Somewhere in a closet in Washington, DC, sits a shoebox-size red plastic bin. A typed paper label reads: “Accident #DCA16MM001. Operator: TOTE Services. Accident date: 10/01/2015.” Inside is what looks like piece of junk, a nondescript steel canister the size of a toaster. Someone affixed a red tag to one of the canister’s handles with wire, like something you’d find hanging off the toe of a cadaver. A cable hangs out from the top of the canister, its end severed and frayed.
This cable once led to the microphones on the bridge of El Faro.
The thing may look completely unremarkable, but the story of how it finally landed here is.
In August 2016, a team boarded the Apache once again to recover the VDR, this time at a cost of $500,000. Once again, they sailed out beyond San Salvador and idled the navy vessel above El Faro’s final resting place. Once again, they used the Apache’s powerful winches to lift the bright yellow CURVE-21 off the deck, swing it over the side, and gently lower it into the ocean. Once again, the vehicle’s navigation system quickly failed, leaving them to rely on dead reckoning as it picked its way through the mountains of debris on the ocean floor.
Fortunately, the NTSB had created a map of the accident scene by stitching together forty thousand photos of the wreckage taken during the two previous undersea operations. The CURV’s operators—contractors from Phoenix International, a marine services company—used this digital tableau to navigate the vehicle, equipped with a small metal basket, safely to the VDR.
When it touched down, the CURV puttered lazily toward its target, continually checking its surroundings to make sure that its fiber-optic umbilical cord wasn’t getting tangled in the junk all around it. This maneuvering took a couple of hours—there was no room for unforced errors.
The CURV operators aboard the Apache were remarkably skilled at manipulating the robot’s Swiss Army–like appendages; they could use their shipboard laptop console to get the thing to tie a knot into a string. They drove the CURV toward El Faro’s twisted mast—now a helpful guidepost in the murk—where the VDR had patiently waited for them for eleven months. When the CURV approached its prize, it set its basket down beside the VDR, kicking up fine particles that dusted everything in the deep, temporarily obscuring the operator’s vision. It waited for things to settle back down. Then it got to work.
Two quick-release clips held the steel canister to the mast. The CURV’s robotic pincers reached out and delicately flipped up the clips. Click. Click. Then it gripped the VDR’s two steel handles and pulled the canister out of its cradle. The VDR was free.
The CURV lowered its precious bounty into the basket, closed and secured the top, and gingerly retreated back to the surface with its prize, like a shopper returning from an underwater market. Doug Mansell, an NTSB engineer, leaned over the deck of the Apache on the night of August 6 squinting into an ocean illuminated by the vessel’s spotlights. In a few hours, the CURV emerged from the deep.
It was hoisted aboard and relieved of its basket. Finally, Doug held the thick steel capsule in his hands. It had taken three missions, six weeks at sea across ten months, and $3 million to get this far.
Doug had to open the canister immediately, which made him unusually nervous. He worked with black boxes all the time, but always under the highly controlled conditions of his lab in Washington, DC. Normally, he’d take his time, carefully documenting his progress. But he didn’t have that luxury. A team from the VDR’s manufacturer in Sweden was waiting to hear that all was well—that the unit had been successfully recovered from three miles down and that it was in decent condition, not waterlogged or corroded. As soon as they got that message from Doug, they’d board a plane for the US and meet him at NTSB headquarters to oversee the downloading of whatever data their device had managed to capture.
All people aboard the Apache crowded around Doug, anxious to see what was inside the canister—whether it had resisted the ocean’s forces for so long—but he wanted to work under the best conditions possible. Changes in pressure from fifteen thousand feet below to the water’s surface could have created an explosive condition inside the capsule.
Doug retreated to a supply room where he’d prepared a makeshift lab, carrying a cooler filled with seawater in which the VDR sloshed around. A few men from the team followed to watch Doug and shoot photos.
Wearing safety glasses, black latex gloves, and his navy blue T-shirt emblazoned with a yellow NTSB on the back, Doug lifted the capsule out of its watery bed and placed it on a table, protected with a white towel. The outer steel cylinder was about fourteen inches high and eight inches in diameter with a screw-on top, sealed with rubber gaskets. Remnants of the reflective tape that Eric Stolzenberg had seen on the second mission still stuck to the canister, though a lot had worn off during the VDR's journey, leaving traces of rough adhesive ringing the can. Two steel handles welded onto one end had made it easier for the CURVE’s pincers to grasp the thing three miles down, and Doug could see fine marks left by the robot’s metal appendages.
Doug used hand tools—a couple of socket wrenches, adjustable pliers, and Allen wrenches—to deconstruct the base of the outer canister and pry it open. Claylike insulating material packed inside to protect the inner canister spilled out, leaving a pile of brown dirt on his towel. A plasterlike polymer—designed to withstand the great pressure of a deep ocean—surrounded the inner steel cylinder. Doug delicately chipped away at it like an archaeologist, keeping the pieces for future study.
The inner steel canister was four and a half inches long, one and a half inches in diameter. Doug clamped the head of it in a vise and slowly unscrewed the top.
A circuit board the size of half a graham cracker slid out. This was where all the information was stored, on one of those microchips. He breathed a sigh of relief. After spending so much time under intense pressure, the inner chamber was dry.
When Doug got the VDR back to DC in August 2016, the Swedish representatives of the unit’s manufacturer met with him to help connect the memory chip to a forensic write-blocker to prevent accidental erasing of the media while downloading it. They made a copy of the information so that Doug could work off a facsimile, not the original, again preventing any possibility of error. And then the whole assembly got tossed into the red bin and filed away. It has sat there ever since. Officially, TOTE owns the VDR. If the company ever wants it back, it’s theirs.
Doug wasn’t sure what kind of data they’d be able to recover. He knew that sometimes things don’t work out. Equipment fails all the time.
Like many twentysomething high-tech specialists, Doug has a matter-of-fact attitude about his work. He’s passionate about the scientific challenges of extricating data from mangled iPads and steel boxes, but he isn’t emotionally affected by what he might eventually hear. He’s listened to pilots’ final words before fatal collisions. He’s heard train engineers in the midst of cataclysmic crashes. He’s heard copilots and crew catching their first and last look at the plane, bird, or mountain that will kill them.
When I asked him what it was like to listen to El Faro’s audio, he said blankly, “It’s my job,” and shrugged, knowing he couldn’t give me what I was looking for, silently wondering why reporters keep asking him this. What’s there to say?
Doug loaded the manufacturer’s playback software and visually scanned the digital representation of the audio. Twenty-six hours of something had been caught on the chip, but would it be clear enough to transcribe? Would it be white noise or gibberish? Or worse, had the VDR stopped recording two years ago or two weeks ago?
Did they even have a recording of the right voyage?
To find out, he put on his headphones and skipped to the end of the tape. There was lots of noise from the wind outside, things crashing. And there was yelling, panic. He adjusted his acoustic filters and listened harder. The sounds of a ship’s final moments were unmistakable. Doug had no doubt they had caught El Faro’s demise.
Twenty-six hours of audio on a black box leading up to a fatal accident was unprecedented for the NTSB. Usually they got thirty minutes, maybe two hours max. This particular VDR model, designed for the marine environment, was supposed to record a minimum of twelve hours. It’s a looped system that writes over as time continues, and the chip was big enough to hold more time, so it simply kept recording.
Doug exported the five audio tracks from the six microphones installed in the ceiling of El Faro’s bridge into separate files. Each track would be heard through software that had various filtering capabilities. He spent the next two days quickly listening to the entire VDR, making notes, creating charts, orienting himself to milestones, like watch changes.
Most accidents aren’t as dramatic and protracted as El Faro. For plane accidents, it’s usually a quick hit. Doug says that the pilot might say “mountain” or “crane” and then hit it and the audio goes dead. But this tragedy unfolded over the course of days. Doug knew that, at the very least, they would transcribe every word of the last four hours.
The NTSB and US Coast Guard assembled a small expert team to listen to and type up what they thought they heard. Each member of the team sat in the listening room wearing a pair of headphones. Doug would play a snippet of the audio, and they would tell him what they thought El Faro’s crew said, and who they thought said it. Danielle was easy to distinguish from the others because she was a woman. But others sounded very similar to one another. If there was a disagreement, there would be a discussion. They wanted to get everything just right. No one wanted to misquote the dead.
They started with the last four hours leading up to the sinking.
The end of El Faro audio was very difficult to hear, not just because it captured a heartbreaking moment, but because as the ship sailed closer to hurricane’s eye, the quality of the audio signal significantly degraded. There was more wind howling around the navigation bridge, and the boat was shaking more, creating a lot of background noise. The blackout curtain around the chart room hung from ball bearings that slid in a metal track; they constantly rattled when the boat shook, making conversations there more difficult to hear.
By the final hour, all the audio channels except one were blown out; Doug was left with two mics—one over the chart table, one over bridge left—feeding into a single channel. If a layperson were just to play it as recorded, they wouldn’t be able to hear much of anything. They’d hear a lot of hissing and some voices in the distance. Fortunately, as the winds picked up, the people on the bridge spoke louder, which helped the team make out their words.
Then Doug went back to the beginning of the audio, before dawn on September 30 when Davidson and Shultz were on the bridge. That conversation was easy to transcribe.
The watch change at midnight at the chart table proved very challenging, obscured by the blackout curtain rattling in its metal track.
At times, Doug had to play the same three words over a hundred times before everyone listening could agree on what was being said. He was constantly putting the audio through different filters to get a slightly different sound to help the team confirm what they thought they heard. They spent one day transcribing a total of nine minutes of audio.
In the twenty-six hours of recording, there were hours and hours of silence while the two people on the bridge—officer and helmsman—drove the ship and watched for hazards, steeped in their own thoughts.
It took twenty-two days, working ten hours a day, for the team to transcribe the VDR. The result was a five-hundred-page transcript, longer than any ever produced by the NTSB.
Doug may not have been affected by the audio, but the ex-mariners on the team were. The voices of those lost haunted them and turned a tragic case into a deeply personal one. The dead whispered in their ears long after they took off their headphones.