Understanding Dive Computers

Dive computers are important and useful underwater instruments across sport, commercial, military, scientific, exploration, and technical diving sectors. We discuss dive computers, 2 basic underlying models, uses and misuses, data collection and correlations, and some observed features of modern diving with computers. Selected


Introduction
Modern digital dive computers [1][2][3][4][5] date to the early 80s, though analog devices simulating tissue gas uptake and elimination through porous membranes date back to the 70s. Analog devices were limited to nonstop diving and had a short shelf life. Digital dive computers proved highly successful and very useful right from the start, progressing from just table emulators to full up algorithmic staging devices across mixed gas, open circuit (OC), rebreather (RB), nonstop, decompression, deep, and shallow diving. Dive computers are moderately expensive items these days, and high end units range beyond $1500. Basically, a decompression computer is a microprocessor consisting of a power source, pressure transducer, analog to digital signal converter, internal clock, chip with RAM (random access memory) and ROM (read only memory), and pixel display screen [5]. Pressure readings from the transducer are converted to digital format by the converter, and sent to memory with the elapsed clock time for model calculations, somewhere in 3-10 second intervals. Results are displayed on the screen, including diver time remaining, time at a stop, tissue gas and bubble buildup, time to fly, oxygen toxicity levels (CNS and pulmonary), and other warnings (model violations). Some 3-9 volts is sufficient power to drive the computer for a couple of years, assuming about 100 dives per year. The ROM contains the model program (time step application of model equations), all constants, and queries the transducer and clock. The RAM maintains storage registers for all dive calculations ultimately sent to the display screen. Dive computers can be worn on the wrist, incorporated into consoles, or even integrated into heads up displays in masks.
Depending on model implementations and ad hoc practices, dive computers can signal divers with audible and displayed warnings for violations. Underwater, modern dive computers can accommodate manual and programmed breathing gas switches in their computational synthesis as OC tank and RB set point changes are made. Some units are equipped with tank-to-computer wireless connections to read tank pressures. USB computer-to-computer connections permit downloading of dive profiles for later analysis and data storage. Dive computer updates from manufacturers are also easily accommodated in the same computer-to-computer mode. With software supplied by the manufacturer, dive planning is seamlessly executed using the same algorithm hard wired into the dive computer. Conservancy levels are user knobs based on age, gender, water temperature, workload, diving experience, and related factors. Dive computers today are sophisticated devices supplying a wealth of information and controls for safe diving. 2. OC: underwater breathing system using mixed gases from a tank that are exhausted after exhalation.
3. RB: underwater breathing system using mixed gases from a tank that are recirculated after carbon dioxide is scrubbed from the exhalant and oxygen from another tank is injected into the breathing loop.
4. RAM: data storage array in modern computers.

5.
ROM: computational array in modern computers that processes information, does calculations, and sends output to registers and displays.
6. Diving algorithm: combination of a gas transport and/or bubble formation model and coupled diver ascent strategy.

7.
Decompression stop: necessary pause in a diver ascent strategy to eliminate dissolved gas and/or bubbles safely and is model based. Stops are usually made in 10 feet increments.
8. Deep stop: decompression stop made in the deep zone to control bubble growth.
9. Shallow stop: decompression stop made in the shallow zone to eliminate dissolved gas.
10. OT: pulmonary and/or central nervous system oxygen toxicity resulting from overexposure to oxygen at depth or high pressure.
11. DCS: crippling malady resulting from bubble formation and tissue damage in divers breathing compressed gases at depth and ascending too rapidly. 13. BPM: bubble phase model dividing the body into tissue compartments with hypothetical half times that are coupled to inert gas diffusion across bubble film surfaces. An exponential size distribution of bubble seeds is usually assumed. Throughout the dive, the cumulative volume of growing bubble seeds is constrained by a single limit point called the critical volume, or phase volume, in all tissue compartments.

Literature Review
There are literally 100s of articles in medical, mathematical, physics, chemistry, and computing science peer re-viewed journals on dive computers. Additional and very useful information about specific operations of any particular dive computer can be found in user manuals, which are extensive, complete, and lengthy.

Key Concepts
Dive computers are useful tools across recreational and technical diving. Able to process depth-time readings in fractions of a second, modern dive computers routinely estimate hypothetical dissolved gas loadings, bubble buildup, ascent and descent rates, diver ceilings, time remaining, decompression profiles, oxygen toxicity, and many related variables. Estimates of these parameters made on the fly rely on two basic approaches [3], namely, the classical dissolved gas model (DGM) and the modern bubble phase model (BPM). Both have seen meaningful correlations with real diving data over limited ranges but differ in staging regimens. Dissolved gas models focus on controlling and eliminating hypothetical dissolved gas by bringing the diver as close to the surface as possible. Bubble models focus on controlling hypothetical bubble growth and coupled dissolved gas by staging the diver deeper before surfacing. The former gives rise to shallow decompression stops while the latter requires deep decompression stops, in the popular lingo these days. As models go, both are fairly primitive, only addressing the coarser dynamics of dissolved gas buildup and bubble growth in tissues and blood. Obviously, their use and implementation is limited, but purposeful when correlated with available data. To coin a phrase from a community at large, all models are wrong, but some are use ful. As research plods forward, computer manufacturers are both quick and flexible in responding to change and update, adding to computer viability as a diving tool. It's reasonable to expect usage in diving to grow with commensurate sophistication.
Presently, some 15 -25 companies manufacture dive computers employing both the DGM and BPM in another 200-250 models by last count. Recreational dive computers mainly rely on the DGM while technical dive computers use the BPM. In the limit of nominal exposures and short time (nonstop diving), the DGM and BPM converge in diver staging. Dive planning and decompression software is also readily available from some 15 -25 vendors.

Important Model Estimators
Instantaneous estimates of parameters needed to stage divers by dive computers rely an mathematical relationships coupled to pressure sensors and clocks in the unit. Important ones follow [3]:

Dissolved Gas Model (DGM)
Diver staging in the classical Haldane approach limits inert gas tissue tensions, p, across all tissue compartments, λ, with halftimes, τ, by a limit point, called the critical tension, M, according to, If M is exceeded at any point on ascent, a decompression stop is required. Helium tissue halftimes are 1/3 nitrogen tissue halftimes. Algorithm is used in recreational and technical diving across OC and RB systems. Algorithm typically brings diver into the shallow zone for decompression (shallow stops). Ascent rates are nominally a slow 30 f sw/min.

Bubble Phase Model (BPM)
Modern bubble phase models (BPM) couple tissue tensions to bubbles directly by assuming an exponential dis-tribution, n, of bubble seeds in radii, r, excited into growth by changing ambient, P, and dissolved gas, p, total pressure, ( ) n N exp r β = − for N and β constants obtained and/or fitted to laboratory or diving data. To date, distributions of bubble seeds have not been measured in vivo. Using the same set of tissue halftimes and inert gas tension equations above in the DGM, diver staging in the BPM requires the cumulative bubble volume excited into growth by compressiondecompression, φ, to remain below a critical value, Φ, throughout all points of the dive and in all tissue compartments, With D tissue diffusivity, S tissue solubility, and γ bubble surface tension. In applications, the critical phase volume, Φ, is taken near 600 microns 3 and surface tension, γ, is taken around 20 dyne/cm. Diffusivity times solubility, DS, is also fitted to diving data. If Φ is exceeded at any point on ascent, a decompression stop is necessary. Algorithm is used across recreational and technical diving on both OC and RB systems. Staging starts in the deep zone and continues into the shallow zone (deep stops). Ascent rates are also 30 f sw/min.

Oxygen Toxicity (OT)
Both pulmonary and CNS toxicity are tracked by dive computers in a relatively simple way. Pulmonary toxicity is tracked with a dose-time estimator, Γ, written, 5. deep switches to nitrogen based breathing mixtures are avoided by technical divers, with a better strategy of increasing oxygen fraction with commensurate decrease in helium fraction in the breathing mixture; 6. RB usage is increasing across the full spectrum of diving; 7. wrist dive computers possess chip speeds that allow full resolution and implementation of the most complex diving algorithms; the computer industry, at large, is becoming increasingly interested in marketing new dive computers.

Validation
To validate computer models [2], data is necessary. In the past, data consisted mostly of scattered open ocean and dry chamber tests of specific dive profiles. In such instances, the surface of correlating model and diving data was only scratched. Today, profile collection across diving sectors is proceeding more rapidly. Notable are the efforts [2,4] of Divers Alert Network (DAN) and Los Alamos National Laboratory (LANL). DAN USA is collecting profiles in an effort called Project Dive Exploration (PDE) here and DAN Europe has a parallel effort called Dive Safe Laboratory (DSL). The focus has been recreational dive profiles for air and nitrox. The LANL Data Bank collects profiles from technical diving operations on mixed gases for deep and decompression diving on OC and RB systems. Some interesting features of the data have emerged: 1. profile collection of diver outcomes is an ongoing effort at DAN USA, DAN Europe, and LANL and has aided in model tuning using rigorous statistical techniques; 2. there are no reported spikes in DCS/OT rates for recreational and technical divers using dive computers; 3. statistics gathered at DAN and LANL suggest that DCS/OT rates are low across recreational and technical diving, but that technical diving is some 10-20 times riskier than recreational diving; 4. data from meter manufacturers and training agencies, reported as anecdotal at recent Workshops, suggests the DCS/OT incidence rate is on the order of 200/3,000,000 dives (underlying incidence) for computer users; 5. the underlying incidences in the DAN and LANL profile data are on the order of 60/190,000 and 23/2,900 respectively.
Profile collection efforts such as these enormously benefit divers and diving science. Without downloadable profile data from dive computers, meaningful algorithm and protocol analysis is very difficult. Profile data banks are important resources for all kinds of diving.
Powerful and useful as dive computers may be, there are some downsides with their usage [1,5].
1. pushing the computer beyond its model limits and correlation envelope; 2. not reading the operating manual;

ignoring warnings;
4. violating ascent rates; 5. diving with a computer that is not properly initialized; 6. ignoring ceilings; 7. using one computer for two divers; 8. improperly entering gas mixtures and pp O set points; 9. twiddling overly liberal correction factors; 10. violating depth restrictions; 11. not performing predive planning; 12. turning the unit off when in ERROR mode (less possible these days).
There are others imbedded, but the above gives the flavor. As with all task loading, risks decline with increasing diver proficiency and computer saavy.