Deco for Divers: RGBM vs other algorithims

[Centrals]

I've been using DSAT since 2002, about 1,150 dives. Conservative and liberal is in the eye of the beholder, I'm perfectly happy with DSAT.

Out of interest, why do you think DSAT is too liberal? I would imagine it is because you compare it to your more conservative Uwatec primary computer. Perhaps the Uwatec is too conservative? I run an Oceanic VT3 primary and a Geo2 backup, they run within 30 seconds of one another, they are both just right. Perhaps you should run a backup computer that runs the same decompression algorithm as your primary computer?

I have no idea if the Uwatec Aladin is conservative or not. From my own experiences I have yet to come across any suggesting that the computer is conservative.

All proprietary means is that it's their secret and they have no obligation to tell it to you. Just like the recipe for Coca Cola is proprietary. If you figure out all on your own a way to produce exactly the same results, you are free to do so - whether it's making your own cola that tastes the same as Coke, or whether it's a dive computer that always gives identical NDLs to someone else's RGBM computer.



Hey Craig,

As you know, I have been diving DSAT (Oceanic Atom 3.0) also, for the last year and a half. More recently, I have also been using a Seabear H3, which is Buhlmann ZHL-16C and I usually use GF of either 30/70 or 35/85 (2 of the default options on the H3). I completed my deco cert 6 weeks ago and have done 16 dives since then, with most of them being deco dives. Most of the deco dives were just doing a few minutes on back gas. A few were with 5 - 15 minutes of deco on 80%.

On all these dives, I have been programming the H3 for all the gases I was using and doing a gas switch on the computer whenever (if) I switch to my deco gas. For all those same dives, I also leave my Atom in Dive mode (versus Gauge mode) and set the FO2 to match my back gas, but I have not been doing a gas switch on the Atom to tell it when I switch to my deco gas.

So far, I have only bent my Atom once. I have put it into deco a number of times. But, even though the Atom thinks I'm doing all my deco on whatever Nitrox I have been using for back gas, its algorithm (I use DSAT) has been clearing my deco in the amount of time it takes me to do the indicated stops that my H3 prescribes. Even when I'm actually doing my deco on EAN80.

As I said, except for once. The day I did my first AN/DP checkout dive, I used 50% for deco gas and, for whatever reason on that day the Atom actually went into Violation mode when I got out and locked itself for 24 hours (though it still showed Depth and EDT the next day when I wore it during my second AN/DP checkout dive)

My conclusion: The DSAT algorithm in the Oceanic computers is pretty darn liberal, even for its deco obligation calculations. OR, maybe I should say that previous comparisons of the Atom NDLs (in the owner's manual) to Buhlmann w/GF 100/100 in Multi-Deco have led me to think that the DSAT implementation in the Atom is straight up Buhlmann with no GF (i.e. GF 100/100). Maybe that is indeed the case. Maybe, for the dives I've done, GF100/100 and deco on back gas would yield a run time that is equal or shorter to the run time for GF 30/70 with deco on 80%. Maybe when I have some time I'll play with Multi-Deco and see.

I think your approach of clearing off the Geo's deco and then staying an extra 3 - 5 minutes is a really good idea - if not staying even longer.



How did you determine it was too liberal? Did you get bent?

I didn’t get bend. The Veo 180 has a lot more ndl available than Uwatec Aladin over the same dive.

 

[scubadada]

I have no idea if the Uwatec Aladin is conservative or not. From my own experiences I have yet to come across any suggesting that the computer is conservative.

I didn’t get bend. The Veo 180 has a lot more ndl available than Uwatec Aladin over the same dive.

There you go, DSAT is more liberal than Scubapro's ZHL8 ADT MB. Scubapro is actually middle of the road when it comes to the spectrum available.

 

[scubadada]

...Hey Craig,

As you know, I have been diving DSAT (Oceanic Atom 3.0) also, for the last year and a half. More recently, I have also been using a Seabear H3, which is Buhlmann ZHL-16C and I usually use GF of either 30/70 or 35/85 (2 of the default options on the H3). I completed my deco cert 6 weeks ago and have done 16 dives since then, with most of them being deco dives. Most of the deco dives were just doing a few minutes on back gas. A few were with 5 - 15 minutes of deco on 80%.

On all these dives, I have been programming the H3 for all the gases I was using and doing a gas switch on the computer whenever (if) I switch to my deco gas. For all those same dives, I also leave my Atom in Dive mode (versus Gauge mode) and set the FO2 to match my back gas, but I have not been doing a gas switch on the Atom to tell it when I switch to my deco gas.

So far, I have only bent my Atom once. I have put it into deco a number of times. But, even though the Atom thinks I'm doing all my deco on whatever Nitrox I have been using for back gas, its algorithm (I use DSAT) has been clearing my deco in the amount of time it takes me to do the indicated stops that my H3 prescribes. Even when I'm actually doing my deco on EAN80.

As I said, except for once. The day I did my first AN/DP checkout dive, I used 50% for deco gas and, for whatever reason on that day the Atom actually went into Violation mode when I got out and locked itself for 24 hours (though it still showed Depth and EDT the next day when I wore it during my second AN/DP checkout dive)

My conclusion: The DSAT algorithm in the Oceanic computers is pretty darn liberal, even for its deco obligation calculations. OR, maybe I should say that previous comparisons of the Atom NDLs (in the owner's manual) to Buhlmann w/GF 100/100 in Multi-Deco have led me to think that the DSAT implementation in the Atom is straight up Buhlmann with no GF (i.e. GF 100/100). Maybe that is indeed the case. Maybe, for the dives I've done, GF100/100 and deco on back gas would yield a run time that is equal or shorter to the run time for GF 30/70 with deco on 80%. Maybe when I have some time I'll play with Multi-Deco and see.

I think your approach of clearing off the Geo's deco and then staying an extra 3 - 5 minutes is a really good idea - if not staying even longer...

Hi Stuart,

You have it right, we're talking about 2 very different scenarios, no stop diving/not exceeding the NDL, and deco diving.

The DSAT algorithm is going to give you a lot more NDL than a conservative choice of GFs. My example was a 60 ft dive on air with a NDL of 57 minutes on DSAT. Simulating this dive on MultiDeco with GFs of 30/70, gives a deco obligation of 0:45 at 30 ft, 2:00 at 20 ft, and 17:00 at 10 feet. GFs of 100/100 give no deco time, 95/95 gives 1:55 at 10 feet, 90/90 gives 3:55 at 10 ft. So, for no stop diving, it takes something like GFs of 95-100 to yield a profile that looks like DSAT.

I don't really know how DSAT handles deco time. The algorithm was developed for no stop diving and deco is for extraordinary use. In my experience, and in a small amount of published data, the algorithm quickly adds shallow deco time when the NDL is exceeded. So, Buhlmann ZHL-16C, with reasonable GFs, will go into deco earlier than DSAT, but will generate a different deco profile. The result may be that the total run time may end up being reasonably similar under some circumstances, such as most of the dives you describe. For example, the same 60 ft dive on air for 57 minutes yields the following deco profiles: 45/95 gives 0:20 at 20 ft and 2:00 at 10 ft, 40/85 gives 0:20 at 20 ft and 7:00 at 10 ft, 35/75 gives 3:20 at 20 ft and 12:00 at 10 ft. It you paid enough attention to your 2 computers, especially when doing deco soley on your back gas, you could see if this is true. I intend on doing just this when I finally break down and get myself a new computer.

Good diving, Craig

 

[stuartv]

I didn’t get bend. The Veo 180 has a lot more ndl available than Uwatec Aladin over the same dive.

If you didn't get bent, then what is the basis for saying the Veo is way too liberal instead of saying that your other computer is way too conservative?

 

[afieldofblue]

I don't really know how DSAT handles deco time. The algorithm was developed for no stop diving and deco is for extraordinary use.

Sorry to resurrect an old thread (I'm reading up on the subject at the moment) but if all Oceanic computers are built on the same model then an Oceanic running DSAT switches to French MN90 tables for deco calculations (as stated in the Geo2 manual)

French MN90 tables are available here:
www.plongee-plaisir.com/fr/pdf/MN90_Mode_Emploi.pdf

Otherwise DSAT does give you longer NDLs between 18 and 12 m, as illustrated here:

Hope this helps

Cheers,

Ben

 

[ajduplessis]

Wanted to get a discussion started on this after browsing through Mark Powell's "Deco for Divers" and how RGBM is one of the newest deco algorithms... it seems to have its merits, even though I've also heard and read that it doesn't seem to have any PROVEN benefits over other models...

What is your preferred algorithm and why? RGBM - too conservative, or the latest in study?

Never had bad experiences with RGBM, VPM, Buhlmann/modified gradient factors or Deep stops in decompression diving. I am not talking about the 20min deepstop profile use by some clever people to make a point. The shorter 2min version used by thousands of divers. Deco is tool, use what works for you.

 

[MikkelBC]

Until recently I used a Scubapro/Uwatec Galileo Sol. Now I use a Shearwater Perdix AI.
The Perdix allows me to set conservatism with gradient factors.
The Galileo allowed me to set "conservatism" with "microbubble levels" (MB0 to MB5, I think). The problem I had with this and the main reason I chose a Perdix over a Scubapro G2 what that I don't know what these levels mean. I could not find any explanation of how these MB levels work. Consequently I had no idea how to make an informed choice regarding a suitable amount of conservatism.

 

[CAPTAIN SINBAD]

Buhlmann with gradient factors.