Prebreathing Survey

[stuartv]

On both of my units I was advised of the scrubber life by instructors (and in the case of my Meg again by the actual builder on a trip we did together)

Those would probably into my personal category of people to whom I attribute some credibility. Pushing limits doesn't have to mean pushing BEYOND what those people say. But, it definitely might mean pushing beyond the manufacturer's published test data (e.g. 3 hours). I'm not recommending that people push and push until they get a CO2 hit from breakthrough. For that matter, I am not RECOMMENDING to push past the OEM's spec. I AM saying that people SHOULD be prepared to deal with a breakthrough - even if they are within the 3 hour limit (or whatever limit their manufacturer publishes), for that matter. And especially if they CHOOSE to expand their limits beyond whatever limit they have started with.

 

[doctormike]

... alternatively looking at the body of evidence that a very typical 5.5 to 6 lbs of sorb in warmer water (sometimes 20+C warmer), with a much lower exertion rate, can last far longer than what the manufacturer's lawyers will allow them to print...

That's all I'm asking for. The body of evidence. You don't need to do CE testing to get numbers. But if you are going to categorically state that a 2.4 kg scrubber will last for a minimum of 6 hours in 25C water, I personally would like something more reproducible than "many excellent divers do that, and they are fine". Because until you actually do a real world prospective diver study, you just don't have that data.

I'm not saying that such studies don't exist. I'm just asking for them. And if they don't exist, then my point stands.

 

[doctormike]

How? You are diligent in sticking to 3 hours. When you have another 1000 hours on your unit, if you continue to be diligent about 3 hours, how would you ever know that you can do more than 3 hours?

I wouldn't. That's the point.

It seems to me that the only way people ever know these things is by first pushing the limits into unknown territory. You won't know if you can dive more than 3 hours until you've tried it. Multiple times.

Or, you can take a preponderance of anecdotal evidence, from very experienced people to whom you attribute some credibility, and begin to push your own limits where it's not 100% "unknown". Just be prepared to FIND those limits....

It seems to me very much like figuring out what works for you when it comes to NDLs, Gradient Factors, etc.. You start with a published "limit" (e.g. PADI tables, or Shearwater Conservatism on High, or whatever), get experience with it, learn what you can from other people's experience, and then gradually expand your own body of experience to figure out what works for YOU.

Are you saying keep pushing the scrubber until you get a CO2 hit? I'm not going to do that.

Also, the comparison to DCS isn't really accurate, for a number of reasons.

1) There is no machine that can model decompression sickness

2) This testing is measuring a simple, easily detectable output variable - CO2 levels. DCS is an incredibly complex and incompletely understood process. You could never generate a graph like figures 3-8 and 3-9 for DCS

3) Safe ranges for N2 loading and decompression were developed over many years by observing the rates of clinical DCS in huge data sets. And the GF ranges are also amenable to testing, since you actually are specifying your input variables. You can see what the DCS rate is for a cohort of divers doing 30/70 when compared to 90/90 or whatever. Unless you are going to set up your experiment to specify CO2 production models in real world diving using a standard nomenclature, it's just guesswork. Two divers in the Caribbean are diving Axial JJs. One runs it for 4 hours, one for 8 hours. Are they both in the same study group?

4) DCS rates can be determined by looking for clinical DCS in volunteer divers. With an on site chamber, that at least addresses the ethical considerations of potentially bending one of your subjects. How would you replicate that for a CO2 hit? In addition to CO2 production variability, there is a variability in the clinical response to a given CO2 level.

 

[rjack321]

That's all I'm asking for. The body of evidence. You don't need to do CE testing to get numbers. But if you are going to categorically state that a 2.4 kg scrubber will last for a minimum of 6 hours in 25C water, I personally would like something more reproducible than "many excellent divers do that, and they are fine". Because until you actually do a real world prospective diver study, you just don't have that data.

I'm not saying that such studies don't exist. I'm just asking for them. And if they don't exist, then my point stands.

They don't exist in the form you are expecting. We are left to extrapolate. But how much Co2 sorb can ideally capture isn't a mystery (for sofnolime 797 its about 150L Co2/kg sorb). So we have bookends.

John's blog John Clarke Online
doesn't get as many hits as it used to but you'll find lots of good stuff here on how temp matters, metabolic rate matters, gas mix matters, depth matters. Even if we had 40m, 1.6Lpm CO2 production, but at 20C temp "CE" testing we would still be forced to extrapolate to actual dives.

 

[doctormike]

They don't exist in the form you are expecting. We are left to extrapolate. But how much Co2 sorb can ideally capture isn't a mystery (for sofnolime 797 its about 150L Co2/kg sorb). So we have bookends.

Right, but that's just one piece of the puzzle. The bookends are moved back by a number of variables, since they are never ideally capturing CO2. And not only is there variability in workload, temperature and CO2 production, but there is variability in what a given level of CO2 does to any human being. So you may know that you are fine at X level of arterial CO2, but someone else will be very symptomatic. Some of that may be addressable by conditioning and training, some may be invariable genetics.

I hav heard that in commercial diving, they used to test for ox tox susceptibility, and not train people who were on the right side of the bell curve. If you seize in a chamber at PO2 of 2.0, it won't hurt you but you might not get picked for class...

John's blog John Clarke Online
doesn't get as many hits as it used to but you'll find lots of good stuff here on how temp matters, metabolic rate matters, gas mix matters, depth matters. Even if we had 40m, 1.6Lpm CO2 production, but at 20C temp "CE" testing we would still be forced to extrapolate to actual dives.

Thanks! Will take a look. Seriously, I'm not trolling or being snarky. I want to learn for more experienced divers, and I appreciate your posts in this thread.

But I also feel that I should question stuff that is justified only by saying that a lot of very good divers do this, when there isn't any science to back it up. Sure, maybe it's fine, maybe those divers are nowhere near the line. But until we figure out where the line is, it's a lot of guesswork, and there may be a survivor bias at play.

 

[rjack321]

I believe the navy quit O2 testing divers as it wasn't predictable - not sure where I heard that, treat that as rumor.

I dunno man, when my instructor and the actual manufacturer of the Megalodon says a 6lb axial in 45-55F water conservatively lasts 4 hours (and their smaller 5.5lb radial is afforded 5hrs) I don't think there is any survivorship bias there. Those are written manufacturer/instructor recommendations.

So do you bail out when you hit 3 hrs on the JJ? I have gone over 4 hrs on my 6lbs axial by small amounts, usually when I'm on the third dive and it runs 5-15mins over what I expected. If I totally screwed the pooch and was delayed in a cave I would do 5hrs but only if I was shallow, sedentary, and had my O2 bottle and BOV readily at hand. I don't like going over 4hrs (at all) but I like shivering and terrible offgassing on OC even less. And in 50F water I would be quite cold at 4+ hrs.

 

[doctormike]

I believe the navy quit O2 testing divers as it wasn't predictable - not sure where I heard that, treat that as rumor.

Yup, could be an apocryphal story. But actually not a bad idea if you are going to put people through expensive training. Sort of like how you wash out of astronaut school if your inner ears can't handle the centrifuge, etc...

I dunno man, when my instructor and the actual manufacturer of the Megalodon says a 6lb axial in 45-55F water conservatively lasts 4 hours (and their smaller 5.5lb radial is afforded 5hrs) I don't think there is any survivorship bias there. Those are written manufacturer/instructor recommendations.

OK, if ISC publishes specs that say that the 6 lb Axial scrubber lasts for 4 hours, that's great, that's exactly what I'm looking for on the JJ! I greatly respect my instructors too, but I do understand that they are human beings, subject to the same thought processes that the rest of us are. If the only evidence of scrubber duration were to be my instructors word, I would still look elsewhere.

Why wouldn't it be survivorship bias? If you do something many times and you never have a problem and you say that it's safe based on your experience, that's the very definition of survivorship bias.

So do you bail out when you hit 3 hrs on the JJ?

No, I plan my dives!

LOL. I know what you mean, and of course I'm not being unreasonable about this - I understand that my risk of a CO2 hit doesn't go from 0% at 179 minutes to 100% at 181 minutes. When I was in Bonaire with some friends and ended up finishing a dive with them with 3:20 on the scrubber, I also remembered what my instructor told me about warm water.

All I want is some good warm water data for my scrubber, to act as a backstop. It's totally doable. If I had a lot of money, I would do it myself. But I don't.

 

[rjack321]

Why wouldn't it be survivorship bias? If you do something many times and you never have a problem and you say that it's safe based on your experience, that's the very definition of survivorship bias.

Survivorship aka selection bias only exists if you are drawing biased or erroneous conclusions. Defined as "deriving an erroneous assumption because dead people aren't reporting"
If 100% of people dive a 6lbs axial for 4 hrs in local waters (45 to 55F) and there are zero Co2 hits (going to ignore packing errors or time counting errors) there is no error there. So you are not using the reporting "survivors" to make an erroneous conclusion. Is there something to overlook is an open question.
Survivorship bias - Wikipedia
"Survivorship bias or survival bias is the logical error of concentrating on the people or things that made it past some selection process and overlooking those that did not, typically because of their lack of visibility."

To confirm we had/have selection bias we have to look for and find a dive (preferably more than one) with a co2 hit, in under 4 hrs. There are two problems with this. The first is that its way outside the norm aka rare like way more than 3 sigma rare. So "has it happened" is unknown, "was it reported" and who it was reported to also unknown. There's no master database of dive vs Co2 events so proving that negative isn't really possible. Either that lone Co2 hit doesn't exist, it isn't reported, it was reported and hushed up, maybe it was reported just not to "us" or publicly - are all possibilities that we can't conclusively know. But not having data is not the same as overlooking contrary data through erroneous selections.

The other issue is what statistical risk are you willing to accept. 4 sigma below the curve would be ~6 CO2 hits in 1,000 dives which seems like we would have heard about since the Meg axial has been in production since about 1996. But that assumes that scrubber breakthrough at the extreme tails of its time distribution is normal - which it probably isn't even under one set of test conditions nevermind under a myriad of depth, time, co2 load, and diluent conditions. Going to guess here that at some point the tails of the distribution have become such rare conditions that there aren't enough CCR dives in recorded history to have a Co2 hit that soon or to have a scrubber last that long (two tails). The CE test data don't really inform us about those tails, so we are left making inferences.

 

[stuartv]

Why wouldn't it be survivorship bias? If you do something many times and you never have a problem and you say that it's safe based on your experience, that's the very definition of survivorship bias.

At what point does "many" change from survivorship bias to "data"?

If I do something dumb once or twice and don't have a problem, saying that it's safe is what I would call SB.

If I do something 10,000 times and don't have a problem and then say it's safe (for me), is that still SB?

 

[doctormike]

Survivorship aka selection bias only exists if you are drawing biased or erroneous conclusions. Defined as "deriving an erroneous assumption because dead people aren't reporting"
If 100% of people dive a 6lbs axial for 4 hrs in local waters (45 to 55F) and there are zero Co2 hits (going to ignore packing errors or time counting errors) there is no error there. So you are not using the reporting "survivors" to make an erroneous conclusion. Is there something to overlook is an open question.
Survivorship bias - Wikipedia
"Survivorship bias or survival bias is the logical error of concentrating on the people or things that made it past some selection process and overlooking those that did not, typically because of their lack of visibility."

To confirm we had/have selection bias we have to look for and find a dive (preferably more than one) with a co2 hit, in under 4 hrs. There are two problems with this. The first is that its way outside the norm aka rare like way more than 3 sigma rare. So "has it happened" is unknown, "was it reported" and who it was reported to also unknown. There's no master database of dive vs Co2 events so proving that negative isn't really possible. Either that lone Co2 hit doesn't exist, it isn't reported, it was reported and hushed up, maybe it was reported just not to "us" or publicly - are all possibilities that we can't conclusively know. But not having data is not the same as overlooking contrary data through erroneous selections.

The other issue is what statistical risk are you willing to accept. 4 sigma below the curve would be ~6 CO2 hits in 1,000 dives which seems like we would have heard about since the Meg axial has been in production since about 1996. But that assumes that scrubber breakthrough at the extreme tails of its time distribution is normal - which it probably isn't even under one set of test conditions nevermind under a myriad of depth, time, co2 load, and diluent conditions. Going to guess here that at some point the tails of the distribution have become such rare conditions that there aren't enough CCR dives in recorded history to have a Co2 hit that soon or to have a scrubber last that long (two tails). The CE test data don't really inform us about those tails, so we are left making inferences.

The word "survivor" implies the original "survivorship bias" scenario, which is that dead people don't report their results. But that's not a strict definition for common use. The point is that you introduce a survivor bias if you are recording two outcomes, but one of them is less likely to be reported than the other.

Despite all of what you said, the fact remains that you are basing your conclusion on the idea that you have some sort of accurate accounting of rebreather mishaps over a large population. And as far as I know, no one really has that for stuff like CO2 hits.

You could assemble a fatality database, since it's pretty rare that those don't get reported somehow. But you said it yourself - "all are possibilities that we can't conclusively know". So the burden of proof is on the positive.

If you want to convince me that as a general rule, it's safe to run your scrubber for 150% of the published stack time if the water temperature is greater than 18C, or something like that, I'm going to ask how you know that. It's not an unreasonable question. If the answer is that all of these CCR divers in the Caribbean are doing that, then I'm going to want to know how I compare to them, both in terms of CO2 production and of sensitivity to hypercapnea. And I'm going to want to know what their dive profiles are and their workload. Imagine trying to publish this. Someone is going to ask for your materials and methods.

We keep going round and round the same issue. I'm not "overlooking contrary data". I'm saying that what you are calling contrary data, I consider to be anecdotal data. I mean, why not work this out and publish it? It would be useful to the rebreather community to have more versions of that breakthrough graph, so that you would actually know scrubber behavior from a reproducible experiment.