Info Deeply Safe Labs: A website for dive computer testing

[cerich]

Hello,

This post raises a number of questions. It implies (though you don't state it definitively) that some of the computers you tested did "take account of aggravating factors". I would be curious to know what factors were taken into account by those computers?

There are many reasons why uptake and elimination of inert gas might be asymmetrical such as being warm during uptake and cold during elimination, or exercising during uptake and resting during elimination. I agree that these sorts of factors might come into play during a surface interval between dives.

The idea that "right-left pulmonary shunt" might be a factor is much less well established. For those who have become confused by the debate on this let me explain. Eric is arguing (or at least citing Buhlmann as suggesting) that venous bubbles passing to the lungs after a dive can impair gas exchange. They could do this by altering blood flow through the lung (we will return to this below). To understand this, think about how the lung works. It relies on well-ventilated alveoli (i.e., plenty of gas moving in and out of lung air sacs) being exposed to good blood flow through the network of capillaries that surround them. This balance between gas movement (ventilation) and blood flow (perfusion) to optimize gas exchange is referred to as the "ventilation-perfusion ratio".

It is easy to conceive how changes in ventilation or perfusion could disturb this. For example, imagine if there was no ventilation of an alveolus, but blood continued to pass through its surrounding capillaries. With no ventilation (gas movement) there will be no gas exchange, and so the incoming venous blood can pass that alveolus into the arteries without getting oxygenated (or eliminating nitrogen). This is sometimes referred to as 'shunt' (as Eric did) which has confused some divers on this thread who are more used to hearing this term in relation to problems like patent foramen ovale. In fact, the correct term where it is happening in the lungs is actually 'venous admixture'. Moreover, in most of the lung, things are not quite as black and white as in my example. Problems are more often a case of ventilation and perfusion not being perfectly matched rather than one or the other being totally absent. Either way, the net effect is a reduction in the efficiency of gas exchange.

So, returning to the debate about whether venous bubbles can cause ventilation-perfusion mismatch with a reduction of gas exchange efficiency and therefore a reduction of inert gas elimination between dives.... Yes, Buhlmann mentions it in the computer workshop proceedings, but he does not say how he knows it to be true. No references are cited. I do not have a translated copy of Tauchmedizin, so I don't know what he says in there.

The only definitive study of this phenomenon that I am aware of was published by the famous respiratory physiologist Michael Hlastala in 1979 [1] (attached). He and colleagues exposed anaesthetized dogs to venous gas (helium or nitrogen or sulphur hexafluoride) infused at 0.2ml/kg/min. This is a lot of gas - equivalent to 14ml / min in a 70 kg human. This is a level of venous gas embolisation that would not be seen in a human diver, and if it was I suspect it would be rapidly symptomatic. Despite this very high gas load the clearance of nitrogen in a single pass through the lung only fell from 96% to 93% over a 15-minute infusion of gas. Perhaps this result is not surprising given that venous bubbles tend to obstruct perfusion rather than reduce ventilation, and blood from obstructed vessels would tend to be 'diverted' through areas of the lung where flow remains less impaired thus maintaining gas exchange. Nevertheless, the usual efficiency of the lung would be impaired a little, hence the fall from 96 to 93%.

Bearing in mind that this was, in human terms, a massive gas infusion, I think it reasonable to conclude that in asymptomatic divers, venous bubbles arriving in the lungs during a surface interval are very unlikely to materially affect nitrogen elimination to produce the sort of asymmetry in gas uptake and elimination that Eric is talking about.

To answer DDM's question, even if between-dive VGE did reduce inert gas elimination, it is not clear to me (given the marked potential variability in bubbling even when the same profile is performed) how a computer would be able to take account of this unless it was interfaced in real time to a bubble counting device. In any event, it seems likely, based on the available evidence, that venous bubbling only exerts a minor / inconsequential effect on inert gas elimination.

Simon M

1. Hlastala M, et al. GAS EXCHANGE ABNORMALITIES PRODUCED BY VENOUS GAS EMBOLI. Respiration Physiology. 1979;36:1-17.

Thank you Simon, I remain as I have been for many years, in awe of your knowledge and jealous of your ability to communicate complex subjects in a manner that is easily understood.

 

[dmaziuk]

Dive safe, don't set 90/90, NOBODY thinks that is a good idea, except some dude screaming it's default when it isn't it seems

90/90 is the "standard" preset on SEAC Guru. Quoting it in this thread is at best misleading as it is not intended for planned decompression dives. They also have a "0" preset at 93/93 that would more-or less roughly mimic DSAT -- designed for repetitive no-stop diving, with about 30 years of practical use "in the wild", and a pretty good track record over said years.

By conflating that with planned decompression "technical" dives, one can easily do what in my profession is called "spreading FUD": Fear, Uncertainty, Doubt.

Edit: PS. And for its intended usage: the multi-day multi-dive no-stop schedules, 90/90 should be somewhat on the "conservative" side of DSAT, although ZH-L16 and DSAT don't match 1-to-1 so... @scubadada knows how they stack up in practice.

 

[cerich]

90/90 is the "standard" preset on SEAC Guru. Quoting it in this thread is at best misleading as it is not intended for planned decompression dives. They also have a "0" preset at 93/93 that would more-or less roughly mimic DSAT -- designed for repetitive no-stop diving, with about 30 years of practical use "in the wild", and a pretty good track record over said years.

By conflating that with planned decompression "technical" dives, one can easily do what in my profession is called "spreading FUD": Fear, Uncertainty, Doubt.

agree 100% with you.

This is why design choices and intended use (and manuals) are so important.

 

[tursiops]

90/90 is the "standard" preset on SEAC Guru. Quoting it in this thread is at best misleading as it is not intended for planned decompression dives. They also have a "0" preset at 93/93 that would more-or less roughly mimic DSAT -- designed for repetitive no-stop diving, with about 30 years of practical use "in the wild", and a pretty good track record over said years.

By conflating that with planned decompression "technical" dives, one can easily do what in my profession is called "spreading FUD": Fear, Uncertainty, Doubt.

Edit: PS. And for its intended usage: the multi-day multi-dive no-stop schedules, 90/90 should be somewhat on the "conservative" side of DSAT, although ZH-L16 and DSAT don't match 1-to-1 so... @scubadada knows how they stack up in practice.

I thought 90/90 or so "matched" DSAT on the first dive only...not on repetitive dives. @scubadada ??

 

[scubadada]

90/90 is the "standard" preset on SEAC Guru. Quoting it in this thread is at best misleading as it is not intended for planned decompression dives. They also have a "0" preset at 93/93 that would more-or less roughly mimic DSAT -- designed for repetitive no-stop diving, with about 30 years of practical use "in the wild", and a pretty good track record over said years.

By conflating that with planned decompression "technical" dives, one can easily do what in my profession is called "spreading FUD": Fear, Uncertainty, Doubt.

Edit: PS. And for its intended usage: the multi-day multi-dive no-stop schedules, 90/90 should be somewhat on the "conservative" side of DSAT, although ZH-L16 and DSAT don't match 1-to-1 so... @scubadada knows how they stack up in practice.

I thought 90/90 or so "matched" DSAT on the first dive only...not on repetitive dives. @scubadada ??

The Seac computers run Buhlmann with presets, no custom GFs. The most liberal preset is 85/85. The GFs corresponding to the six presets are not in the owner's manual, I requested the information from Seac

It is the Ratio computers that have a presets of 93/93 and 90/90, other presets are 80/80, 75/75, 50/85, and 45/80. They have no custom GFs on the Pro or Deep model, custom GFs are available on the Tech+ model.

The Aqua Lung i330R runs Buhlmann with presets of 90/90, 35/85, and 35/70, no custom. The Apeks DSX runs the same 3 presets but also offers custom GFs.

I have been diving a computer running DSAT since 2002, a few over 2300 dives. I have been diving a second computer running Buhlmann ZH-L16C with GF since 2016, about 1150 dives, half on a Dive Rite Nitek Q and the other half on a Shearwater Teric I bought in 2019.

Though DSAT and Buhlmann at a GF high of 95 run similarly, they are somewhat different on both first dive and on repetitive dives. On a first, clean dive, DSAT is more liberal than a GF high of 95 and runs closer to a GF high of 99. I leave my Teric set on a GF high of 95 and find this difference most noticeable on slightly deeper dives, when I am more likely to come close to or exceed my NDL. Here is a table of the NDLs for DSAT and Buhlmann using 32% nitrox:

On repetitive dives, especially with shorter surface intervals and perhaps shallower, a GF high of 95 is often more liberal than DSAT. They tend to even out again with a longer surface interval, like a lunch break before afternoon dives. When I first started diving my Nitek Q, I would change GF highs to see if I could match DSAT. I would dive the first dive at a GF high of 99 and then lower it to 90, or sometimes 85 for the second dive. Though this worked out fairly well, it was a nuisance. For quite a while, I have left my GF high at 95. I simply dive the more conservative of the computers for a no stop dive or clear one or both computers if it is a light deco dive.

I have never kept systematic track of the differences between DSAT and Buhlmann at a GF high of 95. However, I have posted a few examples, here is one:

DSAT vs GF high 95, examples

I have written previously about my experience diving DSAT and Buhlmann ZH-L16C with a GF high of 95. I have an example of what I have previously reported. This dive is a first dive of the day on the Castor in Boynton Beach. It was relatively deep with a max depth of 107 feet and an avg depth of...

scubaboard.com

 

[dmaziuk]

Guru was a rebadged Ratio IIRC

 

[scubadada]

Guru was a rebadged Ratio IIRC

View attachment 834098

You're right, I forgot this . The Guru used a Ratio transmitter for AI also. It's on my transmitter compatibility list,

Are Pressure Transmitters compatible across brands?

I've been a fan of HW for many years, owned 2 of their computers and one of their lights. Customer service rivals Shearwater, just amazing. I had to double check wondering how I missed the introduction to AI. Don't think they'll ever go down that route. To be fair, HW launched an AI computer in...

scubaboard.com

 

[Deeply Safe Labs]

Hi @Dr Simon Mitchell

Thank you for your detailed answer. The study you reference is very interesting. I am unaware of any reference made by Pr Bühlmann to this article, given that it is anterior to his work on ZH-L16, but considering that he was specialized in pathophysiology of the respiratory system, it seems unlikely to us that he was unaware of it.

We have made a machine translated version his book Tauchmedizin [1]. The translation isn't perfect, but gives a general idea of his writing and allows cross checking the translations using different online translators. For obvious intellectual property reasons, we will not share it publicly, but since you own the German original, we will gladly share it with you by private message. I will nevertheless share some short quotes for the understanding of other readers.

It seems that Pr Bühlmann somewhat agrees with the numbers on this paper, [1], page 106:

The proportion of this right-to-left shunt in the cardiac output is up to 10% at rest and drops at work to approximately 2%.

His human subject experiences showed results that corroborate this previous statement, [1], page 147:

In 9 different series of tests with 127 divers, 127 Second dives and an additional 39 third dives were carried out. [...] Symptoms of insufficient Decompression occurred after the repeated dive in both the interval of 20 min as well as at an interval of 120 min after the repeated dive. [...]
The N2 release during the surface interval was calculated without correction for a possible right - left shunt. Table 23 shows that on repeated dives the N2 tolerated pressure values are on average 90-97% of the “theoretical” limits.

Regarding the Uwatec lawsuit, I believe that some corrections are necessary. First, I am not the founder of Uwatec. I was recruited as a technical manager for the French subsidiary. I never owned any shares in Uwatec, and therefore never received anything from its sale to Johnson.
On April 22, 1996, the Uwatec USA Inc. staff tried to make a recall for the faulty Air X NitrOx, unsanctioned by Uwatec AG (Switzerland) headquarters [2], point 79. They were fired by Heinz Ruchti the very next day [2], point 38 & 81, and were replaced by Bret Gilliam (TDI #1), which learned about the aborted recall attempt, but ignored it under the advise of Ruchti [2], point 82.
In France specifically, not a single faulty Air X NitrOx reached the retailers, because I refused to send them.
The person to blame here is Heinz Ruchti, and surely not Pr Bühlmann himself.

Now, concerning the issue at hand, let's leave these physiological, philosophical and ethical debates on the side. Allow me to better explain the point we are trying to make:

The inverse correlation between decompression time and risk of DCS is well known.

Dive statistics show that repetitive dives have a higher incidence of DCS.

Therefore, it seems reasonable to assume that the ZH-L16 C computers we have tested, which compute a significantly lower increase in decompression times for repetitive dives than other computers we have tested, induce a higher risk of DCS than these other computers for repetitive dives.

How much higher is this risk? We do not know, and that is not our job to find out, neither is it that of divers. It's yours. Not yours specifically, Dr Mitchell, but yours as the scientific community as a whole. Our job ends with testing computers, publishing factual results, comparing them to the state of the art and alerting about differences.

Best regards,
Eric Frasquet,
Deeply Safe Labs.

[1] Tauchmedizin, Barotrauma Gasembolie · Dekompression Dekompressionskrankheit Dekompressionscomputer, A. A. Bühlmann, E. B. Völlm, P. Nussberger, 5th Edition, Springer, 2002. Tauchmedizin
[2] UNITED STATES DISTRICT, COURT NORTHERN DISTRICT OF CALIFORNIA, Case 4:03-cv-00513-WDB, Robert Raimo v. Uwatec, attached in Did Uwatec conceal a known computer flaw?

 

[Nick_Radov]

Now, concerning the issue at hand, let's leave these physiological, philosophical and ethical debates on the side. Allow me to better explain the point we are trying to make:

The inverse correlation between decompression time and risk of DCS is well known.

Dive statistics show that repetitive dives have a higher incidence of DCS.

Therefore, it seems reasonable to assume that the ZH-L16 C computers we have tested, which compute a significantly lower increase in decompression times for repetitive dives than other computers we have tested, induce a higher risk of DCS than these other computers for repetitive dives.

That is a plausible hypothesis. It would be interesting to see a large-scale controlled study to test it with a few hundred divers doing chamber dives at the limits of various deco algorithms, with or without the various "additional procedures" that you advocate. But until we have results from such a study, it seems that you have skipped some steps in the scientific process and your rather arbitrary tests of various dive computers aren't telling us anything useful.

 

[Dr Simon Mitchell]

His human subject experiences showed results that corroborate this previous statement, [1], page 147:

Hello again Eric,

The quote you cite from Tauchmedizin does not convince me that Buhlmann actually measured (as opposed to calculated) nitrogen washout (whereas Hlastala measured it), and nor does it convince me (or even imply) that he correlated his calculated changes in nitrogen washout against a measured venous gas load (as Hlastala did). I'm therefore dubious that this makes a convincing argument in support of the hypothesis that venous bubbles after diving create a ventilation-perfusion mismatch of sufficient magnitude to materially alter inert gas washout kinetics.

I agree that this is not a point worth debating 'to the end of the earth' but as others have pointed out, you were emphasising this hypothesis as an explanation for asymmetric gas kinetics and the best evidence I am aware of suggests that it is not likely to be significant effect.

The inverse correlation between decompression time and risk of DCS is well known.

Dive statistics show that repetitive dives have a higher incidence of DCS.

Therefore, it seems reasonable to assume that the ZH-L16 C computers we have tested, which compute a significantly lower increase in decompression times for repetitive dives than other computers we have tested, induce a higher risk of DCS than these other computers for repetitive dives.

How much higher is this risk? We do not know, and that is not our job to find out, neither is it that of divers. It's yours. Not yours specifically, Dr Mitchell, but yours as the scientific community as a whole. Our job ends with testing computers, publishing factual results, comparing them to the state of the art and alerting about differences.

The claim that longer decompression = safer decompression, all other factors being equal, is demonstrably true (many dive table testing / development programs have demonstrated this). However one has to be careful with simplistic claims that computer x must be safer than computer y because computer x prescribes longer decompressions. There are several reasons for saying this.

First, to materially alter risk between two profiles the difference in decompression time may need to be substantial. I remember Wayne Gerth (who ran the US Navy table development program) making this point at the Deep Stops Workshop in 2008.

Second, although longer decompressions may reduce the risk of DCS, they can introduce other risks like increased risk of oxygen toxicity, cold stress, logistic difficulties etc. The limits prescribed by every table or computer represent the manufacturer's interpretation of a 'line in the sand' of acceptable risk, and their sense of an acceptable balance between length of decompression (with its various challenges), risk of DCS, and risk of other things (such as the examples I gave) associated with decompression.

It follows, for example, that a very conservative computer may prescribe profiles that are safer from a DCS perspective, but many divers may not want to use it because the decompressions are too long and the diver may choose to accept the slightly higher risk of DCS associated with using another device after evaluating their personal priorities and circumstances.

Finally, one of the reasons that I emphasised the phrase "all other factors being equal" above is that length of decompression may not be directly related to DCS safety when comparing computers using completely different decompression algorithms. For example, (and speaking hypothetically - I'm not thinking of any particular devices here) if a computer emphasises deep stops, then it may require longer shallow stops to compensate for increased slow tissue loading during deep stops. The resultant longer decompression might not necessarily be safer than a shorter decompression prescribed by a computer that places less emphasis on deep stops. In a situation like that, comparative testing (as you imply) is the only way to definitively answer the safety question.

Simon M