Wreck diver killed by leaking computer - UK

[rjack321]

As an engineer and an O/C diver this worries me a lot. There is an automatic controller, programmed to keep the PPO2 in the breathing loop within a safe range at all times. Here you are saying it is possible the controller might initiate an action that would drop the PPO2 below the level needed to maintain consciousness. That would go against every principle of engineering critical systems.

How would this be possible? Because the percentage O2 in the dil wasn't correctly programmed into the controller? Even then, it is a closed-loop feedback system where the PPO2 is measured and fed back to the controller, isn't it?

Not how it works:
1) The controller does not maintain a breathable loop even if its functioning 100%. If you dont add enough diluent (descending) the controller won't inject O2 but your ppO2 is going to stay too high. The diver has to add the diluent.


2) The theory is:
-The controller flooded and broke
-He was operating the O2 injection himself and injecting O2 as he ascended
-Sometime during the 6m stop he was distracted or perhaps was fighting buoyancy a little or shooting an smb or whatever
-The ADV injected diluent when the loop volume dropped as the O2 was consumed
-He had hypoxic diluent for a 70m dive so he passed our and drowned.

The ADV is essentially a 2nd stage reg, when the loop volume drops because the O2 is metabolized its going to actuate and add (hypoxic in this case) dil. The controller was disabled so the only reason it didnt actuate deeper was because he was ascending (the loop was expanding) and he was adding O2 to maintain an ever increasing fO2 in the loop.

 

[Jack Hammer]

As an engineer and an O/C diver this worries me a lot. There is an automatic controller, programmed to keep the PPO2 in the breathing loop within a safe range at all times. Here you are saying it is possible the controller might initiate an action that would drop the PPO2 below the level needed to maintain consciousness. That would go against every principle of engineering critical systems.

That's not how it works.

 

[John the Pom]

That's not how it works.

Humans are really very good at performing complicated tasks, but they are very bad at monitoring some automatic system performing the same task, because for example they get distracted and attention wanders (like some recent self-driving car accidents where the human supervisor failed to intervene). It sounds to me like it might be safer to dive with a fully manual CCR.

 

[Angelo Farina]

That's not how it works.

And that's the main reason for which I will never use an electronic-controlled CC rebreather.
The point is that they are not truly electronic-controlled, not in the the way an engineer would design a "fail-safe" system, based on feed forward plus feedback for error control...
I am an engineer and an OC diver, too.
For me, CC only means ARO (pure-oxygen, completely mechanic, 10 meters max).

 

[bletso]

Angelo,
I agree with you except that it wasn't just the failure of the ECCR, but diver error/s in addition. He should have been on OC. I don't agree with the philosophy of staying on the loop as long as possible. Electrons and salt water never mix well.

I wrote a long time ago;
"Well Dave Thompson, original designer of the Inspiration CCR, put it very well. He said; "The bad thing about manual CCR's is you "need" to drive them, and to a degree that's task loading... The good thing is you KNOW that if you don't, you will die. It focuses the mind.

The good thing about electronic CCR's is you don't need to do anything because they drive themselves. The bad thing about electronic CCR's is they MIGHT go wrong, and you MIGHT not notice it because you don't need to do anything right!!!!""

The more bells and whistles you add don't seem to improve a systems reliability but just increase the chances of something not going well and you not noticing or acting on it. It can't possibly be....

 

[BlueTrin]

As a person who does not know about CCRs I find this thread really interesting and learned a lot of information from others here.

 

[EFX]

The good thing about electronic CCR's is you don't need to do anything because they drive themselves. The bad thing about electronic CCR's is they MIGHT go wrong, and you MIGHT not notice it because you don't need to do anything right!!!!""

Arn't there audible or visual alarms when something goes wrong?

 

[grantctobin]

Arn't there audible or visual alarms when something goes wrong?

Depends on the unit. Audible/visual/vibration/some combination/none.

 

[Jack Hammer]

Arn't there audible or visual alarms when something goes wrong?

On the JJ the HUD flashes red repeatedly when the PO2 is too high or too low. This is right in the divers field of view and is visible to team members. Of course, this assumes the HUD is on, properly calibrated, placed so the diver can see it, and not being ignored.

 

[uwxplorer]

How can the JJ HUD be "misplaced"? On the rEvo, the clippable adapter can actually pop off (the HUD then dangling from the loop, invisible) or simply be pushed forward on backward, rendering it useless. Sre we talking about that kind of problem? A system like that coming with the Divesoft BOV solves that problem (<OT>it's actually surprising that nothing has changed with the rEvo BOV as far as HUD attachment, as far as I can tell</OT>).
Sometimes as well, bright light from the surface at 20 ft can result in HUD LEDs and even computers to be hard to read without shielding them from ambient light...