On a NDL dive, which computers' NDLs are not affected by GFLo?

[huwporter]

Hi @stuartv , OK, firstly 20/85 was used with those profiles as a good illustration of some of the peculiarities and less-than-optimum profile features that the current GF linear steps implementation can throw up with certain profiles, not as a recommendation. The point I was actually making was that spending all that time getting up to the deco gas switch doesn't make a whole lot of sense.

Generally, on a significant multi-stage deco ascent, the faster compartments are controlling (i.e. are closest to M-value, or have the highest gradient factor) at the start of the deco, medium-fast are controlling during the middle and medium/medium-slow by the end. Using a GF-Hi of 85 means that no compartment will be more than 85% of the M-value at the surface; the faster your ascent up to that point, then the faster the controlling compartment will be at the end of deco - the slower your ascent up to that point, the slower the controlling compartment will be.

Personally, my preference is to limit the maximum GF experienced by the fastest tissues, while balanced against avoiding spending 'excessive' time deep to additionally load the medium-slow tissues; my entirely subjective anecdotal experience is that this leaves me less fatigued after bigger dives. The initial proposal I'd put to my team when planning 40 mins @ 65m for real would look closer to my third profile example than the first, more of a controlling of ascent rate than 'deep stops'.

...I also think that many of the people who say they dive, e.g. 85/85, in actual truth are not really diving 85/85 because I don't believe they achieve 9m/min ascents all the way from the bottom to the first stop. Whenever you hear comments like "my computer was set to 85/85 but the first stop had disappeared by the time I got there" - what this actually means is that the extra time they had spent on their slower ascent had effectively inserted 'deep stops' into the profile which had cleared the first 'calculated' stop during their ascent; if they'd actually done 9m/min (or, whatever their ascent rate setting was) all the way, first stop wouldn't have cleared.

 

[stuartv]

...I also think that many of the people who say they dive, e.g. 85/85, in actual truth are not really diving 85/85 because I don't believe they achieve 9m/min ascents all the way from the bottom to the first stop. Whenever you hear comments like "my computer was set to 85/85 but the first stop had disappeared by the time I got there" - what this actually means is that the extra time they had spent on their slower ascent had effectively inserted 'deep stops' into the profile which had cleared the first 'calculated' stop during their ascent; if they'd actually done 9m/min (or, whatever their ascent rate setting was) all the way, first stop wouldn't have cleared.

I think that depends somewhat on the computer. But, I believe your statement is correct for Shearwaters.

Regardless, the hardcore tech divers that I know that are diving GF85/85 or higher seem to generally be aware of the need to ascend at an actual 30'/min. Actually, I think it is pretty normal for most of them to ascend much more quickly than that, from the bottom up to somewhere near their first deco stop.

Anyway, I think we're on the same page. No further questions, your Honor.

 

[KenGordon]

If you ascended instantaneously to 33', where the ambient pressure would be 2 ATA, you would have tissue tension of 3.2 ATA of Nitrogen and the inspired partial pressure of Nitrogen (i.e. the ppN2 in the gas you're breathing) would be 80% of 2, or 1.6 ATA.

That ratio of 3.2 : 1.6 is the over-pressurization gradient (in your 5-minute compartment). So, at that moment that you arrive at 33', that compartment has an over-pressurization gradient of 2.0. The Buhlmann algorithm is accompanied by a set of M-values. Those M-values are the Maximum over-pressurization gradient for each compartment. I

Stuart,

The m-limit applies to the ratio of the inert gas partial pressure, 3.2in your example, and ABSOLUTE pressure, 2.0 here. So the over pressure ratio is only 1.6, not 2.

Consider a gas switch to 50%, now the ratio as per your calculation would be 3.2 to 1 (50% n2 at 2bar absolute). Does the diver suddenly get bent because it isn’t o2 and not n2 being breathed?

 

[KenGordon]

Yeah exactly. I think one of the specific downsides of the current implementation is that it can occasionally throw up disconcertingly long deep stops that don't really make physiological sense . Example:

=========================================================================
 FIRST DIVE: Standard GF Plan:
 65m for 40 mins OC on 15/55 with 50%+O2, GF 20/85

 Phase                Depth   Time       RunTime   Mix O2/He Ceiling GF
 -----                -----   ----       -------   --------- ------- --
 Descent To:          65m     3.6 min    3:37      15/55     0m      0
 Bottom Phase:        65m     36.4 min   40:0      15/55     27m     0
 Ascent To:           42m     2.6 min    42:34     15/55     26m     20
 Stop:                42m     2.0 min    44:34     15/55     25m     20
 Stop:                39m     2.0 min    46:34     15/55     24m     24
 Stop:                36m     3.0 min    49:34     15/55     23m     29
 Stop:                33m     3.0 min    52:34     15/55     22m     32
 Stop:                30m     5.0 min    57:34     15/55     20m     39
 Stop:                27m     6.0 min    63:34     15/55     18m     42
 Stop:                24m     8.0 min    71:34     15/55     16m     47
 Stop:                21m     4.0 min    75:34     50/0      14m     52
 Stop:                18m     5.0 min    80:34     50/0      12m     55
 Stop:                15m     8.0 min    88:34     50/0      10m     62
 Stop:                12m     10.0 min   98:34     50/0      8m      63
 Stop:                9m      17.0 min   115:34    50/0      5m      71
 Stop:                6m      59.0 min   174:34    100/0     0m      75
 Surface:                                174:34                      85

 Deco Time (21-6) = 103
 Estimated Total Runtime (including switches and gas breaks): 202 mins.
=========================================================================

I mean, waiting eight whole minutes at 24m on 15/55 just to get down to 47% of the M-value, when you have a deco gas waiting for you just one stop shallower? And total 30 whole minutes creeping up between 42 and 24m before you can get on a deco gas? Outside the internal logic of the model, It just doesn't make sense, and means you have to do a whole lot more deco shallower to compensate for it. This model behaviour is at least part of what people are complaining about when the 'deep stops are good/bad' discussions come up.

Some alternatives I've played with are:
Setting a maximum tolerable GF for each compartment separately, which then remain fixed through the dive rather than varying. example, you could set for, "never allow compartment 1 to exceed 20% of allowable M-value, compartment 2 to never exceed 40% of allowable M-value, etc etc.

=========================================================================
 FIRST DIVE: Modified GF Plan:
 65m for 40 mins OC on 15/55 with 50%+O2, GF 85

 Phase                Depth   Time       RunTime   Mix O2/He Ceiling GF
 -----                -----   ----       -------   --------- ------- --
 Descent To:          65m     3.6 min    3:37      15/55     0m      0
 Bottom Phase:        65m     36.4 min   40:0      15/55     27m     0
 Ascent To:           39m     2.9 min    42:54     15/55     26m     31
 Stop:                39m     1.0 min    43:54     15/55     25m     31
 Stop:                36m     1.0 min    44:54     15/55     25m     39
 Stop:                33m     2.0 min    46:54     15/55     23m     47
 Stop:                30m     2.0 min    48:54     15/55     22m     54
 Stop:                27m     3.0 min    51:54     15/55     21m     63
 Stop:                24m     5.0 min    56:54     15/55     19m     71
 Stop:                21m     3.0 min    59:54     50/0      16m     75
 Stop:                18m     5.0 min    64:54     50/0      13m     79
 Stop:                15m     6.0 min    70:54     50/0      11m     77
 Stop:                12m     9.0 min    79:54     50/0      8m      81
 Stop:                9m      14.0 min   93:54     50/0      5m      82
 Stop:                6m      53.0 min   146:54    100/0     0m      84
 Surface:                                146:54                      85

 Deco Time (21-6) = 90
 Estimated Total Runtime (including switches and gas breaks): 174 mins.
=========================================================================

Now reduced to 14 minutes on stops below the first gas switch, a lot better. This approach can give "nicer looking" results for some dives, particularly dives with multiple deco gases (like the one above) but not so well for others - it doesn't work well at all for planning backgas deco (you can get into situations where the model can't ever let you surface), and the limits per compartment aren't based on any kind of science, just totally made up from gut feel.

One rather neat solution is to have rules along the lines of; have an initial GF threshold at which you slow your ascent from say 9m/min to 3m/min, then ascend at this rate til you arrive at a second GF threshold, then do a conventional linear interpolation from there to the surface. Example for this, say, do 9m/min up to GF20, then 3m/min up to GF 65, then a linear interpolation from there to GF85 at surface:

=========================================================================
 FIRST DIVE: Multi-Stage GF Plan:
 65m for 40 mins OC on 15/55 with 50%+O2, GF 20/65/85

 Phase                Depth   Time       RunTime   Mix O2/He Ceiling GF
 -----                -----   ----       -------   --------- ------- --
 Descent To:          65m     3.6 min    3:37      15/55     0m      0
 Bottom Phase:        65m     36.4 min   40:0      15/55     27m     0
 Ascent To:           42m     2.6 min    42:34     15/55     26m     20
 Deep Stop:           42m     1.0 min    43:34     15/55     26m     20
 Deep Stop:           39m     1.0 min    44:34     15/55     25m     28
 Deep Stop:           36m     1.0 min    45:34     15/55     24m     36
 Deep Stop:           33m     1.0 min    46:34     15/55     24m     45
 Stop:                30m     2.0 min    48:34     15/55     22m     57
 Stop:                27m     4.0 min    52:34     15/55     21m     65
 Stop:                24m     6.0 min    58:34     15/55     18m     69
 Stop:                21m     4.0 min    62:34     50/0      15m     70
 Stop:                18m     4.0 min    66:34     50/0      13m     69
 Stop:                15m     7.0 min    73:34     50/0      11m     75
 Stop:                12m     9.0 min    82:34     50/0      8m      75
 Stop:                9m      14.0 min   96:34     50/0      5m      77
 Stop:                6m      53.0 min   149:34    100/0     0m      81
 Surface:                                149:34                      85

 Deco Time (21-6) = 91
 Estimated Total Runtime (including switches and gas breaks): 177 mins.
=========================================================================

This reduces many of the oddities of the normal GF implementation and works well in most situations. I would argue that deep stops are not bad in and of themselves as a principle, but that excessive deep stops are pointless and counterproductive.

I'll have a think about how to implement linear off-gassing.

Bit of thread drift going on here.

That is not the algorithm’s fault, that is a poor choice of gases given the algorithm.

Setting GF per compartment. That Bühlmann bloke, what did he ever know about deco eh?

 

[huwporter]

That is not the algorithm’s fault, that is a poor choice of gases given the algorithm.

Well yeah, to repeat what I said elsewhere, that was an example designed to illustrate an extreme case, not a recommendation for how to plan that dive. I think it's still a valid observation even for less extreme cases. See also: 10,000 forum posts about how deep stops have probably been overemphasised in the past.

(Also, that is kind of a weird response - do you suggest picking your deco gases to work around quirks in your algorithm? Particularly when it is kind of obvious that there isn't any real physiological basis behind it, it's just a mathematical quirk?)

Setting GF per compartment. That Bühlmann bloke, what did he ever know about deco eh?

Well, Bühlmann didn't talk about gradient factors at all, that was that Erik Baker bloke.

Since GFs are already a graft on top of Prof. Bühlmann's work, why not discuss other approaches of grafting...

 

[Jay]

Updated list. Adding Garmin Descent.

GFLo basically does not affect NDL on a NDL dive
Shearwater (all)
Dive Rite's discontinued Nitek - Q (post#2)
Deep 6 (post#93 and via PM)
Garmin Descent (from a Descent diver via PM)

TDC-3 <-- very provisional (see post#69)

GF Lo affects NDL on a NDL Dive
none yet

 

[dmaziuk]

@huwporter : If you are willing to get out of the water with your inert gas loading at 85% of the way to its M-value (in your leading compartment, of course), why are you not willing to let your body get up to 85% of the M-value on the way up to your first stop?

Because the model is built on the assumption that fast tissues can tolerate much greater overpressure that the slow ones.

That's what happens when you learn stuff from reading code: you gain very deep, detailed understanding of how it is done. Not so much of why or what for.

 

[stuartv]

Because the model is built on the assumption that fast tissues can tolerate much greater overpressure that the slow ones..

Isn't that why they have different M-values?

 

[scubadada]

Isn't that why they have different M-values?

Exactly

 

[dmaziuk]

Isn't that why they have different M-values?

Yes, and if you think your leading compartment at the first stop does not need more protection you set your GFs to 85/85. That is up to you.

But what GFs were designed for is ("Confusion about deep stops", page 2, "illustrating the problem"):

... complete decompression profile calculated by the conventional method. In this profile, the fastest compartments have the greatest gas loadings during the initial ascent and are leading. The M-values for these fast compartments permit large overpressure gradients relative to the slower compartments. Consequently, a large and rapid overpressure gradient is created during the ascent to the first stop. This is out of proportion with the smaller overpressure gradients permitted during the rest of the decompression profile ...

The reason he invented GF Low is to add more conservatism to this initial phase of ascent; you are not willing to let your body get up to 85% of the M-value on the way up to your first stop because you subscribe to his reasoning and believe your fast compartments need extra protection.