Buhlmann, Computers, and Helium

[And]

Hi

Disclaimer: Like you, I am continually learning about deco and technical diving. These articles are written primarily to ensure that I have understood what I have learnt and hopefully to encourage others to explore this fascinating subject. I do not guarantee this is correct and I do not recommend you do anything without knowing why you are doing it. I am not responsible if you read anything I write and change your plans and get bent. I am not responsible if your buddy or anyone you know gets bent, and I am not responsible should your dog be run over by a car or because you find a fingernail in your Big Mac. You make your own choices.
Just wanted to make that clear.

In previous threads we have discussed algorithms and such like. What I am hoping to explore today is the decompression model that most computers are based on and why we, as technical divers, are beginning to question these models.

Haldane

Mark Powells excellent article on Caissons leads me to the next step in deco theory, and introduce a clever man from Scotland called John Haldane. He was around at the time when the caisson workers were suffering decompression illness and he started experimenting on goats to find a solution.
Yes, the deco computer you wear on your wrist and the tables in your laptop have their roots in goat anatomy

In my article on air and decompression I talked about perfusion and diffusion. Read that again if these terms are not familiar to you. Haldanes research led him to the conclusion that gasses (in this case air) breathed by the caisson workers at pressure diffused into their tissues and when they ascended from the caisson, these gasses diffused out of the tissues and formed bubbles which caused DCI.

He knew that certain tissues worked differently from others and his model suggested the human body be viewed as a group of tissues, all arranged in parallel, ie they all diffused gasses at the same time. He also realised that the caisson workers were able to work at shallow depths of 10mtrs or so without getting bent, no matter how long they worked. We call this tolerable overpressure ie the tissues are able to withstand a certain amount of pressure without resulting in DCI symptoms. One point to note is that we are talking severe pain and disablity or type II DCS, not tiredness and headache, what we might call type I DCS.

Workman and Buhlmann

This idea of an unlimited level of tolerance was refined by Haldane in his model and was further developed by Robert Workman who created the ‘M Values’ the maximum tolerance of each tissue group, dependent on the depth and tissue ie he started splitting the body into more than one tissue group. Also at the same time as Workman was working on this so was Albert Buhlmann, a very clever fellow focussing on altitude diving, as he was based in Switzerland. Buhlmann published a book which gave instructions on how to calculate decompression and this is perhaps why his model is the most popular as it is readily available and is extremely flexible in that it can be calculated for repetitive diving, altitude diving, flying after diving etc etc.

Ongassing and Offgassing

We know that tissues with good perfusion absorb inert gas quicker but also diffuse quicker. When a tissue is exposed to a higher partial pressure of inert gas than is inside the tissue then it absorbs the inert gas until it reaches the same pressure. The same thing happens with offgassing, when the external partial pressure drops then the gas is diffused from the tissue. This is the main assumption in the Buhlmann model. The Tissues are said to ongas and off gas in halftimes, ie after a certain time, the pressure is halved and then after that same time it is halved again. This forms a decay curve common to many biologists and scientists. If we ascend to fast ie faster than the M value the gas cannot diffuse at its proper rate and forms bubbles in the tissue which cause DCI.

ZH-L16 Model, 'Bend and Mend'

Buhlmann created the ZH-L16 algorithm which split the body into 16 tissue compartments, each of them having a half time, from minutes to many hours. He used M values for each of the compartments to determine the theoretical maximum tolerance or tolerable overpressure as I mentioned earlier. One point to note is that the whole focus on this calculation is the partial pressure in the tissue compartment versus the partial pressue of the inert gas breathed, in this case nitrogen.

How Buhlmann works is that it calculates a maximum depth you can ascend to with each compartment based on the depth and time and also referring to the halftimes and the tolerable overpressure. This would mean that on shallower dives the faster tissues are the ‘controlling’ tissue and on longer deeper dives then the middle compartments would be the controllers. This is logical as the slower tissues will not load up on the shorter shallower dives. One main point about this is that we are using tolerable overpressure ie the amount of abuse a tissue can take before bubbles are formed. This is in essence why we sometimes refer to these profiles as ‘bend and mend’ simply because you bend those faster tissues at the start of the ascent and then fix them later during the shallow stops.

Testing and Versions

Buhlmann did loads of testing with his model, but only with nitrogen as the inert gas. He also padded the model to allow for things like rapid ascents, heavy workload, and poor physical conditioning and a whole bunch of other things. This modified algorithm is called ZH-L16B. He further padded the algorithm again for use with computers as their real time calculations removed some of the conservatism of the table model. This is called ZH-L16C. There was another model even more padded which is used in Aladdins computers, the ZH-L8 ADT model. Other computers will adapt the algorithm depending on how conservative they want to be, and can do this in a variety of methods, by planning dives deeper or longer, or by altering the compartment overpressure value (Gradient Factors is an example of this)

Helium, Giving the Wrong Impression?

And so, how does Buhlmann cope with helium? Well, his calculations are simply modified to include the different partial pressure of the helium as well as the nitrogen. He derived his half times for the compartments on the ones for nitrogen and largely guessed the M-Values for the tissue compartments with regard to helium. He died before he could test his theories properly.

The model we are left with seems to treat helium far too conservatively and gives the impression that decompressions from helium are longer than with nitrogen. Helium is said to be a ‘faster’ gas than nitrogen to the tune of around 2.65 times. This means in straightforward terms that the helium will diffuse quicker into the tissues than nitrogen but also diffuse out of them quicker which Buhlmanns model does not seem to accommodate and is perhaps were the logic of ‘getting off the helium as early as possible’ came about.

Knowing what we now know about both helium and how bubbles are formed during deco we are now seeing a growth in the popularity of deep stops on trimix dives as well as air dives. Some computers have incorporated them into the algotithm, sometimes called microbubble stops, such as in the Aladin computers and the VR3, although the VR3 is reported to have a more effective strategy than the Aladin. This simply slows the ascent to allow the faster tissues to offgas without punishing them and also for the faster helium to offgas properly, although it should be pointed out that these models are still based on the algorithm which we now believe treats helium too conservatively or incorrectly, depending on your point of view. For those of us without computers the deep stops benefit us in that we can reduce the shallow stops and create a more efficient deco curve IMHO.

One other point is that Doppler studies have shown that bubbles form after most dives. Buhlmanss model assumes all gas is eliminated in the liquid phase (diffused in tissues) rather than the gas phase (bubble form). These phases are quite different in how the diffusion process works. Newer deco models are attempting to model this gas phase as well as the liquid phase such as VPM and RGBM. Currently these are models which seem to more realistically model the decompression process.

Feel free to add to, correct or otherwise comment.

Andy

[Mark]

Nice post Andy!

Originally Posted by And

For those of us without computers the deep stops benefit us in that we can reduce the shallow stops and create a more efficient deco curve IMHO.

It is also worth mentioning that by doing the deep & intermediate phases of the deco correctly you have a much better chance of getting out of trouble if things go wrong shallow.

Obviously an understanding of the oxygen window and deco gas selection are critical in acheiving this 'more efficient deco curve'.

I'm sure everyone is looking forward to Andy's next article where maybe he will cover both topics


Regards,

Mark.

[Mark Chase]

Nice article Andy. This is the focus of my investigations at the moment due to higher levels of He used in my deeper dives massively increasing my deco. Opting for an END of >25m rather than <30m costs me dearly on long deep dives. There is also a good argument for He in the deco gas without additional deco penalty, which is being pushed by DIR and Mark Elliot. The He in the deco mix is kinder to the lungs and can arguably enhance off gassing of the nitrogen but it is still early days for this theory.

It should be noted that the VR3 would also reduce the shallow stops proportional to the extension of mid and deep stops. The computer will also give 5min stops on the deep range stops (below 30m) on longer / deeper dives so it is not a good idea to compare it with the 1min bubble stops on the Aladdin. You should have a look at Proplanner to get an idea of the VR3 profiles it might surprise you.

ATB

Mark Chase

[Andy Phillips]

Originally Posted by And

Hi

ZH-L16 Model, 'Bend and Mend'

Buhlmann created the ZH-L16 algorithm which split the body into 16 tissue compartments, each of them having a half time, from minutes to many hours. He used M values for each of the compartments to determine the theoretical maximum tolerance or tolerable overpressure as I mentioned earlier. One point to note is that the whole focus on this calculation is the partial pressure in the tissue compartment versus the partial pressue of the inert gas breathed, in this case nitrogen.

How Buhlmann works is that it calculates a maximum depth you can ascend to with each compartment based on the depth and time and also referring to the halftimes and the tolerable overpressure. This would mean that on shallower dives the faster tissues are the ‘controlling’ tissue and on longer deeper dives then the middle compartments would be the controllers. This is logical as the slower tissues will not load up on the shorter shallower dives. One main point about this is that we are using tolerable overpressure ie the amount of abuse a tissue can take before bubbles are formed. This is in essence why we sometimes refer to these profiles as ‘bend and mend’ simply because you bend those faster tissues at the start of the ascent and then fix them later during the shallow stops.

Testing and Versions

Buhlmann did loads of testing with his model, but only with nitrogen as the inert gas. He also padded the model to allow for things like rapid ascents, heavy workload, and poor physical conditioning and a whole bunch of other things. This modified algorithm is called ZH-L16B. He further padded the algorithm again for use with computers as their real time calculations removed some of the conservatism of the table model. This is called ZH-L16C. There was another model even more padded which is used in Aladdins computers, the ZH-L8 ADT model. Other computers will adapt the algorithm depending on how conservative they want to be, and can do this in a variety of methods, by planning dives deeper or longer, or by altering the compartment overpressure value (Gradient Factors is an example of this)

Helium, Giving the Wrong Impression?

And so, how does Buhlmann cope with helium? Well, his calculations are simply modified to include the different partial pressure of the helium as well as the nitrogen. He derived his half times for the compartments on the ones for nitrogen and largely guessed the M-Values for the tissue compartments with regard to helium. He died before he could test his theories properly.

The model we are left with seems to treat helium far too conservatively and gives the impression that decompressions from helium are longer than with nitrogen. Helium is said to be a ‘faster’ gas than nitrogen to the tune of around 2.65 times. This means in straightforward terms that the helium will diffuse quicker into the tissues than nitrogen but also diffuse out of them quicker which Buhlmanns model does not seem to accommodate and is perhaps were the logic of ‘getting off the helium as early as possible’ came about.

Knowing what we now know about helium and deco we are now seeing the popularity of deep stops, sometimes called microbubble stops, such as in the Aladin computers and the VR3. This simply slows the ascent to allow the faster to offgas without punishing them, although it should be pointed outh that these models are still based on the algorithm which we now beleive treats helium too conservatively. For those of us without computers the deep stops benefit us in that we can reduce the shallow stops and create a more efficient deco curve IMHO.

One other point is that Doppler studies have shown that bubbles form after most dives. Buhlmanss model assumes all gas is eliminated in the liquid phase (diffused in tissues) rather than the gas phase (bubble form). These phases are quite different in how the diffusion process works. Newer deco models are attempting to model this gas phase as well as the liquid phase such as VPM and RGBM. Currently these are models which seem to more realistically model the decompression process.

Feel free to add to, correct or otherwise comment.

Andy

Andy,

Sorry mate, but,

Buhlmann created the 8 tissue compartment (ZH8) which worked fine to depths up to 40M, this is the bases for the PADI tables (and others), he also developed the 12 tissue compartment, but his work on this was incomplete, the work was taken up by others in his group, he had little involvement in the actual development. The ZH16 is a variation on the 12 with further comparments added to the mid range. The B varient was developed for tables as it gives distinct steps (stops) but doesnt like reverse profiles during a dive, i.e. it gives huge penalties for not doing the deepest part first and then riseing, hence the C varient which pads certain areas of the algorythm, but also allows for continuous variation during the dive, hence the reason why dive computers use it. There is no allowance for deep stops, it simply adds the time to the total deco. Also, mend and bend is slightly inaccurate, as the buhlmann solution considers a compartment to have become saturated and does not add to it, this is part of the reason for added "slower" compartments for the ZH16 solution. As for the Helium issue, it does not allow for the fact that HE is a faster gas, rather adds the time to the deco as if it were N2.

The deep stops have nothing to do with HE, rather came about from the research done by Pyle, which is why they are sometimes refered to as "Pyle stops"

Sorry to go on about this, but the work Buhlman, his group and others have done around the "tissue compartments" forms the bases for the vast majority of the decompression we do today, including your "deco on the fly". It is only in recent times we have begun to move away from this and move into the RGBM and VPM theories.

Take care,

Andrew

[And]

Hi

Mark, I have edited the post to distinguish the VR3. If you want to write a paragraph about it I will be happy to insert it as I have not had too much luck in finding out how it handles deep stops other than the profiles you have posted in the past.

Andy, no need to apologise, you are not saying anything I disagree with. The symantics of who developed what over the 30 year period Buhlmann worked on his theories I fully accept your accuracy and I apologise if I have stated Buhlmann rather than Buhlmann and his group. I based my post on the algorithms presented in the 1995 edition of the Buhlmann book so I kind of ignored the initial work to keep the post interesting (for me anyway ).

All deco is based on the decay curve, it has to, the shallow stops are the longest, the deeper stops are the shortest. I haven't said otherwise. VPM and the others simply look at gas in the gas phase as well as the dissolved phase and treat the ascent accordingly. My post is hopefully an introduction to these concepts rather than a criticism of Buhlmanns work. One cannot just jump into the new stuff without understanding the foundations, in fact the B in VPM-B stands for Buhlmann, as you know.

What I have said is that Buhlmann used the overpressure tolerances in his equation which are related to the speed of the initial ascent and form the basis of my bend and mend statement. They are the a and b in the equation Mark posted on his deco for beginners 2 thread.

I have said that the tissues reach equilibrium with the external partial pressures rather than stating saturation points. I did not discuss 'saturation' per se as it is different for each tissue compartment and largely unsuitable in this context. As per the decay curve a tissue can never be saturated due to the half time concept although after 6 half times we say it is to all extensive purposes, saturated. I have edited the bit about deep stops. It didn't read how I wanted it too. Thanks for pointing that out.

I don't think we are disagreeing here, maybe I've just not made my intentions clear enough. This is hopefully an article which will lead people to VPM, RGBM, Pyle etc. and understand a little more where the computer gets its numbers from. It is definitely not a we are right you are wrong post.

Sorry if that didn't come across better.

Andy

[Andy Phillips]

And,

nope, I think we are on the same vein, I just wanted to add a little bit more about Buhlmann, he is after all the man who really brought decompression to the masses (so to speak). If you read all of his book you are a better man than I. It is a tad "dry" to read, I have kinda picked at it over the last few years. I did manage to get copies of Pyle's papers which are a very interesting read.

Andrew

[chrisch]

Guys

I have been reading this stuff with interest. I am curious as to where the Helium M values came into being. If, as I read it above, AAB "guessed" at them why are they in use? Or was there perhaps an input from the military e.g. USN.

Most of us use VPM simply because V Planner is easy to get and not expensive. Baker's model & theories may or may not be any better than others.

To write this software (or Proplanner or whatever it is that others may use) the author must have input the M values and therefore there must (or maybe I am assuming too much) be some "accepted" M values somewhere.

Or have I misunderstood how V Planner et al work? I assumed they start with some form of Buhlmann algorithm as the base from which to calculate, with some modification of either M values or saturation or off-gassing times depending on how the author views either deep stops or bubble mechanics (Weinke) or both?

Chris

[And]

Hi

You have misunderstood how Vplanner works. Thats a different subject altogether and one I alluded to in this piece. One indication is when you choose advanced settings and look at the sheer number of variables you can set. Buhlmanns model is a lot simpler in that it is only concerned with the difference in ambient pressure of the tissue compared with the inspired partial pressure of the gasses.

VPlanner is based on the work of Yount and Hoffman who were based in Hawaii working with the divers there. He used the behaviour of bubbles to arrive at his formula, along with empirical evidence of the Hawaian divers. It is quite different to Buhlmann. VPM-B came about simply because although the stops are deeper in VPM the shallow stops were deemed to short 'for the masses' and some compartment pressures were added to extend the shallower stops.

Andy

[Mark Chase]

Andy M8 when you attempt to compact several book loads of information into a one page article your bound to have holes that will be deemed important overlooked points by some but the effort that goes into getting this threads of the ground is appreciated.

Picking up on a point made by Andyp:

Helium off gassing is a grey area but I thought it was generally agreed that deep stops / very slow ascent from depth was critical for preventing helium bends. One thing with Nitrogen bends is that it is very unusual to feel them in the water most of the time they feel it as they surface or on the boat. Helium is different and two of my friends have had Helium hits on the 6m stop.
All agreed that heavy workload on the bottom and ascending to fast to their first and second stop depths were the likely cause.

Just experience here no technical data to back it up but it was the obvious similarity in the two cases. Both involved paralysis one in the face and one in the arm and were treated with 02 on the boat and no recompression was sought. This demonstrated that the bends were He related and not Nitrogen as the latter would have necessitated a visit to the pot.


ATB

Mark Chase

[chrisch]

Ah! that explains some of the posts I've seen on Scubaboard about V Planner being bend-o-matic.

I also have downloaded a program "XS for windows" which uses a Norway Navy algorithm and gives remarkably short stop times (it has a ZHL option if you prefer it).

Clearly there are many ways to come at the problem and not just the Buhlmann route.

For the sort of diving I do I am happy to use a computer, but the odd deeper dive (35-45m) using mix means cheating. I am unhappy to cheat and realise that the nitrox based calculations don't stop you deep enough for helium. So you sort of make the worst of both and do a EAN profile with a mix but use precut tables for the stops. This tracks the O2 for other shallow dives to follow.

V Planner will always get you out much faster than the computer (Suunto). so you are highly unlikely to have problems as far as I can see.

I am looking at the VR3 but sort of thought it was Buhlmann too and would therefore follow much the same profile as the Suunto but a bit faster as it doesn't have the "fudged" RGBM ammendments.

I'll do some more reading....

Chris