Which Algorithm, Why?

[gee13]

Not sure whether this has been discussed previously but I have been skimming the surface of different dive computers ranges and algorithms used. It seems that different manufacturers have their own adaptation of the early Navy dive tables. I hear of Buhlmann - ZHL8, MB, PG, ZHL12, ZHL16 (Z+), GF, VBPM, Randy boher, then there is Pelagic Haldanean, DSAT Spencer-Powell, Mares Wienke, Suunto Wienke.

Im curious to invite some discussion on what algorithms are used and why you prefer it? How does this affect choice of dive computers? How do they apply to normal limits of recreational diving to 30M? How do they apply to deeper technical dives post 30M?

Im sure most divers who start out choose price and brand first and are left with whichever algorithm the computer uses. That is if they even choose to even understand what the algorithm dictated by the computer means to them?

 

[Hatul]

The liberal Pelagic Hadanean algorithm is being abandoned by Oceanic in their newer computers. Instead they're using the Z+ which is similarly conservative to the Suunto, when I looked at the graphs, so it may indicate a convergence of algorithms.

Personally I dive for recreation so I prefer the extra safety of a conservative algorithm.

Adam

 

[hdugan]

You missed a few variations, Buhlmann ZH-L16A,ZH-L16B,ZH-L16C. The other variations Buhlmann are modifications of ZH-L16A. ZH-L8 and L12 use the same alogarithm as L16, they just use fewer "compartments". The difference between 16A - 16C is Theory provided the M values for 'A'. After actual testing, M values were modified to those in L16B. ZH-L16B is intended for table use, L16C is 'softened' (M values relaxed) for real time computer use. All Bulhmann really did was extend Workmans equations to allow altitude diving, Workmans equations use sea level as a base and can't be used at less than 1 ATM. In addition Bulhmann also provides 'M' values for Helium. I've never seen Workman 'M's for that gas.

GF is Bulhmann with a twist, you never reach the 'M' values when decompressing. Bulhmann assumes that the 'M' value is a hard line, above the line not bent, below the line bent. We know this to be in error now (the reason you often hear straight Bulhmann refered to as "Bend and Treat") It's a way of modifing the 'M' value, you are not allowed to get within a certain percentage of the 'M' value.

I must assume a typo for VPBM. This extends Bulhmanns basic equations from a single phase (dissolved gass) to a dual phase (dissolved gas and micro bubbles) model. The disolved gas part is taken care of by Bulhmann (often L16C), and the model tries to minimize the growth and number of "Micro Bubbles".

RGBM is (or at least was) VPBM with a few propritery twists. Last time I checked, Wienke was claiming it wasn't VPBM. What has been added is the use of statistical analysis to better fit acutal risk of DCS for a given profile. It doesn't work on a divable computer.

Suunto's RGBM "seems" to be the Ideas That underpin VPBM but they aren't actualy calculated, they are "folded" in, a little like 'Gradients'. Now all this is guess work as Suunto keeps the details of SRGBM Very tight to the vest. One advantage of this method is IF you don't do anything to flag the RGBM part, you can dive it very close to a Workman/Bulhmann model. It's only after you make it mad ( ascent rate, sawtooth profile, reverse profile, ect) that you get puinished.

The last time I checked (a number of years ago) Pelagic used Workman's equations with Spencer's 'M' values. This was very close to the 'Navy Tables". It provides the most liberal model also.

 

[jd950]

The liberal Pelagic Hadanean algorithm is being abandoned by Oceanic in their newer computers. Instead they're using the Z+ which is similarly conservative to the Suunto, when I looked at the graphs, so it may indicate a convergence of algorithms.
Adam

They aren't abandoning their customary algorithm, they supplemented it with the Z+ algorithim in their newer computers. By choosing the desired algorithim, along with whatever conservative settings and customizable alarm setpoints are desired, these computers allow a diver to be as liberal or conservative as they may choose and even set the computer to more or less match a dive buddy's computer.

I don't think it is a convergence of algorithims, as much as it was a marketing decision to offer customers the ability to use their Oceanic computers to dive with algorithims similar to Uwatec and Suunto computers.

 

[rossh]

You missed a few variations, Buhlmann ZH-L16A,ZH-L16B,ZH-L16C. The other variations Buhlmann are modifications of ZH-L16A. ZH-L8 and L12 use the same alogarithm as L16, they just use fewer "compartments". The difference between 16A - 16C is Theory provided the M values for 'A'. After actual testing, M values were modified to those in L16B. ZH-L16B is intended for table use, L16C is 'softened' (M values relaxed) for real time computer use. All Bulhmann really did was extend Workmans equations to allow altitude diving, Workmans equations use sea level as a base and can't be used at less than 1 ATM. In addition Bulhmann also provides 'M' values for Helium. I've never seen Workman 'M's for that gas.

I must assume a typo for VPBM. This extends Bulhmanns basic equations from a single phase (dissolved gass) to a dual phase (dissolved gas and micro bubbles) model. The disolved gas part is taken care of by Bulhmann (often L16C), and the model tries to minimize the growth and number of "Micro Bubbles".

Hello,

I would like to correct the statement about VPM-B. VPM does not extend Buhlman or use the Buhlmann model. They have only one part in common - the dissolved gas tracking. This component can also be found in most models in various forms.

The core of a Buhlmann model is to provide pairs of matched Coefficients to halftime values, applied to dissolved gas values, from which M values are created.

For VPM, the dissolved phase gas tracking is via Schriener and Haldane equations. VPM then applies micro bubble physics (the free phase) and controls the ascent through those considerations.

More details of VPM can be found here: VPM Decompression Site

Regards

 

[hdugan]

.... I would like to correct the statement about VPM-B. VPM does not extend Buhlman or use the Buhlmann model.....

"Extends" was a bad choice of wording. The disolved gas functions are based on Bulhmann's half times

 

[gcbryan]

The bottom line is actually it doesn't make a whole lot of difference for recreational profiles as the data in this range is observed and all of the models pretty much converge in the recreational range (they fit their models to the observed data in this range).

The theory part as opposed to observed data is when you get deeper. That's not to say all computers use the same NDL numbers they don't but the numbers used in recreation data are derived less from theory and more from actual observed data. That's my take on it anyway.