ceiling/GF

[lermontov]

this has been a very informative thread id like to hear from @Dr Simon Mitchel to correct any errors in our thinking

 

[KenGordon]

this has been a very informative thread id like to hear from @Dr Simon Mitchel to correct any errors in our thinking

We are only arguing over the meaning of GF and how the interpolation ought to work. These are minute details of the implementation which don’t make a significant difference and rather below Simon’s pay grade. It is actually the stuff that Ross would be able to help with .

However, if people want to put out graphs which are supposed to aid understanding then they ought to be correct or the literal interpretation of the graph will not match what is happening.

 

[KenGordon]

@Shearwater , I'd be interested in your comments. It was my impression that having set a GF of 30/80, the computer runs a continuous determination of the GF Lo ceiling, giving you a first stop depth at the closest rounded depth below a GF of 30 (in this example) and interpolating that on the fly to generate your 80% GF Hi surfacing overpressure. In other words, the Perdix will compute a real-time 30/80 plan based upon your max depth/time theoretical tissue loading. Why can't they do that?

A planning program can consider the whole dive and make assumptions about diver behaviour. In particular it can say that the diver will be at the first stop when planned and the ascent will be the planned ascent. If the diver is later then there will be another plan for that contingency.

So a planner has a well defined first stop at witch to start the interpolation between GF lo and hi.

A dive computer can only look at what has happened and guess what might happen. To create the ascent profile it has to start with a GF lo somewhere. Apparently they will use the deepest ceiling during the dive for that (you can read the OSTC code, or the derived code in subsurface for that). An alternative would be to use the depth when the ceiling is reached. However the diver does not want to reach the ceiling but stay below it. So, chose some arbitrary ‘zone’ below the ceiling which means you are at the first stop? Now what if I ascend gently on a sloping site? I am never near the ceiling so never at the first stop so where should GF lo apply? My first stop might be quite shallow. Do I want to be at 9m with a GF max of 30 when the planner would have had me at a GF max of 65 or so?

So, to keep things simpler it is assumed that the first stop, for the purposes of GF lo, is the appropriate one (eg rounded to the next deeper multiple of 3m) for the deepest ceiling.

However the actual first stop depth is likely to shallower due to the profile.

This will only make much, if any, difference for a planned multi level dive.

 

[doctormike]

I think that we are all getting hung up on terminology, which is good because this discussion helps optimize the graphic (thanks to @Shearwater for being so responsive!). This drawing really does help people understand decompression and GFs. But it’s a cartoon to explain a concept, it’s not a way of precisely measuring specific pressures. The lack of units (as @rsingler pointed out) means that any slope is arbitrary, and is drawn for clarity, not for precision.

The slope of the ceiling line is determined by the slope of the M-value line, because there are two things that change as you ascend: Allowable overpressure (goes from 30 to 80 over the course of the ascent) and the actual M-value (maximum overpressure, which is different at different depths). So the distance from the AP line on the drawing for an overpressure of 30% will be different at different depths - the M-value function is linear but not parallel to the AP line. This diagram is an oversimplification, since there are really 16 lines (and even that is arbitrary). It is just to help with visualization.

So my point was that it’s not necessary to define about the angle of the ceiling line before the first stop, because the ceiling line is just a shorthand way of defining your dive plan and first stop. Of course, you HAVE a ceiling before the first stop, you do at every point in the dive once you have exceeded NDLS. But even then, the ceiling is generated by calculating your first stop if you were to ascend at that moment. That is, the ceiling doesn’t really exist until you reach it, because it would change if you delayed your ascent, descended, switched gasses, etc….

GFLo refers to a stop depth, that is how it is defined, and that (with GFHi) is what generates the ceiling line. I might be wrong, but I just don’t see how the ceiling line has meaning before the first stop, since it’s defined as being from GFLo to GFHi. It’s sort of like saying “what is the path of the race before the starting line?”.

@KenGordon, you are absolutely right to say that maximum depth isn’t a relevant thing because 30% is for the first stop only. It was used, however, to define the right side of the chart, and the chart was designed to depict a square profile, single gas ascent.

 

[J-Vo]

I have always thought that inverting both axis of this chart would have accelerated my understanding of what it is describing. It would more (graphically) represent how we (physically) interact with the model as it is.

 

[rsingler]

Now what if I ascend gently on a sloping site? I am never near the ceiling so never at the first stop so where should GF lo apply? My first stop might be quite shallow. Do I want to be at 9m with a GF max of 30 when the planner would have had me at a GF max of 65 or so?

NOW I see why you have been so persistent on this issue. You have an interesting point! So let's examine that.
A guy plans his dive for a GF 30/80 profile from a certain point of ascent. He decided that his best safety lies in following that offgassing profile (which I don't necessarily agree with, but that's what's in the sample Shearwater graphic). But as you suggested, instead of following his profile, he gets interested in photographing the little crabs higher and higher along the sloping bottom, and stays well below GF30 for a significantly greater part of the dive, despite ascending. He doesn't go below rock-bottom remaining gas as he slowly ascends, so he's content to vary from his planned profile.
Finally, he decides to head up. What should his ascent profile be? You have suggested that since he is significantly shallower, he should be content with a higher initial GF Lo, which the computer won't have computed unless he changes his gradient factors on the fly. I'm not sure that's the safest way to proceed. Since he may have continued to ongas into slower tissues he now has a completely different dive, for which the computer will implement decompression (at 30/80) as soon as the diver reaches this new first stop. Why should that stop be greater than GF 30?
Conversely, if his meandering ascent (below GF 30) has allowed him to offgas along the way the computer shortens times at his first stop, or eliminates them altogether on the fly.
I'm not saying that this sort of deviation would be smart, or safe from a required gas standpoint. But for the purpose of discussion, when the diver finally starts his ascent I think he should still follow a GF 30/80 profile rather than jump immediately to some higher ceiling simply because he is at a shallower depth. It is the gradual progression from GF Lo to GF Hi on which we predicate a safe offgassing progression. The heat maps that Dr. Mitchell has used in his presentations, and which are displayed nicely in Subsurface are helpful reminders in this regard.
Conversely, how will a diver with a depth gauge and timer and a sheaf of plan variances in his pocket pick the correct ascent strategy? Are you really suggesting that this diver who finds himself now at a depth above his planned first stop due to a meandering ascent just jump to the nearest next stop in his decompression plan? I don't think that's what you're suggesting.
For me, the takeaway in what we have discussed here is the beauty of having a computer-generated decompression profile that responds to changes in the planned dive. And when the diver finally recognizes his knucklehead move in getting distracted and slowly ascending rather than following his plan, I would think that he should follow his planned gradient progression, even if it starts from 9m. He won't have very long stops in any case, if he really has been offgassing with his slow ascent. What seems more likely is that he will have significantly ongassed in slower tissues by not following an efficient ascent profile. In that case it seems all the more important to stick to his chosen gradients. Whether he'll have enough gas to surface becomes the obvious next question.

 

[rsingler]

However, if people want to put out graphs which are supposed to aid understanding then they ought to be correct or the literal interpretation of the graph will not match what is happening.

So if they eliminated the dotted line to the right of the first stop, would the graphic be more correct? I think that's all we're debating here.

 

[doctormike]

1) Dive planning software - a tool that generates an ascent strategy based on preselected mix, depth, time and gradient factors, assuming either a square profile or a multilevel dive with relatively few discreet segments. Useful as a "worst case" (i.e. maximum N2 loading) plan, so that you are sure to bring along enough gas.

2) Dive computer - a tool that continually tracks N2 loading in real time and continually generates a safe ascent strategy for any given point in the dive.

3) This diagram - a tool that helps teach new tech divers how decompression works, and how gradient factors can be used to modify an ascent plan by adjusting the tradeoff between DCS risk and decompression efficiency. It is based on an assumption of a square profile, a single gas, and a single compartment model. Any attempt to apply it to more complex dives will result in confusion.


I agree with @rsingler - I also don't think that you should bump the GFLo on the fly just because you have made a slow initial ascent. The algorithm is complex, and thinking that you can juggle the various risks and inputs in your head seems risky. I get what you are saying, Ken, but it just seems like normalization of deviance. Either let the computer do it's thing, or just follow your dive plan. If you want "credit" for the slow ascent staying below the ceiling, are you also tracking in your head the ongassing of any slower compartments that might have happened while looking at the crabs?

 

[doctormike]

So if they eliminated the dotted line to the right of the first stop, would the graphic be more correct? I think that's all we're debating here.

I think that it should be eliminated, as I don't think that it has meaning before the first stop.

Ken, correct me if I'm misinterpreting your post, but I think that you said that it should extend to the right of the first stop but with a different slope to reflect the ceilings that are continually generated between initial ascent and the first stop?

 

[Doppler]

Well, this has been a thoroughly interesting thread to read through. Thanks all!