Dive computers... SO many choices!

[atdotde]

I just realized I'm a doofus and what I said above is dumb. There is no interpolation required. At the point you go into deco, the depth at which you exceed 30% (or whatever your GF Lo is) of the M-value (in your leading compartment) is whatever it is. It can be calculated directly, without regard to what GF Hi is. And, the only way your depth would affect the calculation is if you choose to factor ascent time into the calculation.

This is indeed what subsurface does: Through the dive, it tracks your ceiling assuming GFlow as the relevant percentage of the M-value. The deepest value is taken as "the first stop depth" to anchor the GF interpolation line. The problem in taking ascend times into account is that if you constantly stayed below the ceiling (be it by only one foot say), that would effectively be the same as an arbitrary slow ascent which means you would have to apply GFlow at the surface (since you never stop but only ascend very very slowly).

 

[stuartv]

This is indeed what subsurface does: Through the dive, it tracks your ceiling assuming GFlow as the relevant percentage of the M-value. The deepest value is taken as "the first stop depth" to anchor the GF interpolation line. The problem in taking ascend times into account is that if you constantly stayed below the ceiling (be it by only one foot say), that would effectively be the same as an arbitrary slow ascent which means you would have to apply GFlow at the surface (since you never stop but only ascend very very slowly).

If you are constantly recalculating, and your calculation starts with "if ( direct_ascent_surface_GF <= GFHi ) then clear_all_deco();" would this still be an actual problem?

 

[KenGordon]

Same dive, same ascent. Since you would actually not hit GF 30 until 90' (per the specifics of the example), that means that, with your way of doing it, when you arrive at 100', you will be less than 30, so your stop would end up being for 0 time. It would clear before or as soon as you get there. So, you would actually ascend non-stop on up to 90' (no matter which way you calculate the first stop). And, at 90', the GF value being used to determine your stop time is GF 34, instead of GF 30 (if done my way).

There are two issues. One is how to calculate the ceiling. The other is where you anchor the gf lo. The second is an implementation choice. The first isn't really if you are wanting to claim to be following Erik Baker.

Assuming you arrive at the projected first stop and it has moved shallower, but you anchor the GF lo at the stop you never made then yes, by the time you get to an actual stop you will have a higher GF and be able to move up sooner.

On the other hand if you put the anchor at the first actual stop, when you cannot ascend without exceeding GF lo, (the way the diagrams show) then if your ceiling is shallower you can move up sooner and are on a more aggressive profile.

This impact of this behaviour will also depend on the gap between GF lo and hi and how deep the first stop is.

I am pretty certain that MultiDeco uses the Baker method of assuming an instantaneous ascent. My own software can do either but agrees with MultiDeco when doing it that way rather than allowing for the projected off gassing. Looking at the SubSurface code, and I think confirmed by the author, it anchors at the first projected stop. At first glance it looks like it uses the projected ceiling for GF but the instantaneous one for VPM.

GF is a scheme to mess with the conservatism of the ZHL16 a/b coefficients to ape the bubble models. You could choose to change the a/b coefficients themselves to a similar end but without all the head scratching about implementation details.

It would be helpful if the manufacturers of the computers provided planning software to match their implementation choices.

 

[dmaziuk]

It would be helpful if the manufacturers of the computers provided planning software to match their implementation choices.

Oh, come on. Dive equipment manufacturers are not software houses and I would take a GitHub project over any of them any day. What would be nice is if they provided detailed documentation of their implementations, or the source. And Deep 6 is the only recreational computer where that may actually happen. Right in this thread, if you manage to talk @LandonL into it.

 

[stuartv]

There are two issues. One is how to calculate the ceiling. The other is where you anchor the gf lo. The second is an implementation choice. The first isn't really if you are wanting to claim to be following Erik Baker.

I posed a simply yes or no question, which you seem to be studiously avoiding.

The only "issue" is whether to calculate your first stop the Baker way - i.e. the calculated ceiling is the first stop - or if you calculate it by factoring in the off-gassing that occurs during a 30ft/min ascent. GF Lo is anchored at the calculated first stop either way.

The example I gave is a dive using GF 30/70. At the start of your ascent, the depth of GF30 (i.e. your ceiling) is greater than 90', but less than or equal to 100'. So, your calculated ceiling is 100' and your Baker first stop is also 100' and your GF Lo anchor point is 100'.

The additional stipulation of the example is that during your ascent to 100' you would off-gas enough that by the time you get to 100', the depth of GF30 is now less than or equal to 90' (but greater than 80'). Perhaps it was only at 91' when you leave out off-gassing during the ascent (which gets rounded up to a stop at 100'), but it would change from 91' to 90' as you ascend. So, including ascent time in the calculation results in a first stop of 90', which becomes your GF Lo anchor point*.

The actual ascent, in this example, would be directly to 90', no matter which way you calculate the first stop/anchor point. The only difference is that with your (the Baker way) of calculating the first stop, you would spend less time at 90' before you ascend to your next stop, at 80'. With my way (factoring in ascent time in the calculation of the first stop), you would stay at the 90' stop longer.

Here are the GF gates for each depth:
Depth Your Mine
100 30 --
90 34 30
80 38 34
70 42 39
60 46 43
50 50 48
40 54 52
30 58 57
20 62 61
10 66 66
0 70 70

So, using either method, the diver would actually ascend directly to 90' before making their first stop. From there on, your diver would have a higher GF in his leading compartment at the end of each deco stop where he or she proceeds up to the next shallower stop. Your way gets the diver shallower, faster.

And, again, the simple yes or no question: Do you still think that the Baker way of calculating the first stop is more conservative, as you said earlier?


* Sure, if the dive specifics result in a difference in the first stop that is still in the same stop interval, then the two methods produce identical results. E.g. the Baker method calculates a first stop of 99' and my way calculates a first stop of 91'. In that case, both round up to a first stop and GF Lo anchor of 100' and from there the results are the same. The point is the difference when factoring in the ascent time causes the first stop depth to raise up into the next higher stop interval - e.g. from 100' to 90'. I assert that including the ascent time makes it the same or more conservative - not more aggressive.

 

[stuartv]

Oh, come on. Dive equipment manufacturers are not software houses and I would take a GitHub project over any of them any day. What would be nice is if they provided detailed documentation of their implementations, or the source. And Deep 6 is the only recreational computer where that may actually happen. Right in this thread, if you manage to talk @LandonL into it.

Heinrichs Weikamp has been doing that for some time, with their OSTC computers. Their computers are open source. I have not seen anything from Deep 6 that says the code for their computer will be open sourced.

 

[stuartv]

The problem in taking ascend times into account is that if you constantly stayed below the ceiling (be it by only one foot say), that would effectively be the same as an arbitrary slow ascent which means you would have to apply GFlow at the surface (since you never stop but only ascend very very slowly).

Another thought on this: Correct me if I'm wrong, but you're saying the problem is what if, for example, the diver is using GF30/70 and they ascend slowly, so that they never actually get to a peak GF of 30. Then you're saying the GF30 would have to be applied at the surface.

I don't think that makes sense. GF Hi is always the anchor at the surface. So, the shallowest that GF Lo should ever be anchored is 10'/3m. That should be the boundary case. If the diver comes up so slowly that they don't exceed GFLo at 10', then the only factor would be to give them a mandatory stop at 10' if they will exceed GF Hi at the surface.

 

[KenGordon]

Oh, come on. Dive equipment manufacturers are not software houses and I would take a GitHub project over any of them any day. What would be nice is if they provided detailed documentation of their implementations, or the source. And Deep 6 is the only recreational computer where that may actually happen. Right in this thread, if you manage to talk @LandonL into it.

I think a dive computer is a software product. They ought to be able to do it. It needs to have downloads that work, sometime uploads. Sadly the all seem to be trying to build cloud/social media empires.

Whatever, if they want to claim that the computer implements ZHL16C+GF it would be nice if they made clear what that means.

 

[scubadada]

I think a dive computer is a software product. They ought to be able to do it. It needs to have downloads that work, sometime uploads. Sadly the all seem to be trying to build cloud/social media empires.

Whatever, if they want to claim that the computer implements ZHL16C+GF it would be nice if they made clear what that means.

Shearwater and Dive Rite implementation appears pretty clear, what more do you want?

 

[KenGordon]

I posed a simply yes or no question, which you seem to be studiously avoiding.

The only "issue" is whether to calculate your first stop the Baker way - i.e. the calculated ceiling is the first stop - or if you calculate it by factoring in the off-gassing that occurs during a 30ft/min ascent. GF Lo is anchored at the calculated first stop either way.

The example I gave is a dive using GF 30/70. At the start of your ascent, the depth of GF30 (i.e. your ceiling) is greater than 90', but less than or equal to 100'. So, your calculated ceiling is 100' and your Baker first stop is also 100' and your GF Lo anchor point is 100'.

The additional stipulation of the example is that during your ascent to 100' you would off-gas enough that by the time you get to 100', the depth of GF30 is now less than or equal to 90' (but greater than 80'). Perhaps it was only at 91' when you leave out off-gassing during the ascent (which gets rounded up to a stop at 100'), but it would change from 91' to 90' as you ascend. So, including ascent time in the calculation results in a first stop of 90', which becomes your GF Lo anchor point*.

The actual ascent, in this example, would be directly to 90', no matter which way you calculate the first stop/anchor point. The only difference is that with your (the Baker way) of calculating the first stop, you would spend less time at 90' before you ascend to your next stop, at 80'. With my way (factoring in ascent time in the calculation of the first stop), you would stay at the 90' stop longer.

Here are the GF gates for each depth:
Depth Your Mine
100 30 --
90 34 30
80 38 34
70 42 39
60 46 43
50 50 48
40 54 52
30 58 57
20 62 61
10 66 66
0 70 70

So, using either method, the diver would actually ascend directly to 90' before making their first stop. From there on, your diver would have a higher GF in his leading compartment at the end of each deco stop where he or she proceeds up to the next shallower stop. Your way gets the diver shallower, faster.

And, again, the simple yes or no question: Do you still think that the Baker way of calculating the first stop is more conservative, as you said earlier?


* Sure, if the dive specifics result in a difference in the first stop that is still in the same stop interval, then the two methods produce identical results. E.g. the Baker method calculates a first stop of 99' and my way calculates a first stop of 91'. In that case, both round up to a first stop and GF Lo anchor of 100' and from there the results are the same. The point is the difference when factoring in the ascent time causes the first stop depth to raise up into the next higher stop interval - e.g. from 100' to 90'. I assert that including the ascent time makes it the same or more conservative - not more aggressive.

Ok, you are right. The lower GF at the shallower stops restults in a shorter time to surface and generally shorter shallow stops despite spending some time ascending more slowly once the initial stop is hit. Another artefact is that 'your' way results in an initial stop longer than the following stops. In the conventional way the initial stops are all very short and likely to have disappeared when you get there. (Assuming a reduced ascent rate once the first stop is reached).