Boy I am lost on this. Maybe you could expand.
Both bubble modes (e.g. RGBM) and dissolved gas models (e.g. DSAT) calculate tissue loading exactly the same way. They set limits differently. DSAT sets pressure limits, RGBM also uses volume limits. AFAIK RGBM did no independent testing to come up with the model parameters used, but rather calculated the volumes that corresponded to the pressure limits in the other models to come up with a new type of limit (i.e they backed into values using available data). RGBM makes some volume adjustments for repetitive dives, sort surface intervals etc. That gives RGBM some capability to address novel situations, but it appears given the track record that the adjustments may be overly conservative. If there is anything that documents that different sets of data were used I would like to hear about it. Or at least that is my just so story. If it is wrong someone can point in the right direction.
Real "iterative" RGBM (and other bubble models like VPM) in fact do calculate decompression completely differently from the way DSAT or Haldanian algorithms do. However the Suunto version of RGBM is not fully iterative (nor is any other RGBM computer on the market, with one exception), it is "folded" RGBM- a Haldanian algorithm with various bubble factors added on to penalize receptive or multi-day diving, rapid ascents, or conditions that could create a greater likelihood of bubble formation. It's essentially a way of approximating the results of fully iterative RGBM while not being quite so computationally intensive.
At root, RGBM uses well known physics of bubble formation to come up with a schedule that prevents bubbles from forming. Conventional algorithms using M-values are designed to keep tissue saturations below certain differentials and to allow for reabsorption of gasses. Bubble models are derived from physics and have been well tested clinically, Haldanian algorithms are historically derived from clinical observations, In dive computers, both have had good success.
Because to the relative conservatism of Suunto computers, many people have the impression that the RGBM algorithm is inherently conservative- it isn't inherently
anything, it gives designers the ability to tune the conservatism to match statistical risk factors derived from clinical experience. Suunto chose to be fairly conservative, but that has little to do with RGBM. It's a choice made about how RGBM was implemented on a particular platform. We made somewhat different (but not dramatically different) choices when we designed the Cobalt computer, also using RGBM, with both folded and fully iterative implementations. About all you can say is that RGBM will generally give deeper stops and slower ascents, but it may well provide shorter overall decompression schedules. The issue of whether or not a computer is more "conservative" is a lot more complex and nuanced than simply charting no-deco limits.
I think the issue the original poster is seeing is really something else, which I will deal with in another post. Suffice to say that in a 60' dive, looking at no-deco limits, the signal to noise ratio between very slightly differing algorithms is very high, making them seem much more different than they really are.
Ron
---------- Post added ----------
I am writing this based on my experience with several different kinds of computers and their levels of conservatism, plus the reactions I've seen on the part of some SB users to degrees of conservatism. Recently I've used a Suunto Gekko, a Mares Puck and an Oceanic OC1 (DSAT) and have been able to compare the allowable bottom time with basic settings (no conservatism adjustments) and I see a pretty dramatic range. The Gekko is way at one end and the OC1 is at the other end with the Mares somewhere in the middle. Recently on the second of 2 approximately 60-foot dives on air, as the Gekko ran up against the no deco limit, the Mares was still in the upper single digits and the OC1 had about 20 minutes. I knew there were differences, but this seemed much more extreme than I expected....<snip>
So, my question for discussion is whether the proponents of conservatism believe that some computers are so liberal as to be unsafe? I've seen the analogy of walking up to a cliff, but how far back do you really need to be? Am I being reckless by leaving the Gekko at home and following the OC1 because it "gives" me longer dives?
The biggest factor in what you are seeing is due to the fact that you are doing a long, shallow dive that is not much below the depth where you would never get a deco obligation. This question comes up often regarding shallow dives or dives that are approaching a lengthy no-deco limit. Divers think they are seeing huge differences in algorithm conservatism, when in fact they aren't. It just looks that way. It has to do with the binary assignment of deco- no deco times that all dive computers do.
If you envision the tissue saturation as a rising line across a graph left to right, on shallow dives it has a very shallow upwards slope. That slope is defined by the algorithm conservatism. Exactly where a computer crosses the line into deco can vary a lot in minutes with a very tiny difference in the slope. So you might see a 10-15 minute variation, but the fact is that even if you cross into "deco" on one computer, you are
just barely into deco in it, and
almost in deco on the other- even if the times displayed seem quite different in minutes.
It's a problem that comes up because we treat "no-deco" limits as if they were a binary on-off switch that just got pulled. It's not a cliff- more like a very gradually increasing slope. In fact it's a very fuzzy area, particularly fuzzy on longer, shallow dives. It helps to remember that "all dives are deco dives". I'm betting that if you did dives with a shorter no-deco time- say a much deeper dive- you would see the variation between the computers shrink considerably. The slope of the line defining saturation would be much steeper, even though the two computers might not be the same, in terms of minutes of no-deco time they would seem much more alike. With shallow dives, even though the variation in times that the two computers gave you seems large in terms of minutes, in terms of tissue calculations it is probably very slight. So in real terms, the risk differential between them is probably negligible.
I'll just paraphrase my favorite quote about decompression algorithms, from Erik Baker, that "all decompression algorithms are attempts to draw a bright, clear line through a fuzzy gray area".
Ron
---------- Post added ----------
I have heard the same thing. It is quite debatable. ...<snip>
I suspect that there were problems getting his pure model to work on the computers, and there may have been compromises. That may be what "Suunto RGBM" means. The article cited above refers to the RGBM superiority especially with the use of helium, and the HelO2, which came out less than two years ago year, is the first Suunto computer that I know of that includes helium. That may be why Bruce put his name on the HelO2 and not the others. (Please note all the unfounded speculation in that paragraph.)
My only experience with the HelO2 was last year when a buddy played with it while we were diving. We planned our dives using another bubble model, VPM, with the profile planned on a computer but executed in the water without one. He tried to set all the variables on the HelO2 to make it come out as close as possible to the VPM model we were using, and he just wanted to see how it worked in the water. The first time we tried it, he became afraid of bending it (going into error mode for 48 hours), so when it varied from the plan we had made, he wanted to follow the computer instead. We ended up doing 20 minutes more deco than we had planned. That was also 20 minutes more than anyone else, and people were very curious as to what was keeping us down there that long. The next time he took it along was on a much deeper dive. He said we would follow our VPM plan no matter what. We followed our plan perfectly, and the HelO2 got bent (went into error mode) just as we were starting our 30 foot stop, with more than 20 minutes still to go on our ascent.
People told me later that he must not have set the variables correctly, but I sure got the sense that the new Suunto RGBM by Bruce Weinke will leave you in the water a very long time on a decompression dive.
multilevel, mixed gas, and saturation).
It's not really a debate, more of a marketing based misconception. There are two broad implementations of RGBM. "Folded" RGBM is what Suunto and every other RGBM computer on the market with the (partial) exception of the Atomic Cobalt use. This refers to a Haldanian calculation that is modified with various factors to emulate the results that the fully iterative RGBM algorithm would give. It's a way of getting about the same results without having to run the extremely computationally intensive "fully iterative" RGBM calculations, and it is more appropriate for recreational depths. It's not because there is or was any problem getting RGBM to work- except that the "full" version requires a lot of computational horsepower and some clever coding to be efficient enough to run on anything that is also diveable. The "Suunto RGBM" label is, just like "Mares RGBM", etc., only a marketing indication that the implementation is to some extent custom to that manufacturer.
As to the issue of being more conservative or liberal, RGBM contains factors that can be adjusted for overall conservatism. Suunto has tended to tweak those pretty far towards being more conservative. It's a company decision, not the algorithm itself being inherently more conservative- just as having "lockouts" for violations is a distinct design decision.
On the Atomic Cobalt (disclaimer: I'm one of the designers of this computer) we utilize a folded RGBM implementation for recreational depth dives, and shift to a fully iterative implementation for dives over 150'. Both are legitimately called RGBM.
Ron
