Shearwater Gradient Factors=DSAT?

[victorzamora]

I don't think so. Only kicks in when you exceed deco and then dictates depth of 1st stop. However, MultiDeco takes it into account In every plan. It's interesting that Buhlmann can be implemented in different ways. It makes me wonder if there are and differences in how it is implemented in dive computers and/or other planning software. If someone was compulsive enough, they could put it to test.

It definitely IS implemented differently on different planners/computers. For a simple test, check two different planners against each other. Come up with a pretty deep/long dive for exaggerated differences. If you have access to a ZHL-16 dive computer, you can also use the "planning" function on it and check its plan compared to the software, or other ZHL-16 dive computers. There's also a difference in ZHL-16B and ZHL-16C in that ZHL-16B is technically "better" but -16C was developed for dive computers of lower processing power.

 

[dmaziuk]

You've said this several tmes, I think based on the original paper on the DSAT work some 25-30 years ago, no? Since then, it would seem the empidemiology would suggest DSAT is much safer than its developers thought it might be.

What others said.

If you are not diving them to the limits, then what does it matter if/how two algorithms match up? If your regular diving is within the limits of the more conservative of the two algorithms, fine, who cares. And if your buddy's scary conservative computer is cramping your style, you should ask them whether they're comfortable diving on your more aggressive schedule. Or find another buddy.

 

[aquacat8]

You’re right, it doesn’t matter, I try to get out of the water with plenty of no deco time left anyway, I was just curious.

 

[dmaziuk]

You’re right, it doesn’t matter, I try to get out of the water with plenty of no deco time left anyway, I was just curious.

That wasn't addressed at you personally, it was a generic "you". It irks me when people quote scubalab studies like they mean something, or say an algorithm has been well "tested by time". Six simulated dives on a few off-the-wall profiles don't make the statistics, and if you don't dive an algorithm to the limits, you're not testing the algorithm. Since many of us get out of the water well within our computers' NDLs, those NDL mean even less than what they're originally supposed to mean: that the statistical chances of (a generic) you developing DCS go from two in ten thousand to three in 9,999. Or whatever the M-values programmed in a given computer translate to.

 

[scubadada]

It definitely IS implemented differently on different planners/computers. For a simple test, check two different planners against each other. Come up with a pretty deep/long dive for exaggerated differences. If you have access to a ZHL-16 dive computer, you can also use the "planning" function on it and check its plan compared to the software, or other ZHL-16 dive computers. There's also a difference in ZHL-16B and ZHL-16C in that ZHL-16B is technically "better" but -16C was developed for dive computers of lower processing power.

Hi Victor,

I have thought that Buhlmann ZH-L16A was found to be too aggressive, so that some of the intermediate compartments were revised and that B was recommended for use in tables. I thought some of the compartments of B were further revised to yield C, recommended for implementation in dive computers. I think the specs are published in Deco for Divers, but do not have access to it while on vacation. I have never heard the alternative, computer processing explanation.

I have wondered why Ratio chose to implement Buhlmann ZH-L16B. Divesoft, Garmin, Shearwater, Divecomputer.eu and Heinrichs Weikamp all use C. Anybody have any ideas about this?

 

[scubadada]

...It irks me when people quote scubalab studies like they mean something...

They do mean something, it's how the computers tested behaved on the 4 hyperbaric chamber, simulated dive profiles. The profiles could not have been much more aggressive or the most conservative computers, Cressi, Mares, Scubapro, would have gone into deco and could not have been included in the report. I believe the profiles were chosen with care to not exclude any of the currently available deco algorithms.

There is very little controlled data on repetitive dives, do you have better data?

 

[victorzamora]

Hi Victor,

I have thought that Buhlmann ZH-L16A was found to be too aggressive, so that some of the intermediate compartments were revised and that B was recommended for use in tables. I thought some of the compartments of B were further revised to yield C, recommended for implementation in dive computers. I think the specs are published in Deco for Divers, but do not have access to it while on vacation. I have never heard the alternative, computer processing explanation.

I have wondered why Ratio chose to implement Buhlmann ZH-L16B. Divesoft, Garmin, Shearwater, Divecomputer.eu and Heinrichs Weikamp all use C. Anybody have any ideas about this?

ZHL-16A was the original. ZHL-16B used revised a values in the middle compartments. ZHL-16C is simplified to reduce computational requirements. That's been my understanding from a few sources and is backed up by the almighty repository of never-incorrect knowledge, Wiki. How accurate the Wiki page is, of course, up for debate. ZHL-16C may well be a modification of tissue values for computer vs table diving.....but it's not how I remember it.

That's why I thought Ratio had ZHL-16B - They've got a chunkier processor and don't mind spending the battery capacity on the more "full fat" variant of the algorithm.

 

[dmaziuk]

This, for example: firegnom/mvplan (which I'm sure is also never wrong) seems to differ only in numbers between -B and -C. They don't look like one set should be noticeably more or less computationally intensive than the other... I'm sure "can be used with almost all low-level processor units but it is less flexible compared to the ZHL16B" meant something to the guy who wrote it.

Edit:

Eric Baker's explanation (http://scholar.google.com/scholar_url?url=http://tsa.kapsi.fi/files/deko/Eric%20Baker%20-%20Understanding%20m-values.pdf&hl=en&sa=X&scisig=AAGBfm1kdd9XKs9v1ZEG4kuisEctIzVDHw&nossl=1&oi=scholarr) is the one that makes sense: the -B set is less conservative because when diving tables one plans for max. depth and that adds the conservatism. In a dive computer that runs off from actual time & pressure that extra conservatism needed to be added to the numbers instead. That's the -C set for computers.

 

[gopbroek]

Until last summer I used the Oceanic DSAT algorithm but then I purchased a Perdix and had the same question. On Scubadada advice I set the GF to 45/95. The first two dive trips I did I wore both computers side by side and then downloaded them and compared the NDL values at various points over multiple dives. Initially the DSAT algorithm seemed more liberal but on repetitive dives the GF 45/95 became more liberal. I posted the comparisons last fall at Algorithm question (Perdix AI vs DSAT)

My take is that on an aggressive repetitive dive schedule the GF45/95 is very liberal and I try to leave a healthy margin off the calculated NDL.

 

[uncfnp]

Do you happen to know where that was said? I've not seen it ... (and here's a link to my reading list on this which I think covered most of those threads

Ha! Disregard my question. I was sure I read something somewhere on SB about when GF Lo could come into play with NDL’s but I have no idea where else to look. My bad.