https://twitter.com/LadyDoctorSays/status/167170098942929715...
The issue as far as I have read is that the hull was made of carbon fibre. There hasn't been any submersible that has reached those depths before made of that material. The effect continued pressurization/depressurization had on the carbon fibre wasn't understood. Composite materials are so much more complicated to model and understand. There was no non-destructive testing to see what effect the repeated cycles had on the hull, no way of knowing whether cracks could form beneath the surface. The failure mode at depth is catastrophic, there's no room for error. Someone pointed all this out to them and was fired https://newrepublic.com/post/173802/missing-titanic-sub-face....
https://edition.cnn.com/2023/06/21/us/titan-sub-safety-ocean...
In response to them knowing the sub wasn't fit for purpose they opted to install a "real time health monitoring system" which acoustically checked the integrity of the hull. But it's pointless. By the time any acoustic monitoring system picked something up it would be too late, because carbon fibre just shatters into a million pieces. It's not like Steel where it can gradually fatigue, it's crack BOOM dead.
Using carbon fibre for the hull is like rolling your own crypto. Maybe you can get it to work but unless you properly scrutinize it there is most likely fundamental flaws in your implementation and it's just better to use tried and true methods. In the sub world that tried and true method is just thick steel.
I wonder if this event will cause some changes, or if it is an expensive step in figuring out how to properly test this material.
No. Not really. It's a horrible idea. It's irresponsible. It's negligent.
Consumer grade components are not designed for the kind of reliability and failure tolerance requirements these kinds of applications have to face every mission. A simple example of this would be no conformal coating anywhere to prevent failures due to condensation on the circuit board. Also, tin whiskers due to RoHS solder. No redundancy or failure tolerance in any portion of the design. It isn't enough to say "we have extra game controllers". That's not how you build reliable failure-tolerant systems where people's lives depend on things working perfectly 99.9999% of the time.
Also, it is a fallacy to say that millions of people use these controllers. The most obvious problem with this statement is that we have zero data on failure rates and failure modes. And, of course, nobody dies. As a parallel example, you would be hard-pressed to find toy controllers on medical equipment, even in situations where there might be time to deal with failures.
> the hull was made of carbon fibre <snip> In the sub world that tried and true method is just thick steel.
Yup. Fundamental principle: Carbon fiber is best used under tension, not compression. Not a good choice for a life-support compartment under incredible external pressure. One way I think of it is: In compression you are mostly relying on the epoxy to keep things together. Without epoxy carbon fiber will support as much compression as a common string. Zero.
The carbon fiber seems to be the critical nail in the coffin.
That's what it sounds like.
> The carbon fiber seems to be the critical nail in the coffin.
If the CF vessel failed it would have instantaneous and violent beyond comprehension. What little reliable info is out there seems to indicate that's what happened.
A lot of this is bewilderment at the choice of carbon fibre for the pressure vessel, which is sensitive to impact damage and wear from repeated load cycles, damage is hard to diagnose, and as it's very brittle, prone to catastrophic failure. It's not commonly used for this sort of application, and there may not be good data on (1) when it would eventually fail and (2) whether you'd be able to tell before use.
As I understand, the choice was motivated by wanting the sub to have the large interior space necessary to bring along that many passengers. Deepsea subs usually don't attempt that either.
All more or less untested and uncertified. Throw in the reported comms issues that were common during past operations, a highly inexperienced engineering team, a culture promoting unsafe practices and you get a death trap of a vessel.
I think there was a fair bit of other concerns. I thought this video was informative and the kind of thing the HN crowd would appreciate, so take a look.
These design flaws weren't an unfortunate mistake, they were part of a very deliberate pattern that went predictably wrong.
Some highlights include:
* Design inspired by the "DeepFlight Challenger" - but DeepFlight said their craft was only rated for one dive, and weakens with each cycle, and could not be used for five dives.
* Hull designed by subcontract manufacturer in 6 weeks - scarcely enough time to do much testing.
* In an early dive the CEO performed solo, lost contact with the surface ship for approximately one hour
* While diving with a journalist, lost contact with the support vessel for 5 hours.
* It was impossible to open the sub from inside (admittedly, this is only relevant if you first manage to ascend to the surface)
* Hull started showing showing signs of cyclic fatigue in January 2020 (they got it repaired)
* Assembly and testing procedures so sloppy they managed to attempt a dive with a thruster installed the wrong way around.
* Employee called for a stronger front window, and nondestructive testing of the hull. Company fired him and sued him.
* CEO on record as saying ship safety laws "needlessly prioritized passenger safety over commercial innovation" and that "At some point, safety just is pure waste. I mean, if you just want to be safe, don't get out of bed."
* Couldn't get certified by a ship classification society, claimed that was OK because most marine accidents are operator error not mechanical failure; and the standards didn't give them adequate credit for their corporate culture of safety.
The other thing that indicates it was a deathtrap is the deaths.
With that said, personally I support the right of people to expose themselves to the risk of death in search of adventure. Normal folks can buy motorbikes and quadbikes, millionaires can buy cessnas, why shouldn't billionaires have deadly entertainment options befitting their wealth?
I watched one video that complained about off the shelf parts being used, but the two examples was an RV light (not safety related) and the gaming controller (which they had multiple back ups).
They also claimed to have been reviewed by Boeing and University of Washington. There was 7 different mechanisms that could force a return. Some of those were purely mechanical.
Clearly the sub wasn’t safe enough, but I’m not seeing anything that makes it obviously badly built.