Comments about technological history, system fractures, and human resilience from James R. Chiles, the author of Inviting Disaster: Lessons from the Edge of Technology (HarperBusiness 2001; paperback 2002) and The God Machine: From Boomerangs to Black Hawks, the Story of the Helicopter (Random House, 2007, paperback 2008)

Thursday, March 29, 2012

Deepsea Challenger: Timeline of tests and missions

When following a complex and interesting event, I assemble a timeline. Reporters often use timelines to track incoming facts and come up with new questions.  

For those who like a matrix, here's a draft list of dates and locations that I sifted from the Deepsea Challenge website and independent reports, such as pieces filed by news outlets in Guam and Australia.

If there are errors or omissions, let me know!

Date (Local) Label Location Max depth (feet) Notes
01/26/12 First submersion –    no pilot Dock at Garden Island Naval Depot, Sydney Harbor < 5 Checks for buoyancy and water tightness
01/31/12 First piloted submersion Dock at Garden Island Naval Depot, Sydney Harbor < 5 First time pilot in sub, in water
Probably first half of Feb Open-ocean test Jervis Bay, S of Sydney Not reported Successful piloted tests
02/21/12 First attempt, test dive to 3K feet Jacquinot Bay, 3 mi offshore of Papua, New Guinea 0 Sub launched with pilot, but aborted after 1 hr. Stayed on surface
02/22/12 Second attempt, test dive to 3K feet Jacquinot Bay 0 Scrubbed due to camera-control problems
02/23/12 Test dive to 3K feet Jacquinot Bay 3,250 Successful dive, photography, moved horiz at 3 kt, and rendezvous with lander Mike and also ROV operated from above
02/27/12 Attempted test dive to 12K Jacquinot Bay 0 Scrubbed
02/28/12 Test dive to 12K Jacquinot Bay 12,000 Successful rendezvous with landers Mike and Andrew, maneuvered 5 km horiz with guidance from ship
03/04/12 First Dive, New Britain Trench New Britain Trench 23,818 On way to NB Trench bottom, tried to stop 5 kt descent with large dumps of trim shot, then ascended slowly; then dropped ascent wts
03/07/12 Second Dive, New Britain Trench New Britain Trench 26,791 Successful dive to trench bottom for photography – maneuvered 1.5 km horiz, 300m vert
03/21/12 Unpiloted test dive to Challenger Deep   Not reported “nearly seven miles” tested to “almost 8 tons per sq inch”
03/25/12 First solo trip to deepest point Challenger Deep (probably vicinity of 11.3632, 142.5770) 35,756 Successful dive – first solo dive to bottom of hadal zone; 3D photography; core sample brought up; surfaced at local noon

Wednesday, March 28, 2012

Mermaid Sapphire: Back on the map

Have been visiting the MarineTraffic page to see when the triumphant Deepsea Challenge fleet will rematerialize; I could find nothing last night but this morning, Mermaid Sapphire was again visible, anchored at Ulithi, southwest of the Trench:
The engineering team is based out of Australia, so perhaps they're taking the sub back to deepsea HQ, the innovative workshop in Sydney known as Acheron Project Pty.  

I checked for Barakuda, mother ship of the twin landers Mike and Andrew, but it's not back on the AIS map yet. Since we don't know where it is, I can speculate freely, like so: maybe Barakuda went back to the most productive spot to date, the New Britain Trench, and is using the unmanned platforms to do more photography, attracting animals large and small with bait so as to snap their portraits. 

That would help perk up the scientists on the project, who perhaps were discouraged about the difficulties Deepsea Challenger had with its non-photographic science-gathering gear during evolutions: the manipulator arm, super-slurper, core sampler, and science door. 

The sub and landers are proving to be champs at video and photography ... as long as there's something to take pictures of.

Because so little technical detail has been released -- such as info about the Cameron-less test dive to the Challenger Deep, a few days before the March 25 solo dive -- I'll keep researching and let you know what turns up. Till then.

Sunday, March 25, 2012

Deepsea Challenger: Halfway through the science mission

Update: Paul Allen, whose yacht Octopus is currently keeping station near Mermaid, reported a half-hour ago that James Cameron is on his way back to daylight. So, for whatever reason, the bottom time on this first solo dive will be closer to three hours than the anticipated six hours. If both ascent weights slid out on command, he should be surfacing within 20 minutes.

All systems are go, reports James Cameron from the Challenger Deep. He touched down about three hours ago, which leaves him three more hours of exploration time in the mission plan. A throng of followers are following Facebook and Twitter feeds, accessible through the website.

That's enough time to work with the two landers, if they came down within a couple of miles of his touchdown point. With the caveat that the ships turned off their AIS transponders shortly before the launch, here's the flotilla's last reported position, on a Google Map. It should be pretty close to the center of activity.
This link at Deepsea Challenge is a sobering summary of the risks the team has had to think about, and overcome. Some stresses were well-tested in the lab, such as the effect of pressure on the steel pilot sphere, its viewport, and the point that is probably of most concern, the penetration plate (see posts below). Deepsea Challenger came with a generous margin of safety and the sphere did fine during its unmanned test, so it's hard to see much residual risk, pressure-wise.

Other external hazards can't be dealt with so precisely, such as entanglement. Entanglement seems to pose no risk during this mission. For one thing, the two landers apparently have no tethers to the surface, so that avoids the risk of fouling the sub's thrusters with such lines. 

But if the sub is used to explore more locations in coming years, as is likely, giving a wide berth to any tangling hazard will be a part of every mission plan. An ultra-deepwater wreck -- say, some storm-wracked ship that happened to come down in the Challenger Deep years ago -- would be a tempting, but dangerous, location for investigation by a manned sub.

Meanwhile: so far, so good!

Deepsea Challenger: Now more challenging to follow

The Deepsea Challenge project tweeted five hours ago (#DeepChallenge) that contrary to the  previous "OK to go" note from James Cameron, it didn't launch on Sunday, Guam time, after all. The latest Twitter note:
and the team are still waiting for the right diving conditions. We will share updates as they come in.
That explained ship movements otherwise inexplicable. That is: Mermaid and Barakuda's AIS positions suggested the project had launched the sub and/or landers sometime Sunday, but then AIS feeds indicated that late Sunday night (Guam time) the ships packed up and moved rapidly away, stopping several dozen miles to the east, where Mermaid then went into a pattern suggesting a recovery effort. 

Here's a screen shot of AIS data from Mermaid at that time:

What caught my attention, and seemed to support some kind of recovery, was the zig-zag on the left. When a submersible is coming up as fast as Cameron's does, a mother ship backs off a half-mile or more, so it seemed to suggest a recovery.

But that didn't make sense either. Puzzled followers were trying to figure out how Cameron's sub (chiefly designed for vertical movement) could be launched and then surface many miles away, far beyond the capability of its horizontal thrusters and batteries. 

So where did our Waldos go? Here's a screenshot putting the two relevant locations as arrows on a Google map:
Note that the previous location, and the newer one, are over the Mariana Trench, according to Google's depth profile. That supports current speculations by followers that the project is planning another launch attempt on Monday, Guam time. 

Given the paucity of updates from the project, Facebook fans and bloggers like BoneInItsTeeth were depending on AIS to follow developments, because it was the only real-time information available, other than NOAA weather reports.

About AIS: Drawing data from their GPS receivers, ocean-going vessels have transponders on board that relay their whereabouts through the Automatic Information System (see my post on Concordia for more background on AIS). Typically we get updates every few minutes, showing up as points on a course.

But an hour ago, all four ships in the fleet (three associated with the Deepsea effort, plus Paul Allen's megayacht Octopus, which joined up on Sunday) stopped beaming up their AIS positions. Humph! Maybe they've moved somewhere else since then.

For a string of posts and screenshots on what the AIS data was showing until it was shut off, check out BoneInItsTeeth. From one recent post:
Right now his big support ship, the Mermaid Sapphire, is the better part of 50 miles away from where they were last night.  Interesting, too, is the fact that Paul Allen's superyacht, Octopus, has been keeping pace with the Mermaid Sapphire throughout the transit period.
When AIS data is up again, you can monitor movements at MarineTraffic

Now for Speculation Corner. The bits of information I've seen would be consistent with a scenario like this: The team starts the multi-ship process to launch the mission on Sunday at mid-day, beginning with the landers, since this has to be done well in advance of Cameron's plunge. Barakuda drops a lander, which starts to descend. But then there are problems handling Deepsea Challenger given the sea state, or perhaps some technical problem. The team scrubs Deepsea Challenger's launch late in the day. On Sunday evening, Mermaid relocates to another part of the trench. Once Barakuda has its lander back on board,  a few hours later, it leaves the site to rejoin the others.

A problem with my scenario is that it's hard to see why the team would move 40-odd miles. Maybe new information, just in, indicates the mission goals would be better served in the new spot. Any ideas?

Saturday, March 24, 2012

Deepsea Challenger: Possibly today, Guam time

Update, 10 pm CDT: Cameron sent this note via Twitter about three hours ago: 
 "I only tweet when I have something worth saying. Today is the culmination of a 7 yr project. It's finally dive day. Follow us"
The weather forecast looks better today off Guam than over the next several days, when the trade winds will increase by 5 knots. It's Sunday morning there now. The sea is choppy with superimposed wind waves and long swells, but that's par for this time of year, so I doubt that Cameron will hold out for ideal conditions. If the sub is ready, I'd bet that he's ready too.

So if you're a Deepsea-Cameron buff like I am, stay tuned for news of the first solo dive to the Challenger Deep.

Spelling correction: the ship's physician doing most of the expedition blogs is Joe MacInnis, not McInniss. Sorry, doc!

Possible correction on last post: I'm told by an engineer not on the team that the unmanned trial of Deepsea Challenger would, more likely, have been carried out without any physical link to the surface, whether by tether (for communications) or cable (for lifting). In that case my "unlikely scenario" is the most likely one. 

If that's the case, the team would have sent the sub down relying on its on-board instruments and programmed responses, supervised by an acoustic link to the surface.

On the subject of tethers vs. cables. Long ago, mother ships used armored steel cables to keep tabs on their unmanned rovers. The sturdy ROV known as Jason, for example, depended on a composite cable more than a half-inch in diameter, with three layers of steel for hauling strength. 

WHOI's unmanned Nereus, which touched the Challenger Deep in 2009, dumped the lift cable entirely. It was tied to the surface only by a light fiber as delicate as a fishing line. Nereus carried its own power supply, and took itself up and down the water column.

A growing population of underwater craft are even further along in breaking free of the surface, and people too. These AUVs, short for autonomous underwater vehicles, can rove the sea for hours, days, or even weeks at a time before returning to base. Some AUV's can recharge on their own, at solar-panel platforms in mid-ocean. 

Ultra-efficient designs like underwater gliders are built for long range travels and don't rely on acoustic directions from the surface. Here's a paper on the subject, and a photo of whale-tracker components from Wired:
A more typical mission for AUVs has been ocean-floor mapping and water sampling. New military uses include hunting for sea-mines and enemy submarines. 

Deepsea Challenger's unmanned run: Good idea

I mentioned in previous posts that an unmanned trip of Deepsea Challenger might precede the one with James Cameron aboard. The project announced today that Challenger did make the trip on its own, earlier this week.

If the seas are calm enough to allow a safe launch, the world's first solo manned trip to the Challenger Deep could come this weekend.

Running it to the bottom for an all-up systems check was a good idea, since the team had only been able to verify the integrity of individual components in Penn State's 60-inch pressure chamber at ARL. Testing components can't verify the myriad of connections between them (such as breakers, buses, and logic controllers). These connections have often the source of problems. As a New York sandhog once told me about his job, "None of this is worth dying over."

Though no details on the unmanned trip have been offered, one can speculate that the team dropped the sub at the end of a steel cable, perhaps exercised the mechanical systems for a few hours, and hauled it back up to peek for leaks. 

Less likely: they just heaved the sub in the ocean and waited for the automatic systems to send it up from the bottom. The reason I say that is that it's absolutely critical to slow the speedy descent before hitting the bottom. That's something the pilot normally does about a half hour beforehand, by dumping ballast shot from the trim hopper (see previous posts).

I haven't heard whether this week's dry run pointed to any tuneups, but it's likely. You've heard the term "pushing the envelope" from The Right Stuff; it's perhaps wiser to nudge the envelope, rather than push it. That's what the team has been doing, proceeding in nudges rather than heaves.

The mission will jump off from two ships: Mermaid Sapphire for Cameron's sub, and Barakuda. Barakuda hasn't gotten much attention, as a Google search will tell you. 

There are several vessels under the Barakuda name, but it's probably this support vessel, photo courtesy of MarineTraffic:
Why a second mother ship? One of the important innovations in Deepsea Challenge is the use of separate undersea platforms called landers, which need their own mother ship. 

They'll give a "you are there" feeling to the mission video, unlike anything Trieste was able to do. Think of the difference: Trieste was certainly historic in its achievement, but brought back very little information from the tiny spot it could survey. 

In fact, the apparent lack of interesting stuff in the abyss was one reason that further ultra-deepwater exploration efforts languished for years afterward. Trieste had two men peering out of a tiny porthole for less than half an hour, as they sat in one spot. No pictures were taken. Visibility was bad to horrible, given the narrow field of view, the harsh and short-ranged lighting, distortion from the thick curved Plexiglas, and silt clouds. Trieste was a "flag on the peak" event, with most of its value delivered in the form of technological improvements. 

Deepsea Challenger not only has enormously better imaging power, it will work in tandem with two landers. The mission plan is to drop the two landers ahead of time from Barakuda, a couple of miles from Mermaid's location, and park them there to attract sea animals with bait, so that Cameron can cruise over and sneak up on them later. The landers are able to spend more time on the bottom than Cameron can. 

Here's a pic of one:
From what I gather on the project's technology page, the landers have no tether to Barakuda, so I'm not expecting real-time video. Cameron will be able to trigger some lander functions from his sub, presumably using acoustic signals; I would class them as non-mobile remotely operated vehicles, or ROVs. 

One challenge built into the mission, and one which Cameron discussed in the expedition journal, is finding the things off in the dark. As he wrote in his email to Don Walsh, one time his sub's sonar malfunctioned and he couldn't complete the lander rendezvous.

The landers have the potential to add much value to the science mission. For example, a lander can set out a baited trap, and the trap will go up later on its own, after the separation of a burn wire (aka galvanic timed release -- see my previous posts). 

The landers will float to the surface after releasing their ascent weights, and a set of beacons will lead Barakuda to them for recovery.

One of the landers has a full set of 3D cameras similar to those on the Challenger, and will film the manned sub in action.

So there are a lot of moving parts in this manned mission, but also a lot of money and expertise. Looking down the road, I think a fleet of such unmanned landers would be a boon to oceanography.

Thursday, March 22, 2012

Deepsea Challenger: Waiting or going?

Still waiting to hear what Deepsea Challenger is down to. Given the lack of word on the website's latest news page, friends on its Facebook page are speculating whether Explorer-Pilot James Cameron is on his way down, on his way up with a smile on his face, or just waiting for calmer seas. 

Following my first post, I've been wondering whether they might have gone ahead with an unmanned test dive for a full systems check; that would explain the lack of publicity. But that's just speculation. 

Meanwhile, here's more information on some of the ultra-deepwater tech in use. First, a photo of the Penn State Applied Research Lab's pressure facility where the Deepsea Challenger's pilot sphere underwent its stress test:
The four-ton cap to the pressure cylinder is moved by overhead crane, but the two employees twirl it into place by hand, using a two-man lever arrangement. From the website for ARL's high-pressure lab, here's a photo sequence. 
On the subject of big hunks of metal, here's a photo of the pilot sphere that ARL tested, early in its life, just after forging. 
We're told in this IEEE Spectrum article that it's "traditional steel." Well, maybe as in "traditional for submersibles." I doubt if it's the kind of mild steel used for girders. Higher-strength steel is essential to keep the size and cost of the craft in control. Syntactic foam is pretty expensive. I've seen figures of more than $10,000 for one cubic meter of standard syntactic foam, which would be enough to displace one ton of water.

The personnel cabins on the Russian Mir submersibles, which have to be the most time-tested manned submersibles in the world, use an nickel alloy called maraging steel. Here's the business end of a Mir in action.
Two of them will be busy during the upcoming Titanic season, taking two ticket-holders on each trip.

Maraging steel has been used for decades in submersibles. So it's likely that Deepsea Challenger used a similar nickel alloy for its Cameron-protective sphere. 

Now, a little more about the proprietary syntactic foam used for buoyancy on the craft, called IsoFloat. The Deepsea Challenge page on the sub's technology, including IsoFloat, says that it makes up 70% of the volume of the craft, and has twice the tensile strength of standard syntactic foam.

Standard syntactic foam is not known for its tensile strength (the ability to resist pulling forces). Rather, it's remarkable for its compressive strength, its ability to resist water pressure without crushing. That's due to millions of micro-spheres embedded in a plastic matrix. Here's a photo of the l'il wonders:
Without microspheres (glass, ceramic, or aluminum), there's no compressive strength down deep. Here's what happens to polystyrene foam when immersed to 9,500 feet, from the National Geographic Society's My Wonderful World site:

Here's an article about failure tests carried out on a standard syntactic-foam slab, illustrating how it fails in tension before compression. We're told that IsoFloat foam has twice the tensile strength of standard syntactic foam. 

Why is tensile strength relevant? As I mentioned in the previous post, about the last thing the team wants to see is the craft breaking in two when it's suspended over the ocean. Here's a photo from the project's Facebook page. (The lifting point is at the top, suspended from the red line; the two white straps in the foreground are taglines, to keep the load from swinging.) The lifting bags are also visible.

If that were to happen, the heavy end (the pilot sphere, trim-shot hoppers, and 1,100 pounds of ascent weights) would drop into the water. Even if the ascent weights came free as the power failed, no one would want to find out if there's enough syntactic foam still adhering to keep the pilot sphere afloat. 

So I'd guess the Cameron design team didn't put all their bets on the tensile strength of IsoFloat to hold things together. In this photo from Deepsea Challenger website, notice the white lifting straps that surround the capsule:
Presuming the straps weren't some kind of temporary arrangement during manufacture, it suggests to me that these straps assist in tying the crane's lifting point (at the top) to the sphere and the ascent weights (below). 

That could be done, for example, by gathering the straps into a longitudinal tube and running them the length of the craft, up to an embedded lifting frame.

Use of internal straps would still leave the IsoFloat as the structural beam the website says it is, because it serves as a backbone that unifies the flotation, the heavy iron, batteries, lights, camera arms, and lots of other equipment. But the foam beam wouldn't be the only key structural element. The Cameron team is big on redundancy, so I'd guess that there's structural redundancy also.

Tuesday, March 20, 2012

James Cameron Plunges Off the Deep End

Watching the Deepsea Challenge project … James Cameron of movie-making fame intends to be the first man to return to Earth's deepest waters since the single two-man mission in 1960 by Trieste. Cameron's solo attempt could be any day now. He'll be riding the 13-ton submersible Deepsea Challenger to 36,000 feet below sea level. The goal now is plumb the Challenger Deep, but other deepwater destinations are likely to follow.

Here's a map of the big ditch, from Wiki page; it's a product of tectonic subduction.
The mother ship, Mermaid Sapphire, is in transit to the launch location, having left Guam. That follows a long delay for the weather to moderate. Technicians and engineers used the time to tune up the craft given lessons from previous dives. The bright-green submersible rides on the deck of Mermaid inside a climate-controlled hangar, bolted a rigid frame.

Some info and thoughts, in FAQ format:

Q. How much has the submersible been tested beforehand?

A. The pilot sphere was tested in a Penn State pressure chamber to a crushing force well in excess of the ocean depth of 36,000 feet, so that's pretty dependable. A point of particular interest is the penetration plate, where control and instrument cables must pass through the barrier. Other key components were also pressure tested individually.

The operational craft has not been so tested (the Penn State chamber wasn't big enough), but Cameron has safely taken it down to nearly 27,000 feet. Not surprisingly, some glitches needed attention. Some were minor and had no direct connection to the pilot's safety (difficulties with sampling, and closing a hatchway to store samples). Some were peripherally related to safety (problems with the cameras, one of the 12 thrusters going off line). 

A few of those gremlins had safety implications: an uncommanded dump of trim shot (due to low voltage), uncommanded intermittent operation of thrusters (following a change in programming), and the tripping of an electrical bus that cut off voice communications (possible seawater leakage into electronics). Some other problems were detected during inspections, prior to launch. All concerns have been addressed, we are told.

I haven't seen firm plans about an unmanned test to full depth first, but it's a possibility. Why? Testing and trials so far don't establish how it will perform as an all-up system at depth.

For those interested in technical stuff, there's much of interest in the expedition journal by Dr. Joe McInniss. As he writes: 
 “If something goes wrong, we have to fix it, and fix it quick. There’s no such thing as waiting for the rescue team. We deal with it, or else. And all of these 'or else’s' are not very appealing.”
The journal also includes an email that Cameron wrote to Don Walsh, narrating this month's successful plunge into the New Britain Trench (a record-setter for a solo mission).

Q. How does Deepsea Challenger differ from the bathyscaphe Trieste, the only vessel ever to make a successful dive to the bottom of the Mariana Trench?

A. The new sub's flotation is much more compact, it's got a great deal more equipment for science-gathering, and it's more maneuverable. And its pressure sphere is barely big enough for one man. But engineering is always about trade-offs. For one thing, the small size allows the mother ship to raise it via crane and carry it on board. 

Here's a size comparison, from Gizmag:
Trieste had to be towed to the site, which sometimes caused damage ... as happened before the 1960 expedition.

Here's a diagram of the new sub from the Deepsea Challenge website. What might be called the bow of the craft is on the left side.
Q. What do we know about plans for the manned deep dive?

A. Here's what I gather from the website: The sub starts out in horizontal fashion in its cradle on deck, braced by four steel arms, Cameron climbs in and the hatch is closed. After he finishes the checklist, a crane lifts the craft off the deck and puts it in the water. The craft is now oriented vertically with the pilot's sphere at the bottom end. 

With the aid of inflatable lift bags, it floats upright on the surface for a systems check. On command from Cameron, the divers release the bags and the sub drops at a speed of about 5 mph for about 90 minutes. Cameron slows its descent over the next half hour by dropping steel “trim shot” from a hopper. His goal is to arrive at the bottom with neutral buoyancy … almost. He'll probably aim to keep a little positive buoyancy, that is, a tendency to float upward, so that the craft's optimal depth can be maintained by pointing the six vertical thrusters to push downward, ever so slightly. That avoids stirring up silt, which blocks vision until it settles.

Here's the control panel:
He'll use the six horizontal thrusters to rotate the craft in the desired direction and then go forward. For visibility he depends on the bank of LED lights and the underwater 3D video cameras, rather than the viewport.

The craft is pretty maneuverable, according to the expedition journal. Cameron might roam a couple of miles from his starting point, over the next five or six hours. His trip into the New Britain Trench had him snapping photos of sea creatures, while maneuvering around cliffs he compared to the Grand Canyon.

When Cameron is ready to head up, he'll alert the Mermaid and jettison the two steel ascent weights, which are stowed in slots at the bottom. He has a switch to do this, but there are several backup systems that can dump the ballast as well. One is a corrodable "galvanic timed release," now used in the fishing industry to avoid killing fish in ghost traps. If Cameron were rendered unconscious, the timed release would drop the ascent weights within 13 hours.

Why such redundancy? The mission is a free dive; the sub has no tether or cable that the Mermaid could use to haul it back to the surface. So once stabilized on the bottom, the sub can't come back up until it drops a half-ton of ballast. Nor is it an option to somehow separate the pilot's capsule from the rest of the craft and have it float free. The pressure capsule is so heavy with thick-walled steel to resist the 16,500-psi pressure that it won't float by itself.

The buoyancy that the pilot depends on to get him to the surface is syntactic foam, in this case a patented form of it called Iso-Float, that serves as the green-painted upper body of the craft. (Trieste didn't use it; it depended on thousands of gallons of gasoline for its buoyancy, held in big tanks above the pressure capsule.)

I don't know the unique details of Iso-Float, but in general, syntactic foam is a composite of very small glass or ceramic spheres, bonded into a shape with tough resin. It's also very expensive, per cubic foot. Iso-Float is said to be an improvement on conventional syntactic foam because it's strong enough to serve as girder that connects top to bottom. Tensile strength is no small matter when raising the craft by crane, given that the bottom of the craft is heavy with ballast and the pilot sphere, but the crane lifts from the top. To have the craft break in half at the surface, letting the lower half fall free, could be catastrophic. I haven't seen the safety margin but presume that it's substantial.

Okay, now it's time to go up. Dropping the two 550-lb ascent weights gives the craft enough positive buoyancy to get to the surface in about an hour, depending on how much trim shot Cameron retains in the ballast hopper. It makes a spectacular arrival on the surface; Cameron calls this moment splash-up, as opposed to the space program's splash-down.) 

Good communication with the mother ship avoids a collision with the hull. If standard communications go out, say, due to a power failure, the mother ship will be able to track the craft's rise passively.

Q. What was Cameron referring to during interviews about the mission risks, when he said that two men had once died in a submersible accident?

A. Cameron most likely was referring to the deaths of two men in Johnson Sea Link off Florida in June 1973. At a depth was 360 feet, JSL fouled a cable while at the wreck of Fred T. Berry, a Navy destroyer sunk as an artificial reef. A rescue team pulled the craft up but by that time the two men in the diver's compartment had died from carbon dioxide poisoning. The other two men survived because they were in a separate compartment forward, with different equipment. Here's the NTSB report.

Perhaps reassuring to Deepsea Challenger's crew is that a common cause of submersible mishaps and near-misses has been entanglement, and we can presume there won't be tangling hazards in the Mariana Trench. 

But past lessons are worth reading, and anyone planning to build or board a submersible today would be well advised to read section II of this 1974 safety panel report on submersibles, which summarizes submersible incidents and near-misses during the boom years.

A flip through that report and other articles indicates that one risk is seawater intrusion -- conductive water getting into powered gear such as an electrical bus, sensor, or battery pack. 

Clearly Deepsea Challenger's 67 lithium-ion battery packs are thoroughly engineered for depth -- filled with oil to resist pressure, and having compensation bladders to allow for fluid compression. I asked Robert Wernli Sr. what he thought about the pressure-compensated approach. (Wernli is a long-time submersible engineer, co-author of the ROV Manual, and author of the submersible techno-thriller Second Sunrise. I interviewed him two years ago for my article on the history of ROVs.) Says Wernli:
Lithium ion batteries are common in undersea vehicles today. The ROV/AUV Nereus, that visited the bottom of the Mariana Trench last year, had lithium ion batteries, although they were housed in ceramic pressure housings at one atmosphere. The pressure compensated (PC) housings for the batteries are common in undersea vehicles to save weight by eliminating pressure housings. With proper design, PC systems will survive extreme pressures.
The Cameron deep-dive team knows all about batteries. Among its many experiences was a close-call event in 2001, while descending in a pair of Mir submersibles to the Titanic wreck. One of the two Mirs was carrying a small ROV in a basket when a battery in the ROV violently overheated, belching out superheated bubbles inches from the Mir and raising concern among the three occupants that thermal stress might crack the Mir's viewport. Fortunately they headed for the surface and managed the situation without having to dump the ROV, and the danger passed.  

Fire is also a concern in tight quarters that are jammed with electronics, but there are extinguishers and a closed-circuit breathing system to cope with that contingency.

And means to cope with many others. Given that Deepsea Challenger has multiple and independent methods to jettison its ascent weights, pilot Cameron should be able to get back to sea level in good shape, toting lots of interesting data.

So from all of us, to one of you, bon voyage