Computers that don't lock out tangent.

[loosenit2]

I’ll just offer some observations from the sidelines:

Violation/ no violation is binary. The real world of decompression is analog.

When a deco violation occurs, all that means is that for the particular settings and the particular algorithm, at that point the computer determines the risk level has passed over some threshold, set by the manufacturer and/ or the algorithm as acceptable for their market.

This doesn’t mean that the computer has gone “tilt”, overrun anything, or lost its ability to calculate a revised schedule. If any of the available conservatism settings (in our case these include age, workload, risk level) were set differently, or if the spacing between stops were set to a different value, or if the computer included additional risk level options (as, effectively, Buhlmann with GF does) the computer might deliver a result that is below the risk level selected and not show a violation.

A revised schedule might include additional time at shallower stops. Obviously, in the case of an extreme violation, the shallowest possible stop might end up being below the diver’s current depth. We can calculate these schedules consistent with the algorithm to a point, but the risk level increases with with the scale of the violations. We do warn, post dive for 24 hours, with a prominent and unavoidable page that must be acknowledged to reach the dive screen, but leave it up to the diver to determine how they handle any violation- re-descent, continuing diving, or seeking aid.

That's just our approach for our intended customers. Locking out post dive after a violation is not an unreasonable approach, but it's not our choice. I find it hard to understand how locking up during a dive improves diver safety under any circumstances.

-Ron

Thank-you for chiming in from the sidelines. This is an eloquent and well reasoned methodology that I can absolutely support.

 

[RonR]

Doesn't mean it didn't either. We assume that under normal circumstances those "garden variety" violations will not push a device out of bounds, but we can't actually know without knowing the exact circumstances and the exact code running inside the computer in question at the point when violation occured.

I think it is very unlikely that any computer on the market is not capable of calculating its algorithm past violations, these are pretty simple functions. However, past some bounds the algorithm will get into territory where it is not validated as to risk- or where risk is clearly unacceptable. But I don’t think that is what most here are talking about. The decision about what information to display when the risk level envelope is exceeded is basically a user interface choice. So long as people understand what their computer does, there are multiple reasonable solutions.

Doesn'tIf you missed a mandatory deco stop and are continuing on your dive because you know better, then you obviously don't need no stinking computer. Why should the computer keep computing when it's obviously not needed? A computer, locked or not, is useless at this point, so who cares if tries to figure out exactly how bent you are or goes playing Global Thermonuclear War with itself from here on.

If you miss a stop, or fail to complete one for some reason, then of necessity your dive is continuing- since you still need to surface. Our feeling is that even if (by not completing a stop) a diver has pushed past the risk level set by the computer and created a violation, that providing them with the best schedule to surface is preferable to just cutting them off during that dive, and providing only gauge information.

-Ron

 

[KenGordon]

If you miss a stop, or fail to complete one for some reason, then of necessity your dive is continuing- since you still need to surface. Our feeling is that even if (by not completing a stop) a diver has pushed past the risk level set by the computer and created a violation, that providing them with the best schedule to surface is preferable to just cutting them off during that dive, and providing only gauge information.

While that is a fine position to take, the consequences for “cutting them off” are not actually severe, the diver will be shallow and simply has to fallback to the next plan which is breathing what gas they have.

The consequences of enabling a diver to get back in the water after a dive that might be bending them I think are worse, and the chances of that seem much greater to me.

To be clear, on a Suunto ascent with stops at 30m, 24m, etc if you miss the first stop then you will still have a working deco schedule. The computer is not giving up unless you violate the hard ceiling for a while. That hard ceiling is actually much shallower than the 30/70 or even 45/85 profile will give.

I know this because I dive Suuntos at the same time as Shearwaters set to GF and understand that Suunto RGBM is a marketing thing. With a real RGBM model, such as yours, or other deep stop models like GF then I agree that loosing the computer because you cut 30 seconds at 30m would be excessive, I don’t think that actually happens though.

 

[rongoodman]

So, to generalize a bit, you want the computer to keep computing you "to safety" after its little brain has been overflown and pushed out of bounds?

Why are you assuming that a missed deco stop means the computer has been "overflown and pushed out of bounds" instead of just having decided that the calculated inert gas load has moved into dangerous territory, especially when discussing computers running proprietary algorithms? There is simply no way of knowing what they're doing, or why. Even Suunto's high end EON Steel will lock itself out and whatever else you might say about it, "little brain" probably doesn't apply.

 

[dmaziuk]

I think it is very unlikely that any computer on the market is not capable of calculating its algorithm past violations, these are pretty simple functions. However, past some bounds the algorithm will get into territory where it is not validated as to risk- or where risk is clearly unacceptable. But I don’t think that is what most here are talking about. The decision about what information to display when the risk level envelope is exceeded is basically a user interface choice.

Why are you assuming that a missed deco stop means the computer has been "overflown and pushed out of bounds" instead of just having decided that the calculated inert gas load has moved into dangerous territory, especially when discussing computers running proprietary algorithms?

Because to me these sentences straight out of the Zoop manual sound like it actually can't compute its RGBM bubble voodoo once "the micro-bubble build-up" hits some kind of ceiling:

when the ascent rate exceeds 12 meters/min [40 ft] momentarily or 10 meters/min [33 ft] continuously the micro-bubble build-up is predicted to be more than allowed for in the decompression model. The Suunto RGBM calculation model responds to this by adding a Mandatory Safety Stop to the dive.

And y'all may have noticed that one or two persons here actually diving Steels claim to have never locked them up in hundreds of dives, and the only people complaining about Dread Lock-outs are talking lower-end devices. Gotta wonder if there might be a correlation of some sort here.

 

[Ghost95]

Morning all.

Wanted to say thanks for the replies. Interesting reading for sure. I do appreciate the opinions and they do show both sides of the argument, uh, I mean discussion. Reading through all the posts would allow someone to see the issues and make that decision based on their needs.

Seriously, thanks for the feedback.

Maybe for my next question I'll resurrect the issue of how bad is it, really, to regularly overfill a current VIP'd and Hydro'd steel 72 from 1963 to say 2750-3000 psi. Should be a general consensus right?

Good day all. Safe diving.