Tec Dive computers and dive plans

[Wibble]

For the sake of clarity, i am in no way suggesting that having a dive computer that can vector you in to your preplanned profile means you have to either

1) look at it
2) follow it


All it does is give you a frame of reference. Like having a speedometer in your car, most people simply drive at a speed that "feels about right", very few drive along looking at their speed continuously, but sometimes, it's rather useful to have a known reference instrument where hitting a target becomes more important.

I'm already using multideco for a reference. My deco PC software is calibrated and compared against this, as it's a proven reference, and my embedded deco code (running on an std 32bit ARM development board) is validated by using a software in the loop type approach against the PC software referenced against multideco

Can't help thinking that your plan is to, errm, planny. Dogma's not a great thing in mother nature. Unless you're diving to utter extremes -- bloody silly mega deep bounce dive to pick a tag off a line and spend 18 hours decompressing -- you simply don't need that amount of detail.

The only things you need are the MOD for your gas (which if it's OC it'll be specific for that depth); the minimum gas required after which you drown; the max time to surface which constrains your deco gas requirements (including factoring in failures); and the max runtime (after which the skipper bans you from his boat).

It's that old adage; the first casualty of a battle with the enemy is the plan. Too much planning == too rigid an approach. If you don't have the skills for a specific dive, then don't do that dive, do it after you've gained the skills.

 

[mc42]

The Bühlmann model, through Erik Baker’s implementation is what I’m familiar with, so I’ll give your deco stops question a shot.

As you've probably found, the original algorithm calculates the ceiling based upon the current compartment pressures, and only when the ceiling rises above the next integral stop depth does it plan an ascent. It doesn't take into account the future pressures from the ascent so the ceiling will have risen further due to the off gassing during the ascent.

I believe Subsurface does this particular ascent optimization to help you clear the ceiling just in time. Extra time spent deeper would in theory lead to slightly longer shallower stops, and I strongly suspect that the first stop depth in deeper dives could also be affected due to the ascent times to the first stop, which may disappear entirely by the time you get there, but is inserted by the original algorithm nevertheless.

I haven't quantified this effect other than noticing differences in shallow stops by a minute or so for the small tech dives (~30 min total deco) and an extra 10' deeper stop (a 200' dive generates a first stop at 110' vs 100' in Subsurface) on some profiles. But this may just be small implementation differences.

On to loading profiles into the computer...
It’s an interesting idea to be able to load a baseline plan then be able to see the variation on the computer. I would think the computer would need to essentially track three plans: my baseline plan, the current plan or how well I'm tracking to the baseline (an electronic slate), and some sort of bailout or normal dive computer plan to act as a safety net.

Tracking all of this in a usable interface with so many human factors considerations is the more complex issue here. You'd need to ensure you didn't load the wrong plan if you had planned multiple dives, for example. Other questions I have:

What happens when you go off the plan? Practically speaking, and others have alluded to it, few of us are probably executing the plans perfectly, like sticking to a 60fpm descent all the way down to 150' or 30fpm ascent. You deviate almost instantly. It's not enough to know I'm 5 minutes off the bottom time. I need to know what the impact is to the plan - do I now need 5 minutes more at 20'? Or it is a wash because I spent the time at 120' checking out something cool instead of at 130'? Will the computer be nagging me to get back down to 130' even if I don't want to? Would it be able to tell me that my original plan is no longer safe even if I got back on track after some big deviations?

I could support an electronic slate or notes viewing mode without any of the tracking features. I could see flipping to it as a mental reminder or backup, then flipping back to the regular computer mode. Maybe even dedicating a line on the precious display are for what depth I should be according to the time but I don't know for me if it's worth the sacrifice of displaying other more relevant information. Any more than this might be too complicated to handle at depth.

There is a potential danger of relying too much on the technology for this and turning off parts of the brain that should be fully engaged during the dive rather than handing it all over to a device that could fail. The biggest danger to a tech diver isn't the inability to follow a plan precisely to the minute but things like switching to the wrong gas, running out of gas (through bad planning, loss of reference, etc.), or blowing through deco entirely.

 

[MaxTorque]

The real time deco is still always calculated for the actual profile you have dived, ie with your actual depth. Stops are therefore set by this calc and not by the plan.

However, the plan deco, which is pre-calculated by the same algorythm that is running in realtime on the dive comp, should match the actual deco if you dive close to the plan. A simple + and - display would give you your depth delta to the plan depth at this moment (based on a dive timer started by wet contacts and say 1m depth being sensed, and the planned depth profile being held in memory in say 10 sec intervals)

It may make sense to also include some form of tissue tension comparison, perhaps a delta to surfacing GF, as per the plan, and as per the real time calc. That means at any time you can see if you have a higher or lower inert loading than your plan. Given that you make a plan with an appropriate amount of conservatism based on personal prefernce and experience, this would tell you simply, through out the dive if you are more or less risky.

On other potential safety advantage is to allow the plan to prompt for gas switches. and potentially to alarm / signal at the switch point.

Again, all that really falls under "useful info" because at no time does failing to follow the plan change the real time deco obligation calculated and displaid. TTS, Surfacing GF, integrated tissue tension, all can be used as normal, only now you can see the value as planned.


As i have said throughout this thread i have no idea if this is actually useful, but mearly suggesting that as dive computers gets smarter, perhaps we as divers should get smarter. Having the planning s/w be able to generate a paper plan also seems sensible, for use in kit setting up, gas filling, predive briefing and support crew briefing. Just give the dive plan print out to your boat skipper and he has all the information as to your basic plan, rather than try to explain in words what you are doing etc


The reason i came into this was because i wanted to explore the way decompression is modelled and calculated, but also because it seems there was somewhat of agulf between the planinng and the diving the plan. Ie i couldn't find any dive computer that had any more than a basic planning capability, and yet plenty of PC software exists to do just that, but in an unconnected way to the embedded computer we carry on all our dives !

 

[MaxTorque]

In terms of display space, which is always critical, the plan delta prompt could literally be a square block of pixels for a + or a - to show which side of the plan you are on at the current time. A second screen with detailed info could then by implemented for divers that need to reference the plan, to check a gas switch or similar, accessed by pressing the change display button as normal

 

[dmaziuk]

As you've probably found, the original algorithm calculates the ceiling based upon the current compartment pressures, and only when the ceiling rises above the next integral stop depth does it plan an ascent. It doesn't take into account the future pressures from the ascent so the ceiling will have risen further due to the off gassing during the ascent.

The fundamental problem is at the point where it calculates a ceiling, the ascent hasn't actually happened yet. You can't compute the next deco obligation until the ascent with all its on- an off-gassing is done.

That works well enough on a "very deco" profile with several stops and significant decompression times, where off-gassing (*) during ascent to the next stop is not significant relative to existing deco obligation. The first stop will take care of the fast TC and for the rest of them your ascent time to the next stop should not be significant relative to the controlling TC's half-time.

It gets pretty wobbly in "the grey area" where you transition from no-stop to yes-stop dive and/or back, though. It can be particularly problematic on deeper no-stop bounce dives where fast tissue is controlling all the way (first stop is the surface) and "safe" ascent time is of the same order of magnitude as that TC's half-time (4-5 minutes) -- i.e. off-gassing during ascent can be sufficient to clear the ceiling before you reach it.

*) and on-gassing too, provided you're not doing something stupid like GF 05/150.

 

[boulderjohn]

Just to go a bit off topic, I am truly amazed as I read this thread because I have been reading threads on ScubaBoard about planning tech dives for many years. If you haven't been around that long and have lots of spare time on your hands, seek out a thread on this topic from, say, 7-8 years ago. You will see a very different tone and a very different consensus. A few of the same people might be involved, and you will see some of them (including me) with a different mindset. The participants in this thread have been nearly unanimous in a way that would have been hooted down years ago.

 

[MaxTorque]

You see the effects of descrete stops (currently set at 3m seperation) in the model, where the off-gasing during the relatively rapid ascent is enough to occasionally just push the thresholds down and force a short stop 3m lower, and actually significantly below the limiting M value tissue tension

See attached pic, black arrow points to first stop being lower than the limiting M value threshold, then sorts itself out on the next stop as the off gassing raises the ceiling. The tissue tension line curves downwards slightly during the ascent and hence the predicted intercept with the supersaturation limit line doesn't actually occur

 

[dmaziuk]

I doubt it's about off-gassing during ascent, my guess would be the next 3m up is "too high": your stop is not at M-value, it's at the nearest x*3m below. Unless of course you have the exact same profile with slower ascent and the first stop 3m up.

Buhlmann in Decompression - Decompression sickness wrote that 3m separation is likely an artifact of practical difficulty of holding the exact depth in open water, and that in a chamber one could go for 0.1 atm spacing instead, or even "ride" the M-value line to maximize decompression efficiency.

 

[PfcAJ]

as dive computers gets smarter, perhaps we as divers should get smarter.

That’s not how it works. Not even a little.

 

[Blackcrusader]

What plan? Max operating depth, Min gas, max TTS, total runtime. That's it. If you can't remember those four numbers, do question why you're diving. The computer does the constant calculation of the TTS. Backup computer does it too. This isn't the 1990's or naughties. We have Shearwaters now.

Well I was 1980's trained with J valve watch and tables and brought a slate to write on for the deco stops.

Yeah we have our shearwaters now or at least I do. No diver will ever match the computing power of the shearwater on the dive in real time which is why we use dive computers. Does it make me lazy, no. It makes me pay attention more than when I dived using tables as so much more information at hand, surf GF, N2 loading, CNS PPO2 TTS using gradient factors you want....