+ Reply to Thread
Page 3 of 5 FirstFirst 1 2 3 4 5 LastLast
Results 21 to 30 of 43

Thread: Serious incident (hyperoxia) when 2 cells fail simulatenously and identically

  1. #21
    RBW Member Joe is a glorious beacon of light Joe is a glorious beacon of light Joe is a glorious beacon of light Joe is a glorious beacon of light Joe is a glorious beacon of light Joe is a glorious beacon of light Joe is a glorious beacon of light Joe is a glorious beacon of light Joe is a glorious beacon of light Joe is a glorious beacon of light Joe is a glorious beacon of light Joe's Avatar
    Join Date
    Oct 2005
    Location
    Florida, USA
    Posts
    485
    Megaladon

    Re: Serious incident (hyperoxia) when 2 cells fail simulatenously and identically

    Correction of linear to "LIMITED"

    Quote Originally Posted by Joe  View Original Post
    That is exactly why I dive my unit manually and use the electronics as a fail safe. I also keep my setpoint at 1.0 because that is what I have calibrated to so the possibility of cells being (current) limited within that range is reduced.

  2. #22
    RBW Member EAC is on a distinguished road EAC is on a distinguished road EAC's Avatar
    Join Date
    Mar 2011
    Location
    Ft Lauderdale vicinity
    Posts
    90
    optima

    Inspiration

    Re: Serious incident (hyperoxia) when 2 cells fail simulatenously and identically

    Quote Originally Posted by tmccar1  View Original Post
    Scary on so many levels.

    This is the Open Circuit equivelent of saying, "I reached bottom, manifold exploded and raging torrents of gas continued to escape over my shoulder. Luckily the explosion didn't kill me so I shrugged it off. I then saw a turtle and decided to chase it."
    Agreed- Abort Abort Abort-to dive on them again?
    Now we see too that all manufacturer's of cells have their issues at times! It is just a tough environment
    -

  3. #23
    Explorers Exodus Hanssing has a brilliant future Hanssing has a brilliant future Hanssing has a brilliant future Hanssing has a brilliant future Hanssing has a brilliant future Hanssing has a brilliant future Hanssing has a brilliant future Hanssing has a brilliant future Hanssing has a brilliant future Hanssing has a brilliant future Hanssing has a brilliant future Hanssing's Avatar
    Join Date
    Nov 2005
    Location
    Denmark
    Posts
    912
    rEvo-II

    Re: Serious incident (hyperoxia) when 2 cells fail simulatenously and identically

    First off, let me start by saying I am glad your buddy is fine. From the report, it sounds as if he was seconds away from dying!

    My comments on what should done by divers and by manufactures, to avoid killing more divers:

    * Dont change all cells at one. This is bad practice in general. Change cells by rotationg them. And bying them from different batches. This minimizes batch-failures correlations (as this one), and it also minimizes failurs-correllations due to dive-history. If you want to change 3 cells a year, change one cell every 4 months.

    * O2-algorithm should inject according to highest PPO, and make deco according to lowest PPO. As long as all cells above life-supporting range - ie. above 0.2.
    It takes longer to decay the loop below 0.16, than it takes to inject it beyond 3PPO. And the risk for double cell-failure is too high.
    This has been described extensively by Andrew something. I know it adds deco, but your machine is running on reduced cell-count, and you no longer have redundancy. Get out of the water, and do the extra few minutes.

    * O2 at depths greater than 20'ish meters is a deadly gas! However eCCRs like JJ, Meg and Inspo have the solenoid adding a much high mass-flow at depth, and thereby very quickly exceeds lifesustaining PPO-levels when something goes wrong in open state. The CE requirements forces a very high minimum flow near the surface to ensure a blow-up has breathable gas, and this is way to high when diving deep with pure O2, in my opinion.
    If I were to dive a pure eCCR to deep depths, I would severely limit the solenoid O2 flow to simething like 2-3x my O2-consumption.
    Besides this I would also cap the regulator, unless I was going below 80msw. Take a look at the Birgitte Lenoir thread, who died at what was speculated to be over-dosing of O2.
    Since a solenoid is just a likely to fail open as it is to fail closed, then I dont think this is the only death at greath depth caused by this, but thats merely my own speculative oppinion.

    I personally run my mCCR with 5 cells, rotationg one every 6 months. Been doing it that way since 2007. Never had more then one cell go out on me, and I am doing it this way to avoid common failure between cells.

    Hanssing

    PS: Its been a shitty year for CCR-deaths. Glad your buddy was not one more to the list.

  4. #24
    RBW Member Danyboy will become famous soon enough Danyboy will become famous soon enough Danyboy will become famous soon enough Danyboy's Avatar
    Join Date
    Mar 2011
    Location
    Switzerland
    Posts
    288
    Poseidon Discovery

    Re: Serious incident (hyperoxia) when 2 cells fail simulatenously and identically

    Quote Originally Posted by jaksel  View Original Post
    Or, one could just accept that the controller will give a warning also in those cases when O2 injection rate changes due to "natural causes" (not because cells are current-limited). After all, if we're at constant depth and not doing random dil flushes, we would expect a constant O2 injection rate.
    ...
    A properly designed eCCR could, (just as your finger does with an mCCR, tell you the approximate O2 injection rate.
    Hello Jaksel!

    Also from my side: kudos!
    I've had my two moments underwater too, and I'm glad for everyone to be able to tell the tale...!

    I would also like to thank all the others that outed themselfs about the sensor-reliability...!
    Some people talk like such a thing never happens - these informations paint now a whole different picture! Thank you!

    At first, I thought that the monitoring of O2-use is a quite interesting idea...
    After all, my good-old-fashioned AladinAir from the early 90's already had a detection that gives an audible alarm if the tank pressure drops too fast - which is assumed for a too high breathing rate.
    Easy, just implement the same for oxygen and you're done...
    Unfortunately, the answer would be: Nope, you won't...

    Problem in my opinion is the variety of users and the way one single specific unit is being used. Even if the thresholds for the alarm could be set to the user's gusto, you would have alarms when you wouldn't want it.
    In example whenever you would do a dil-flush, as the unit would try to establish your setpoint by injecting oxygen.
    The envelope in which such a system would work properly is that much restricting, that it becomes impractical.
    And that is just the scenario for when the unit is used as a real eccr... If the eccr is being used as mccr, the electronics as backup with setpoint 0.4, then the scenario becomes even more complicated, as there's not just the unit injecting oxygen, but also the diver by manual-add.

    And when I then look into the disco-corner, how some people already have huge problems to understand the difference of a layout with one controller as opposed to a network of four processors, it's pro's and con's... Uh boy, you better not implement such a feature in a 'breather... ;)

    And I think that such a feature does not really solve the problem in the first place, it only patches the results. I strongly believe that only an automated verification of at least one of the three sensors is the solution.

    Cheers,
    Dany

  5. #25
    RBW Member PeterVICEG is a glorious beacon of light PeterVICEG is a glorious beacon of light PeterVICEG is a glorious beacon of light PeterVICEG is a glorious beacon of light PeterVICEG is a glorious beacon of light PeterVICEG is a glorious beacon of light PeterVICEG is a glorious beacon of light PeterVICEG is a glorious beacon of light PeterVICEG is a glorious beacon of light PeterVICEG is a glorious beacon of light PeterVICEG is a glorious beacon of light PeterVICEG's Avatar
    Join Date
    Jan 2010
    Location
    Vancouver BC Canada
    Posts
    440
    Meg

    Fenzy PO68 Dolphin

    Re: Serious incident (hyperoxia) when 2 cells fail simulatenously and identically

    Quote Originally Posted by Danyboy  View Original Post
    Hello Jaksel!


    I strongly believe that only an automated verification of at least one of the three sensors is the solution.

    Cheers,
    Dany
    I couldn't agree less...

    First off I am most greatfull to to OP for posting this incident.

    As I see it the diver failed to do or understand a dil flush. Had he/she done so properly the real problem probably would have been clear.

    Others have also pointed out other errors that were made and I have no interest in repeating them.

    My point is simply that if one is to play this game, diving beyond recreational depths, one needs to understand what cell behavior really means.

    I am not interested in the my re breather is better than yours game. I am happy, however, that my unit allows me to view the raw cell Mv output at any time. All and any electronic gizmos can and will fail eventualy. Especially in a marine environment.

    Any kind of significant depth or overhead calls for a full bag of skills. The ability to anylize cell behaviour is principle among them.

    Peter

  6. #26
    RBW Member Lancer4545 is a splendid one to behold Lancer4545 is a splendid one to behold Lancer4545 is a splendid one to behold Lancer4545 is a splendid one to behold Lancer4545 is a splendid one to behold Lancer4545 is a splendid one to behold Lancer4545 is a splendid one to behold Lancer4545 is a splendid one to behold Lancer4545 is a splendid one to behold Lancer4545 is a splendid one to behold Lancer4545 is a splendid one to behold Lancer4545's Avatar
    Join Date
    Jun 2007
    Location
    Cairns QLD Australia
    Posts
    1,974
    Meg/Shearwater Pathfinder

    Re: Serious incident (hyperoxia) when 2 cells fail simulatenously and identically

    Rebreather manufacturers invest countless hours testing and modifying their units to make them as reliable as they possibly can and then fit these Galvanic Sensors into the heart of it.

    No matter what unit you dive you'll be using Galvanic Sensors to monitor the PO2 whether its a mCCR, eCCR or whatever. The fact remains we are entrusting our lives on these totally unreliable sensors to give us a PO2 reading for the loop gas.

    Never ever take what they are showing to be true and correct unless you verify and test them during every dive. I've seen so many sensors start off reading reliably only to fall over during the dive and they can go either way showing higher or lower than the actual loop PO2

    Unfortunately until something better comes along we're stuck with them so be careful and always test and verify their output on every dive!!

    Regards,

    Lance

  7. #27
    Jakob Selbing jaksel will become famous soon enough jaksel will become famous soon enough jaksel will become famous soon enough jaksel's Avatar
    Join Date
    Oct 2008
    Location
    Stockholm
    Posts
    178
    AP Evolution+

    Re: Serious incident (hyperoxia) when 2 cells fail simulatenously and identically

    Quote Originally Posted by Danyboy  View Original Post
    In example whenever you would do a dil-flush, as the unit would try to establish your setpoint by injecting oxygen.
    The envelope in which such a system would work properly is that much restricting, that it becomes impractical.
    I don't quite agree. Yes, you would get alarms when e.g. doing diluent flushes, but I don't see why that would be such a big problem. For example, on the Inspo/Evo, changing to high setpoint or flushing with diluent triggers cell warnings quite often. As a user, I notice the warning and understand why it occurs. Usually the condition resolves within a few seconds. Further, diluent flushing will also cause the HUD to blink green, warning the diver that the measured ppO2 deviates from the setpoint by more than 0.2 (I think). Why not use the same type of warning when the injection rate drifts out of a reasonable range? The diver then decides if the warning is expected (dil flush, ascending, wokring hard, etc) or if there's something wrong with the CCR.
    If the eccr is being used as mccr, the electronics as backup with setpoint 0.4, then the scenario becomes even more complicated, as there's not just the unit injecting oxygen, but also the diver by manual-add.
    Yes, if the eCCR is used in such a way, the expected O2 injection rate would instead be close to 0. A simple controller configuration choice (eCCR or mCCR/hCCR mode) would solve this.
    And I think that such a feature does not really solve the problem in the first place, it only patches the results.
    It does not solve the problem, but at least it gives you a hint that there might be something wrokng.

    I am just saying is that the setpoint controller could be a little more intelligent, possibly catching double cell failures like this one (which could just as well have resulted in a fatality), or other types of problems like ADV triggering due to inadvertent loop gas exhaustion (hasn't there been a couple of such accidents?).

    Why not compare this "intelligent" controller with the diver's brain when diving an mCCR? A reasonably smart diver wouldn't happily keep injecting O2 like crazy at depth even if all 3 cells tells him that the ppO2 is lower than the setpoint. He would rather understand that something is wrong; either the cells have failed or the O2 injection has failed (e.g. blocked MAV or O2 tank is depleted). Either way, he/she could do a diluent flush, check the cells, and take a minute to work out appropriate action.

  8. #28
    Capt/ Instructor
    tmccar1 is a jewel in the rough tmccar1 is a jewel in the rough tmccar1 is a jewel in the rough tmccar1 is a jewel in the rough tmccar1 is a jewel in the rough tmccar1 is a jewel in the rough tmccar1 is a jewel in the rough tmccar1's Avatar
    Join Date
    Mar 2006
    Location
    Long Island, NY
    Posts
    406
    SF2 | rEvo

    PRISM Topaz

    Re: Serious incident (hyperoxia) when 2 cells fail simulatenously and identically

    Don't get me wrong, I agree that it CAN be done but why add layers upon layers of convoluted flow meters, volume sensors, and computers to arrive at an educated guess that tells you "is the solenoid firing too often?"

    I don't really see it being worth while except for in your friends case where audible alarms were masking the sound of it opening. (another issue in itself) When I dove ECCR, I could hear the solenoid firing and once you had some experience you knew if it was firing too often.

    I think an alarm that sounds every time you add dil, change depth, clear a mask, etc... would become the boy that cried wolf and be ignored after a while.

    Additionaly, the range of .5/l min to 5/l min is a huge margin in a system like this, you'd need equipment to measure VO2 in order to be accurate, so now you're talking about heart rate monitors and/or O2 sensors in or around the mouthpiece measuring expired FO2.

    I think the best routes are more reliable O2 cells (or alternative methods of gas analysis) and diluent (known gas) based cell validation (whether that means you flushing the loop or the Poseidon automatic validation system).

    Once cells start acting wacky the easiest way to tell which one is working for sure is just a dil flush.


    Quote Originally Posted by jaksel  View Original Post
    Don't most dive computers monitor depth?

    I don't see why a warning given due to a clogged/failing solenoid would be bad?

    Further, diluent addition (or manual O2 addition) will change the ppO2 -- but not gas exhaustion itself.

    The metabolism changes of course, but the upper and lower limit would be fairly easy to figure out. At least we know that the rate is larger than 0.5 L/min and probably less than 5 L/min or so.

    Figuring out how the flow rate changes while descending and ascending is not too difficult either (only a few simple equations to solve).

    Or, one could just accept that the controller will give a warning also in those cases when O2 injection rate changes due to "natural causes" (not because cells are current-limited). After all, if we're at constant depth and not doing random dil flushes, we would expect a constant O2 injection rate.


    That is definitely an advantage for mCCRs, but I am talking about eCCR now. A properly designed eCCR could, (just as your finger does with an mCCR, tell you the approximate O2 injection rate.
    Last edited by tmccar1; 11th December 2012 at 14:21.

  9. #29
    RBW Member Paul S is on a distinguished road Paul S is on a distinguished road Paul S's Avatar
    Join Date
    Sep 2008
    Location
    UK, Twickenham
    Posts
    147
    rEvo III Mini hCCR

    Quote Originally Posted by tmccar1  View Original Post
    Once cells start acting wacky the easiest way to tell which one is working for sure is just a dil flush.
    I think you're overstating the diagnostic usefulness of a dil flush.

    You're at 30 meters (4 ATM) and have air diluent. Your cells are reading 2.10 - 1.20 - 1.25, so you do a dil flush. They all drop to 0.89 - 0.80 - 0.85. We're now on known gas and the readings look about right for it, but its not clear what the problem is (at least to me). Is cell 1 running high or are 2&3 both limited? "It can't be 2 AND 3 surely, they've been so close to each other the whole dive. "

    Unless you hear the solenoid it's easy to make the decision that cell 1 is bust and acting odd, and if we're full eCCR then we're going back hyperoxic very soon. I'm thinking the smart choice, if your set on staying on the loop, might be to bring the set-point into a range we can verify, which in this case would be about 0.85. We now know the cells work at that level.

  10. #30
    Explorers Exodus Hanssing has a brilliant future Hanssing has a brilliant future Hanssing has a brilliant future Hanssing has a brilliant future Hanssing has a brilliant future Hanssing has a brilliant future Hanssing has a brilliant future Hanssing has a brilliant future Hanssing has a brilliant future Hanssing has a brilliant future Hanssing has a brilliant future Hanssing's Avatar
    Join Date
    Nov 2005
    Location
    Denmark
    Posts
    912
    rEvo-II

    Re: Serious incident (hyperoxia) when 2 cells fail simulatenously and identically

    Quote Originally Posted by Paul S  View Original Post
    I think you're overstating the diagnostic usefulness of a dil flush.

    You're at 30 meters (4 ATM) and have air diluent. Your cells are reading 2.10 - 1.20 - 1.25, so you do a dil flush. They all drop to 0.89 - 0.80 - 0.85. We're now on known gas and the readings look about right for it, but its not clear what the problem is (at least to me). Is cell 1 running high or are 2&3 both limited? "It can't be 2 AND 3 surely, they've been so close to each other the whole dive. "

    Unless you hear the solenoid it's easy to make the decision that cell 1 is bust and acting odd, and if we're full eCCR then we're going back hyperoxic very soon. I'm thinking the smart choice, if your set on staying on the loop, might be to bring the set-point into a range we can verify, which in this case would be about 0.85. We now know the cells work at that level.
    Yup, you don't know, which is right, for sure. Run all cells within 0.2-1.6 bars (ie. higest cell at 1.6 maximum). Do the deco according to the lowest, and get out of the water. Your CCR is ****ed up, go home, live another day.

+ Reply to Thread
Page 3 of 5 FirstFirst 1 2 3 4 5 LastLast

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts