The Cromwell’s sonar logs and action reports pinpoint the engagement window as lasting from 15:37 to 15:42 local time. At 15:37, the sonar operator detected a faint, intermittent return. It was a whisper of sound barely distinguishable from the background noise. The target was designated S-1. The initial signature was inconsistent with a full-sized Foxtrot or any known biological source. It was slow, estimated at less than four knots, and its acoustic return was maddeningly inconsistent, appearing for a few seconds before fading back into the clutter of the shallow channel. The sonar officer, alerted to the faint contact, ordered the SQS-23 system into a high-repetition search pattern, flooding the bearing with active pings. The return sharpened slightly, resolving into a classification that sent a jolt through the CIC: probable non-biological, mechanical origin. The target was real.
The fight unraveled almost before it began. At 15:38, as the fire control team worked to develop a firing solution, the first failure occurred. The primary plan position indicator (PPI) scope for the SQS-23 sonar flickered and went dead. A post-action analysis pointed to the burnout of a single 6SN7GT vacuum tube within the display’s sweep generator circuit, a minor component failure with major consequences. Without the visual display, the sonar officer was now fighting blind, relying entirely on the raw audio feed in his headphones and a secondary, less precise analog needle on the bearing deviation indicator. He could hear the target, a faint swish of a slow-turning propeller, but he could no longer paint a clear picture of its location for the rest of the CIC. The tactical plot became a mess of grease-pencil approximations based on yelled instructions from the sonar shack.
Seconds later, at 15:39, this cascade of failure spread to the ship’s primary weapon system. The commanding officer, having decided to engage based on the tenuous audio contact, ordered the forward Mark 112 ASROC launcher to train toward the estimated bearing of S-1. The fire control computer had a solution. The launcher itself remained frozen, aimed dead ahead. A frantic report from the launcher captain confirmed the second malfunction: a hydraulic leak in the traverse mechanism. The launcher was stuck. With the main sonar display dark and the ship’s primary anti-submarine weapon disabled, the Cromwell was in a desperate situation. The midget submarine, still methodically closing, had effectively neutralized the destroyer escort’s advanced systems without firing a shot.
This forced a turn to improvised tactics. The commanding officer had one weapon available: the experimental QH-50C DASH drone helicopter on the fantail. The drone was armed with a single Mark 44 lightweight torpedo. Under normal doctrine, the drone would be guided to a precise point identified by the SQS-23 sonar. With the sonar display gone, this was impossible. A new plan was devised out of necessity. Between 15:40 and 15:42, the ship’s drone controller in the CIC was patched directly into the sonar operator’s audio feed. The drone was launched. Its controller attempted to fly it “down the bearing” based entirely on the sonarman’s shouted descriptions of the audio contact’s strength and character. The Cromwell itself began a hard turn to starboard, trying to unmask its own hull and get a different acoustic angle. The drone controller had to visually estimate the drone’s position relative to the ship, while the sonar operator tried to interpret the shifting sounds of the unseen target. The final command was a calculated guess, ordering the drone to drop its torpedo at a timed interval, hoping to place the weapon in the submarine’s path.
The detonation of the Mark 44 torpedo was not a direct hit. The Cromwell’s post-action damage report indicates the weapon, dropped from the DASH drone on an estimate, exploded on a proximity setting approximately seventy-five yards off the destroyer escort’s port quarter. The underwater shockwave was powerful. It lifted the ship’s stern several feet, slamming it back down into the resulting void with a structural groan that resonated through the hull. While the hull itself remained intact, the whip-like concussion tore through the vessel. In the aft engine room, sailors were thrown against machinery. In the Combat Information Center, the few remaining operational displays shattered. The most severe human cost occurred on the exposed main deck, near the ASROC launcher.
The order to fire the disabled ASROC had never been given, but the launcher crew remained at their station. They were caught completely exposed by the nearby underwater blast. The ship’s lone Hospital Corpsman, a Second Class Petty Officer named Reynolds, established a triage station on the slippery steel of the mess deck. His initial actions, documented in the ship’s medical log, focused on controlling severe hemorrhage and treating traumatic injuries. Two sailors from the ASROC detail had been hurled against the launcher’s blast shield, suffering compound fractures and internal injuries. Another was knocked unconscious, sustaining a severe concussion. Reynolds, with only his standard medical kit and two stretcher-bearers, worked under the dim red glow of battle lanterns to apply tourniquets and pressure dressings. His formal reports note the difficulty of administering morphine in the violent, unpredictable motion of the damaged ship.
Compounding the chaos was the effort of Photographer’s Mate Third Class Collier to document the engagement for intelligence purposes. After-action interviews reveal the technical challenges he faced. Armed with a bulky, manually operated Graflex Speed Graphic press camera, Collier’s primary objective was to capture an image of the submarine’s periscope or any surface debris. The shock of the torpedo blast had knocked him to the deck, cracking the casing of his primary light meter. Acrid, white-grey smoke from the still-smoldering rocket motor of the jettisoned ASROC weapon clung to the deck, reducing visibility. His attempts to get a clear photograph of the torpedo’s splash point were frustrated by this chemical haze and the fading light of the Caribbean dusk. Each shot required him to manually change a flashbulb and reset his focus in a pitching, smoke-filled environment, all while damage control parties shouted warnings. Not a single one of his photographs would prove useful for post-mission analysis.
The combined effect of smoke and blood transformed the Cromwell’s weather decks into a hazardous environment. The deck plates, already slick with sea spray, were now coated in patches with blood from the wounded being moved to the mess deck. This created treacherous footing for the damage control teams attempting to assess the ship’s structural integrity. The thick, choking smoke from the ASROC motor was not merely a visual obstruction; its chemical composition caused uncontrollable coughing and watering eyes among the crew, hampering every effort. A review of the damage control officer’s log notes multiple instances of sailors slipping and falling, complicating the evacuation of the injured and the deployment of firefighting equipment, even though no major fire had broken out. This sensory assault created an atmosphere of extreme confusion.
An examination of the declassified after-action report filed by the Cromwell’s commanding officer, designated CINCLANTFLT AAR 14-63B, reveals a narrative stripped of its chaotic reality. The official accounting presents a version of events that is technically accurate in its sequence but misleading in its character. It records a sonar contact, a decision to engage, and the deployment of a DASH-delivered weapon, resulting in the presumed neutralization of the threat. A close reading of the document shows conspicuous omissions. There is no mention of the failure of the SQS-23 sonar’s primary display scope. The hydraulic failure of the ASROC launcher’s traverse mechanism, which left the ship’s primary anti-submarine weapon useless, is similarly absent from the summary. The report frames the use of the QH-50C drone not as an act of desperation, but as a deliberate tactical choice. This sanitized version served a clear institutional purpose, portraying a textbook engagement that validated the Navy's investment in FRAM upgrades and experimental drone programs. Admitting that a new midget submarine had nearly neutralized a modern destroyer escort through a cascade of minor technical failures was an unacceptable conclusion.
The arbitrary nature of injury aboard the Cromwell is a brutal footnote to the engagement, one absent from the tidy official summaries. A review of the ship’s medical logs and crew muster reports shows that survival was dictated by a sailor’s physical location at the moment the Mark 44 torpedo detonated. The three most grievously wounded men were all part of the ASROC launcher crew, stationed on the exposed 01 level, directly above the point on the hull that absorbed the main force of the underwater shockwave. The explosion, though not a direct hit, transmitted the full violence of the concussion up through the ship’s structure. These men were incapacitated not by enemy action, but by the physics of their own weapon’s blast. In contrast, sailors in the aft engine room, deeper within the hull and shielded by multiple steel decks, were violently thrown but largely protected from the direct water-hammer effect. Personnel in the Combat Information Center were likewise spared the primary shock but were subjected to a shower of shattered glass from instrument panels. Survival was not a function of rank or training, but of proximity to the concussive force traveling through the ship’s frame.
The historical record of the five-minute battle is defined by what is missing. The lack of comprehensive documentation made a full reckoning of the event impossible. Photographer’s Mate Collier’s failure to obtain any usable imagery left a critical intelligence void. There were no photographs of a periscope, no images of a debris field, nothing to visually confirm the nature or even the existence of the target. The technical documentation was equally compromised. In the absence of the SQS-23’s plan position indicator scope, no definitive plot of the submarine’s course and speed was ever generated. The sonar logs from the period consist of the operator’s handwritten notes, based on his subjective interpretation of audio-only signals under extreme stress. Without magnetic tape recordings of the sonar feed, a capability not yet standard on destroyer escorts in 1963, the sound of the propeller was a ghost, existing only in one man’s memory. The ship’s main deck log, the primary legal and historical document, reduces the engagement to a few lines of sterile, passive-voice naval jargon, leaving a permanent and unresolvable gap in the official account.