Info Deeply Safe Labs: A website for dive computer testing

[dmaziuk]

It may be worth remembering that there is no empirical data below a certain level of conservatism, and when it comes to calculating probabilities, that is not the same thing as zero incidence of DCS. I.e. increasing decompression time past a certain point becomes a matter of faith rather than math.

It's not entirely clear how that meshes up with "to materially alter risk between two profiles the difference in decompression time may need to be substantial". The difference between GF100 and GF80 does materially alter risk, we know that, the difference between GF75 and GF55 is anybody's guess as far as I know.

 

[scubadada]

It may be worth remembering that there is no empirical data below a certain level of conservatism, and when it comes to calculating probabilities, that is not the same thing as zero incidence of DCS. I.e. increasing decompression time past a certain point becomes a matter of faith rather than math.

It's not entirely clear how that meshes up with "to materially alter risk between two profiles the difference in decompression time may need to be substantial". The difference between GF100 and GF80 does materially alter risk, we know that, the difference between GF75 and GF55 is anybody's guess as far as I know.

Using the SAUL Recreational Dive Planner, here is the probability of DCS at a variety of bottom times for a 1st clean dive to 100 ft using air. NDLs were calculated with my Shearwater Teric. The dives all include a 3 min safety stop at 15 feet but NO decompression stop time. You can easily calculate missed decompression times using your own deco planner with whatever GFs you choose. Using my computer set at 80/95, the dives for 19-22 min only have 2-3 min of deco and would be covered by the "safety stop". The dives for 23-25 min have 4-7 min of deco time.

SAUL Recreational Dive Planner | Modern Decompression

 

[dmaziuk]

Yes, and if you follow the references, you'll end up finding that there is no data below GF60, or whatever it was I forget, and any probabilities assigned in that range are made up by curve-fitting a line to the data that does exist, and extending it down to where there is no data. There are certain problems with that approach. For example

Edit: graph stolen from Robert's write-up @ Probabilistic decompression models: Probably problematic – The Theoretical Diver

 

[scubadada]

Yes, and if you follow the references, you'll end up finding that there is no data below GF60, or whatever it was I forget, and any probabilities assigned in that range are made up by curve-fitting a line to the data that does exist,and extending it down to where there is no data. There are certain problems with that approach. For example

View attachment 834616

Edit: graph stolen from Robert's write-up @ Probabilistic decompression models: Probably problematic – The Theoretical Diver

The Saul probabilistic decompression algorithm can also be found on The Theoretical Diver, in case you had doubts.

 

[Nick_Radov]

Using the SAUL Recreational Dive Planner, here is the probability of DCS at a variety of bottom times for a 1st clean dive to 100 ft using air.

Those numbers look suspiciously precise. I don't believe we have enough data to calculate DCS probabilities to 3 significant figures. Where are the error bars and confidence ranges?

 

[dmaziuk]

Those numbers look suspiciously precise. I don't believe we have enough data to calculate DCS probabilities to 3 significant figures. Where are the error bars and confidence ranges?

I have a strong feeling that Buhlmann's a and b coefficients are given to 4 decimal digits because they were printed to %5.4f format mask. Welcome to Scientific Computing for Life Sciences.

 

[scubadada]

Those numbers look suspiciously precise. I don't believe we have enough data to calculate DCS probabilities to 3 significant figures. Where are the error bars and confidence ranges?

Sorry, I'm the messenger, I would suggest you do your own background research including the Saul Decompression Algorithm.

 

[Deeply Safe Labs]

Hi @Dr Simon Mitchell

I suppose you will agree that all the dive tables that include a repetitive dive procedure increase the decompression time of repetitive dives according to the time spent at the surface. In the beginning, I seem to remember that Haldane recommend adding the time of both dives. The US Navy used this method, but it was very restrictive. All the navies around the world started using the 120min compartment to calculate decompression of the repetitive dive. Some commercial diving companies preferred to artificially fill all compartments at the end of the first dive. The French company COMEX company used the three methods and selected the most conservative. On the recreational side, DSAT selected the 60min instead of 120, as it seemed to be sufficient in that field.

These methods imply a longer decompression than the one you can calculate if you use the load of inert gas in the leading tissue of the second dive. I’m aware that you will not learn anything reading this, but you can eventually correct me, for the benefit of other readers and mine.

Our tests demonstrate that the dive computers that are implementing an algorithm based on the ZH-L16 C parameter set only use the inert gas load to calculate the decompression of repetitive dives. This point was very easy to confirm by a simple calculation based on the same parameters.

We used a first protocol which can be seen as “heavy” but showed significant differences. We compared the behaviour of some “old” computers to ZH-L16 C dive computers. As we used square profiles for these tests, we could compare to some dive tables, but the result will be very similar to that of “old” computers.

The point is: some ZHL16C dive computers gave a decompression time divided by 2 for the second dive. Maybe claiming that computer Y is not as safe as computer X because computer Y prescribes shorter decompressions is also simplistic, but I would be scared should it be true. I’m not the only one, in my country, some hyperbaric chambers representatives are worried as well.

I have no doubt that technical divers can easily counteract this with GF. But the marketing target has changed with the arrival of Garmin, Mares and Seac. Even Shearwater now clearly aims at the recreational diver. That’s what we can be worried about. What do you think @Duke Dive Medicine ?

Best regards,
Eric Frasquet,
Deeply Safe Labs.

 

[dmaziuk]

I have been the source of some of those postings. A couple of years ago a representative of DAN did a presentation at a social gathering at our dive shop. She said that over 80% of DCS accidents happened on the first day of a dive trip, and over 80% happened on the first dive of that day. I have repeated that information several times on ScubaBoard. Unfortunately, there is no way to link to her presentation.

If true, that would suggest the exact opposite of DCS risk increasing with repetitive dives.

 

[tursiops]

I suppose you will agree that all the dive tables that include a repetitive dive procedure increase the decompression time of repetitive dives according to the time spent at the surface. In the beginning

Can you state this more clearly?

These methods imply a longer decompression than the one you can calculate if you use the load of inert gas in the leading tissue of the second dive.

I don't follow your argument.

Our tests demonstrate that the dive computers that are implementing an algorithm based on the ZH-L16 C parameter set only use the inert gas load to calculate the decompression of repetitive dives. This point was very easy to confirm by a simple calculation based on the same parameters.

Not a new or amazing result; you did not need to do your work to verify this.

The point is: some ZHL16C dive computers gave a decompression time divided by 2 for the second dive.

I do not understand your statement. It seems you are saying that if the first dive requires (say) 10 minutes of deco at (say) 3m, for a certain square profile, that there is a computer that requires just 5 minutes of deco at 3m for an identical repetitive dive, regardless of the Surface Interval? If you are not saying this, what ARE you saying?