Sponsored by:
Submerge Scooters 
click here...

Video from January 2005 Rodney's latest "home built" back mounted, 
bellows, KISS valve CCR. Footage in Jackson Blue Spring, scootering "Middle Grounds", 
back out to the mainline then up to the "Banana Room" and beyond to about 
4,500 feet/1.3Km in, total distance scootered 10,000 to 12,000 feet, 
3 to 3.5Km with UV-18 NiMH scooter. Click here....

Rebreather Articles

Rebreather training standards

Carbon Dioxide and the Rebreather Diver- a Survival Guide

Unique Failure Modes of Constant Mass Flow Rebreathers

Unique Failure Modes of Fully Closed Circuit Mixed Gas Rebreathers

Full Face Masks and rebreathers... is it for You?

Deep Decompression Stops and rebreathers

Andrea Doria, RMS Republic, and cave diving dive report

The Advantages of Helium

Partial Pressure of Oxygen selection in Closed Circuit Rebreathers

Passive Addition Rebreathers

Halcyon Operating Principle

Mandatory Reading: Oxygen Sensors

More on sensors

Should you sell your open circuit equipment?

Return to Homepage










AARG rebreather standards: background definitions

Copyright February 1998.

revised 3rd February 1999,

& April 2000.

FO2 Fraction of oxygen in a gas mixture.

Partial Pressure
The pressure exerted by any constituent gas in a mixture; the most significant measurement with respect to the physiochemical and physiologic behavior of a gas. As described by Dalton's law, the total pressure of a gas mixture is the arithmetic sum of all the individual partial pressures of the constituent gases.

PO2 Partial pressure of oxygen.

C02 carbon dioxide

V02 volume of oxygen metabolised by the body in 1 minute

loop The loop describes the counterlungs, breathing hoses, C02 absorption canister/system and mouthpiece with upsteam and downstream check valves. Basically all gas filled parts of a rebreather that contains gas rebreathed by the diver.

flushthru or flushthrough technique of venting all the gas in the loop and the replacement of this
gas from a high pressure gas cylinder. A common procedure used especially but not exclusively by semi-closed circuit divers.

setpoint the chosen operating PO2 of fully closed circuit rebreathers

hypoxia physiological symptoms precipitated by insufficient oxygen

hyperoxia physiological symptoms of excess oxygen, generally central nervous symptom involvement possibly including grande mal type convulsions, and unconsciousness

hypercapnia physiological symptoms of increased carbon dioxide tension in lung gas and blood usually encountered in rebreathers due to scrubber failure/exhaustion and the resultant elevated CO2 in the breathing loop

scrubber the carbon dioxide absorbent canister, a common component in all self contained rebreathers.

mixed gas in relation to rebreathers any gas mix containing >0 and <100% oxygen, the remainder of the gas mix made up of any single or combination of inert gas. Note air is classified as mixed gas as the FO2 in the loop rarely equals FO2 in diluent supply gas.

For the purposes of this standard, rebreathers will be distinguished as one of 2 types by the following classifications:

Semi closed circuit

Semi closed is defined as a mixed gas rebreather where the metabolised oxygen is replaced via a source of mixed gas. Designed to operate below the safe depth of pure oxygen rebreathers. Gas is pre-mixed in high pressure cylinders with oxygen levels to suit the depth of the dive.

Fully closed circuit

Fully closed is defined as a rebreather in which the metabolised oxygen is replaced from a source of pure oxygen. Designed for depths exceeding the safe depth of pure oxygen diving, a source of diluent gas is used to keep the PO2 within life support range at any depth. The monitoring/control of oxygen in self contained units invariably requires some sorm of electronics, which usually includes 3 galvanic oxygen fuels cells used as oxygen sensors.

Fully closed rigs have relatively stable PO2 in the loop, i.e. a variable F02 which decreases in direct and inverse proportion to increased absolute pressure. This is in contrast to semi-closed units in which the PO2 increases in direct proportion to depth.


Silent submersion distinguishes between 2 ratings:

o Semi closed circuit (unit specific)
o Fully closed circuit (unit specific)

These ratings are unit specific as rebreathers unlike OC equipment may have widely varying operating procedures.



Common core theory and practice requirements, semi closed and fully closed rating.

Theory, common core:

the diver must demonstrate a sound understanding of the:

o effects of oxygen, both hypoxia and hyperoxia, including PO2 life support ranges.
o rate of oxygen metabolism in divers
o calculation of the duration of CO2 absorbent
o variables that effect canister duration
o disadvantages of using nitrox supply gas in rebreathers, especially at depths exceeding 20 meters, including ventilatory requirements of the diver, inert gas narcosis, decompression and thermal considerations and canister efficiency.
o physiological effects of elevated CO2 and the ramifications for the rebreather diver.
o physiological effects of hypoxia and hyperoxia including warning signs
o planning for adequate bailout to the surface, including all required decompression stops
o reasons why an OC second stage is required as the 1st option for safety in the event of a loop flood or scrubber failure.
o inherent difficulties of starting the CO2 absorbing reaction in-water and it's implications for divers who have a second rebreather as their only bailout.



Practice, common core

The diver must demonstrate competency in:

o pre-dive preparation of all equipment.
o pre-dive checks including functioning of check valves, loop integrity via both positive and negative pressure checks, and scrubber functioning via pre-diving the loop for a minimum of 5 minutes.
o unit specific pre-dive checks
o in water bubble and leak checks.
o providing adequate buoyancy in the event of a loop flood. This is simulated by the addition of weight equal to the buoyancy of the loop volume of the rebreather being used, and where appropriate additional weight to simulate suit compression due to depth.
o a blue water ascent maintaining a safe ascent rate, and adequate buoyancy control. This is more difficult on rebreathers as there may be 3 pieces of equipment that need buoyancy control: the BC, drysuit, and the loop.
o the ability to plan for the bailout requirements of their dive partners, including any required decompression stops.
o choice of bailout gas's to provide a safe ascent to the surface with a PO2 of no less than 0.16
o re-filling the scrubber to minimise chance of channeling
o accurate record keeping of dive-time on each canister fill.
o cleaning of the unit between dives

Semi-closed circuit specific requirements:

Theory, semi closed:

the diver must demonstrate a sound understanding of the:

o calculation of theoretical FO2 in the loop with different flow rates and supply gases.
o importance of accurate pre-dive flow rate testing for constant mass flow semi-closed units
o duration extension of CO2 absorption advantage in semi-closed units.
o effects/risk of flow rate reductions caused by foreign particles in the constant mass flow orifice.
o calculation of equivalent depths, for decompression calculations

Practice, semi closed circuit rating

o For constant mass flow units, the diver must demonstrate competency in pre-dive flow rate testing. This includes introduction of back pressure equal to planned diving depth.
o For other types of semi-closed units, such as passive addition systems, the diver must demonstrate competency in the unit specific pre dive checks supply gas flow rate.
o the diver must demonstrate a flushthru before any ascent
o in the case of constant mass flow units with fixed intermediate pressure 1st stage regulator, a secondary cylinder with compensated intermediate pressure 1st stage, and a OC second stage regulator is required for bailout

Fully closed circuit specific rating requirements:


Theory, CC

the diver must demonstrate a sound understanding of the:

o calculation of adequate diluent volume for the depth of the dive
o parameters for choosing a setpoint PO2
o ways to avoid PO2 spikes on descents, part of choice of correct diluent.

The diver must demonstrate an advanced familiarity with:

o partial pressures of gases
o mixed gas decompression theory
o the ramifications of using helium diluent
o the specific unit being dived
o the common causes of death and inherent weaknesses of the make of unit they are diving, and ways to avoid them.
· Oxygen sensor function and the vulnerability of the diver to hypoxia due to condensate on the sensors, especially in automatic oxygen add rebreathers


Practice, CC

o the diver must demonstrate calibration of the oxygen sensors. The sensors must be calibrated to 1.0, with pure oxygen. If setpoints greater than 1 are used, the sensors must be calibrated at that setpoint using a pressure pot, or verified using pure oxygen at the relevant depth before descent. for example if the setpoint is 1.3, verify sensor readings with pure oxygen at 3msw.
o the diver must show competency in the manufacturers pre-dive checklist.
o demonstrate an adequate loop flushthru with both diluent and oxygen, as they may differ in procedure due to placement of addition point in the loop.
o the diver must demonstrate the ability to control the unit in semi-closed mode, both with the aid of the oxygen sensors, and simulated without sensors using pre-dive manual addition semi-closed calculations.

Return to Homepage

Return to index


Carbon Dioxide and the Rebreather Diver: A SURVIVAL GUIDE

Rodney Nairne (c)copyright 2000

Hypercapnia is the term applied to the effects of increased tension of carbon dioxide in the blood. Open circuit divers are at little risk of this unless deep diving on air or by deliberate hyperventilation to conserve gas supplies.

By comparison rebreather divers are more commonly effected due to re-breathing exhaled CO2.

Fact: If you dive a regularly dive a rebreather you will experience hypercapnia at some point. As we can't eliminate this problem the goal for
rebreather divers should be effective management by taking both preventative action and corrective action.

Some ways to help minimize the risk of hypercapnia:

The rebreather should have fresh absorbent. If you intend to use one absorbent fill for several dives, it is imperative you keep accurate records of dive time on each canister fill. Some divers even log the amount of oxygen metabolized to get a closer estimation of the amount of CO2 which has been scrubbed, but for most purposes dive time is sufficient.
Always be conservative when deciding if a change of absorbent is necessary: the stuff is relatively cheap.

Check the scrubber is packed immediately before the dive: Some settling of the absorbant material is to be expected in transit to the dive site. Many canisters have a way of compensating for this reduction in the volume of the absortbant, however a check before the dive is always wise.

Keep water out of your scrubber: if your absorbent becomes wet it just won't work as well. If you do wet your scrubber and it seems OK it's illusory as any increased output of C02 will just not be absorbed. In other words it will work until you really need it. The cause is water soaking the porous granules where much of the CO2 is absorbed, and nothing short of an oven will dry it out again.

Other things to watch: ensure you have the gas flow in the correct direction, some units are less tolerant of this mistake than others. The mouthpiece check valves are also a common failure point if they become "hung up" or just old. A good pre dive check is to block one side of the mouthpiece hose off and check the valve can hold a little pressure.

This method is preferred to just listening for the "slap" of the valve as many valves will still slap when hung up.

When it happens: what you can do.

Switch to open circuit is sound advice as a first step, just to get a clear head and evaluate the situation whilst breathing uncontaminated gas. Carbon dioxide can hinder cognitive processes. After this you may decide to go back on the loop if the cause of the problem was just poor breathing patterns, heavy exertion, or if necessity demands it even a failing scrubber.

In many cases, switching to open circuit does little to alleviate the symptoms of severe hypercapnia in the first 10-20 minutes. Therefore it is imperitive if diving in overhead environments to carry at least 30 minutes of open circuit gas, even if you have a redundant rebreather.

Always restrain and minimize any exertion if you get in this situation, and STOP and catch your breath BEFORE it gets out of hand. There is a lag between exertion and the output of CO2 so don't get overconfident that you have sorted a situation; in reality it's going to catch up to you in a big way real fast.

A good technique if your scrubber is not performing is switching to manual semiclosed mode, exhaling each 4th or 5th breath will allow you to continue even if the scrubber is almost completely inoperative. Clearly this should not be a common practice however it may get you out of a serious situation unscathed.

Lastly just be plain professional and maintain a high standard of equipment, training and awareness.

Carbon Dioxide Levels

 Carbon dioxide level


 Description of symptoms


 US Navy cutoff limit for closed circuit system CO2 absorbers


 Most divers sense stimulated breathing rate and depth


 Breathing stimulus does not increase much above this level


 Resting patient is conscious, but unable to use self rescue due to depression of the central nervous system.


 Becomes anesthetic

Return to Homepage

Return to index

Unique failure modes of semiclosed circuit constant mass flow (CMF) rebreathers.



(note... examples of CMF rebreathers are the Drager FGG, FGT, Atlantis and Dolphin and Drager Ray.

CMF semi closed circuit rebreathers are simple and reliable devices, and may be second only to oxygen rebreathers in terms of common usage. The whole system is dependent on the correct and accurate functioning of the CMF orifice, which meters the flow of the mixed gas into the breathing loop.

As the whole purpose of a rebreather is greater gas efficiency, the CMF is usually set with a small margin of safety, most commonly to give a fraction of oxygen in the loop equal to 19-21% (to avoid hypoxia on the surface) assuming a maximum oxygen uptake (VO2) of 3 liters per minute. The F02 in the loop can be estimated by the formula:
F02= ((flow rate *F02 in supply gas) - Max Vo2)/(flow rate -V02)
For example a nitrox mix with 60% oxygen assuming a flow rate of 6 liters per minute and a diver with a V02 of 3 liters per minute will give us a F02 of 20%.

The upper VO2 limit represents a reasonable upper limit of oxygen metabolism that the average diver will not be able to exceed for prolonged periods. Actual VO2 uptakes have been recorded twice this level however, which would result in hypoxia if repeated for a reasonable period underwater. Instantaneous hypoxia does not occur due to the loop acting as a oxygen reservoir and the increased pressures whilst diving giving an increased PO2 for a low fraction of oxygen.

Whilst this formula can be manipulated to calculate the flow rate required for any gas mix, in reality the flow rate will be limited to one of 3 as determined by the CMF orifices available in the rebreather. Each of the 3 orifice sizes flows a different mass per unit time of gas, and is designed to be used with a supply gas with a specific fraction of oxygen and a specific inert gas diluent, in all but rare cases this is nitrogen. The most common mixes are nitrox 60, 40 and 32 for diving to 16, 30 and 40 msw respectively. This depth limit is calculated by the maximum safe PO2 of 1.6ata of oxygen. (note: military divers may use a max safe depth calculated on a P02 of 2.0 ata, or 23, 40 and 52msw respectively. )

The CMF orifice will operate according to the following formula, provided the supply pressure is 2 times the ambient pressure, and the supply gas is composed of only nitrogen and oxygen.

flow rate in cubic feet per minute = 11(supply pressure)(diameter of orifice in inches)squared.

e.g. for a supply pressure of 450psi and a required flow fate of 6 liters per minute as in the example above, the approximate orifice size is =0.0065 inches or 0.185mm.
This formula leads us to a failure mode unique to CMF rebreathers: particle impingement (PI). PI of an orifice as small as 0.188mm is possible with a wide variety of substances commonly found in diving equipment: corrosion particles can either block the orifice directly of indirectly, as can sand particles, and even (and most commonly) salt water that can bypass the
usual block for particles (brass sintered filters) in soluble form to form sodium chloride crystals either between the filter and orifice, in the orifice or after the orifice upon evaporation of the solvent.

To demonstrate the effect of particle impingement, assuming a moderate blockage of 10% of the orifice diameter and using the above formula we have a reduction of flow from 6 to 4.85 liters per minute. Substituting this into our original formula gives a negative value for F02, this means the gas mix will be insufficient to sustain life at a V02 of 3 liters per minute.

Unfortunately should blockage by salt crystals result in a fatality, it would be difficult to establish this as a cause. Firstly, if salt crystals were responsible, they would almost certainly redissolve in the time intervening the accident and the recovery of unit as it would have flooded. Secondly, if salt crystals are found, it is most likely they formed once the unit was recovered and sufficient time had elapsed for the water solution to have evaporated leaving the crystal solute.

The insidious nature of such a flow reduction, is that it is not reliably detectable by the diver once in the water (there is no reliable warning). Unfortunately pre dive checks will not prevent either subsequent formation of crystals or other particle impingement's which may occur during the dive. This is in stark contrast to open circuit equipment where a failure of the life support apparatus is instantly apparent by inability to take a breath. The rebreather diver may be able to continue to breath normally, unaware of any problem, until oxygen reserves in the loop are depleted and hypoxia occurs. This outcome is more likely in the deeper range of diving depths due to the oxygen fraction making up less of the loop volume.

Particle impingement is only one example of hypoxic outcomes possible in semiclosed circuit rebreathers. Should the supply cylinder run out un-noticed during a dive (an unfortunately common event in scuba diving) again the diver may not be aware of the flow stopping, as the only indication would be a cessation of a small amount of bubbling from the unit which reduces with depth. This could easily go un-noticed, especially if the diver is pre occupied with another task, is wearing a hood in cold water, or diving with comparatively noisy open circuit divers. Again the comparison with open circuit: if the cylinder runs out the diver can not breath so is instantly alerted.

Pre-dive flow rate checks may be compromised if a supply mix with a fraction of oxygen less than that required for the orifice size is used for
a dive. The flow rate check in this case will pass the rebreather for use whereas in fact the rebreather will give a mix containing insufficient oxygen. A possible mistake might be if the 32 mix is connected to the 60 orifice, of if the nitrox mix is substantially oxygen deficient, for example the 60 mix contains only 50%. (note should this occur, and hypoxia result it may be difficult to test for the nitrox mix in the tank after it has been totally discharged).

The responsible and properly trained semiclosed rebreather diver will be aware of these risks, an be vigilant in pre dive checks. The diver will also be aware, (despite pre dive checks) of the risk of reduced flow during the dive, and always flush the loop with fresh gas before ascents, where hypoxia is most likely due to the reduction of pressure and the resultant drop in PO2 to hypoxic levels should a flow impingement have occurred.

Return to Homepage

Return to index

Unique failure modes of electronically controlled fully closed circuit rebreathers
By Jason Rogers (c) 1998

In the last issue, Rodney Nairne dealt with some failure modes of constant mass flow semiclosed circuit rebreathers. Now I'd like to look at similar situations, in electronically controlled fully closed circuit rebreathers.

First (as always seems to be the case) some background information.
Electronically controlled 'breathers tend toward similar lines, and while there are of course many model specific differences, they usually have these major components:

A system of mechanical devices such as;
An ambient pressure loop
A system for providing medium pressure diluent and oxygen
A diluent add valve
A over pressure dump valve

As well as the electronic components that we will be dealing with here;

O2 sensors
These produce a tiny voltage in proportion to the partial pressure of oxygen present on the actual sensor.

An O2 add valve
This controls the flow of pure oxygen into the loop, and is usually driven electrically

A primary electronics package
This is the "smarts" of the rebreather, and it decides when to open the O2 add valve based on information it receives from the O2 sensors

A primary display
A simple display, intended to stay in the diver's field of vision at all times, it is normally limited to displaying that the O2 level is in one of three states, "low" "correct" or "high". Some displays are set up to show other things, but this comes as a trade off with readability.

A secondary display
Often erroneously called "backup", this more complex display, it usually there to provide a numeric display of the sensor outputs, and is useful during calibration of the setpoint, and so the diver can check the outputs from the multiple sensors for agreement.

A system of cabling
This connects the elements of the electronic systems together, and allows them to send information from one element to another, usually in the form of an analog variation in voltages. Most cables need to carry more than one type of information, and so they are multiple wire cables, with several information carrying lines, and one common ground line.

Now to examine the possible failure modes, we can play the "what if" game with each element that makes up the electronic system. The main failures for systems like this will be, flooding or breakage, in other words, short circuit and open circuit. The effects of these failures unique to each element, and I'll discuss them in turn.

O2 sensors

These are the beasties that evaluate the O2 levels in the loop. Known for being unreliable they're almost always used in sets of three, with the primary electronics programmed to make some type of evaluation of the information that each sensor is giving, and work out which one to believe
The most common type used has a semipermiable membrane covering an electric cell. This type of cell uses oxygen as part of it's workings. When there is no O2, the cell stops working and the output voltage falls to zero. As the partial pressure of O2 in the gas around the cell increases, more O2 diffuses into the cell, and the output voltage rises.

Because it is consumed by the reaction that generates the current, it can go 'flat' like any other battery. When it is going flat, it will become non-linear, in other words, the amount of voltage produced will no longer be in direct proportion to the O2 level of the gas in which it is immersed. What happens is that the cell is no longer able to create high currents, and high O2 levels will result in less voltage, or conversely, higher O2 levels will be required to create the same output. This is of particular significance to users who check the calibration of the set at an O2 level less than the setpoint. For instance, there is no guarantee that sensors that worked at say a PPO2 of 0.95 will remain linear all the way up to 1.4 Since it is impossible to calibrate O2 sensors at 1.4 on the surface, there is no way to tell what actual O2 level is going to be needed to make the sensors output a voltage corresponding to 1.4

The sensors are also subject to variations due to the collection of water droplets on the surface of the sensor. If the membrane becomes covered in water droplets (whether due to condensation or water ingress) the loop O2 level will be uncoupled from the sensor voltages. If the coverage happens when the loop O2 level is higher than the setpoint, the rest of the electronics will assume that all is well, and the O2 add valve will not be opened. Both the primary and secondary displays will show that the O2 level is within range, and the diver will deplete the loop O2 and pass out when it falls below life support range. If the O2 level is below the setpoint when the sensors are covered, the O2 add valve will remain open. Both the primary a secondary displays will show that the O2 level is within range, but the O2 level will increase unchecked, rising above the life support range.

Water ingress has obvious causes, and obvious solutions, but condensate on the sensors can be quite subtle. Mature designs will have had it determined that the sensors remain warmer than the dewpoint of the gas passing over them under all conditions. Experimental designs may not have determined this factor. Additionally, such things as reversing the direction of the loop may mean that there are local variations in the temperature and dewpoint within the loop. Where the sensors may have been protected from condensate, they may now be at a point where large amounts of water will form. This can arise from something as simple as installing the mouthpiece check valves incorrectly, or fitting the hoses back to front.

Unexpected water formation may also lead to a partial short circuit of the sensor leads, resulting in a suppression of the signal from the sensors. This will mean that an increase in loop PPO2 will be required to maintain the setpoint voltages.

O2 Add valve

Some type of electromechanical valve. As with all valves it can stick open, closed or partially open. Perhaps the least dangerous of all failures in an electronic rebreather, as any one of these three failures should be readily apparent to the user through the displays. In the situation of a stuck open valve, the user should have 1-2 minutes in which to spot the problem and react if they are using a sane setpoint, somewhere close to the middle of the life support range. This is because a sensible design will only flow a couple of litres per minute more than the greatest metabolic requirements, and so it will take some time to push the O2 level to the point where it will cause acute oxygen toxicity. Of course when using a setpoint up around 1.4 or 1.6, near the top of the life support range, the delay before reaching lethal levels is shorter, and due to the diver's exposure to hyperoxic mixtures, the delay before the onset of gross symptoms will be reduced.
Leakage of water into the electrical parts of the valve may overload the primary electronics, disturbing it's other outputs.

Primary electronics package

Normally mounted in some type of pressure proof housing, this is where the decisions are made in the rebreather. The housing will normally also contain the main battery, that is used to drive the primary electronics, the primary display and the O2 add valve.
Since it is inside a strong housing, it is normally immune to breakage, however since it needs to be regularly dismantled for battery changes and calibration, it is quite prone to flooding. Major leakages are not normally a problem, as the entire electronics package will fail suddenly and totally, making the diver instantly aware that something bad has happened and that they need to abort to the preplanned bailout. Of course a sensible diver will realize that the O2 sensors can no longer be trusted and that a manual bailout on the secondary display readings is foolhardy. The sensors can only produce tiny voltages, and after a primary electronics flood out, they'll be connected to a lot of wires, big batteries and salt water. Not the ideal state for a sensor that can only output milliwatts.
Minor floods are more serious, as they are difficult to identify, and their effects are difficult to predict. Small amounts of water somewhere in an analog system are just plain bad. They may effect the readings, but not the results, they may effect the results, but not the readings or they may effect both. It can damage the operation of the software in the primary, leading it to alter the way it handles (and displays) the results of disagreement amongst the sensor voltages. It may suppress the sensor voltages themselves, causing the set to increase the O2 levels or increase the sensor voltages, causing the set to allow the O2 level to drop.

Essentially, if the set begins to do anything weird, such as adding a lot of O2, fluctuations in either display, sudden variation in primary battery voltage or in fact anything unusual, then you should suspect primary electronics flood.

Primary display

Usually a very simple device using flashing lights, or shrinking/growing bar graphs. Beware that if it floods the short circuit in the display can effect the operation of the primary electronics. Analog circuits are naturally sensitive to variations in voltages, and the unexpected load of a flooded display will probably cause unexpected voltages to appear in the wrong places. If this effects the sensor voltages, then the secondary display may start giving you incorrect information.

Secondary display

Usually with it's own internal power source, in most designs it displays the sensor voltages (converted to a PPO2 reading) and some indication of the state of the primary battery. A floodout in this display can connect the sensors to the primary battery voltage, or to the secondary battery voltages, or both. It can also suppress the sensor voltages by providing a short circuit. Of course if this happens, then the primary electronics can no longer control, or display the O2 level correctly. The O2 level may go up a little, or a lot, or down a little or a lot. No way to tell.

Cable system

All the information that is carried within the rebreather's electronics goes via cable. If any of the cables leak, then the information is reduced to garbage, and the components that it's connected to (at both ends of the cable) will not behave correctly. For example: a cable carrying only O2 sensor voltages. If it leaks, then it may tie all the voltages together. All three sensors will read the same value, even if one or two of them fail. Clearly the O2 level would have to be much higher to maintain the same average voltage if one of the sensors dies. While this is happening, the apparent O2 level will remain constant, the electronics will perform as normal, and both displays will read normal levels, with agreement between the sensors.

As I think you will see, the electronics of a rebreather depend on the correct function of every part. No element can be considered to be working independently of the others. Often you will hear statements such as; "If the primary electronics fail, you can fly the unit manually just off the independent backup display, with some practice it's easy!". Now you will have the tools to judge for yourself the value of that person's knowledge of the subject.

Return to Homepage

Return to index

Some comments on failure modes in electronic CCRs By Walter Stark

Hypothetical failures in hypothetical designs are somewhat hard to talk about. The possibilities are limitless. I will instead suggest some ways to avoid or minimize them as much as possible.

Electronics and sensors situated together and using short fixed Teflon insulated cables present minimal opportunity for cable failure. Electronics situated atop a back pack and operating at ambient pressure are least likely to leak and if so most likely to let gas out rather than water in. If transparent plastic is used for the absorbent canister, electronics housing and breathing bag simple visual inspection for absorbent condition, condensation and water leakage is possible. Cylindrical canisters and electronics housings can be strong, rugged, easy to make in a lathe and lend themselves to highly reliable radial o-ring seals.

The greatest danger of water entry comes via the mouthpiece. A breathing circuit whereby exhalation goes direct to the counterlung and inhalation is drawn in through the absorbent canister past the electronics provides double water/condensation removal before the sensors. Outlets from both counterlung and absorbent canister should be free standing and away from the bottom to reduce the likelihood of picking up any water which may have gotten in. Counterlungs should be provided with drains which can be used underwater.

Readouts should operate at high impedance so that any failure of them or their cables have minimal effect on the control circuit. Redundant sensor circuits should incorporate signal limiting or clipping to prevent failure in one from dragging the others past safe limits. Switchable redundant batteries should be used. Only batteries constructed with hard wired connection between cells should be used. An extra ppO2 readout which is totally independent of the control circuit is worth considering.
With a well designed unit by far the greatest source of failure is operator error and failure to pay attention.
Some common errors to watch out for:
1. Trying to squeeze more time out of absorbent, batteries and sensors.
2. Careless or rushed setup. Lack of thoroughness.
3. Black box syndrome. The calibration is way off so just crank up the trim pot until it reads what you want rather than finding out why it's so far out.
4. Not thinking about what you are doing. Leads to mistakes like valving in pure O2 at depth instead of inert gas.
5. The She'll be Right assumption. The alarm is sounding and the readouts are high but if I take a few more breaths it will probably drop back.
6. Leaky mouth washout. This is all too common. Experienced OC divers are pretty casual about letting some water in around the mouthpiece. Even when it's gurgling away with each breath in a CCR they may ignore it until the system is well on the way to flooded. When a mouthful of pH 13 seawater wakes them up they blame the unit, saying, "It flooded!".
7. Drunken diving. Related to 4. above. If you have N2 on the inert side whether you know it or not you are impaired at depth and more prone to not pay attention or to make simple mistakes. There is a lot of individual variation in this. Using a mix which keeps the ppN2 below 3 ata is a good idea.
Other than water leakage by far the most likely
equipment failure is manifested as solenoid not operating or stuck open and readouts too high, too low or one is drifting away from the others. For solenoid problems over ride with manual control. You should be able to hear if a solenoid is working or not. If all readouts are high cut off O2 and flush with inert. If this bring them back use manual control and abort dive. If it doesn't you have major electronic problems. Go to manual SCR mode and abort dive.
If all readouts are low use manual O2 control. If the readouts respond properly continue with manual control to abort dive. If they do not. Flush with inert premix and abort in manual SCR mode. If one sensor is drifting but the others are responding normally simply abort dive permitting the remaining two to effect control. If the errant one is erratic or extreme go to manual SCR mode.

This all presumes the inert gas supply is premixed with a breathable % O2. Manual SCR simply means valving in the inert premix, taking two to four breaths (depending on depth and %O2), exhaling overboard and refilling from the inert. If a decompression obligation is involved you can switch to O2 at 40 fsw. If your first stop is deeper than that then you should have a standby breathing system available at the first stop.

Return to Homepage

Return to index

Full Face Masks

Jason Rogers (c) 1998

Given the inherent risk of unconsciousness that is a part of rebreather diving, many users are tempted to use full face masks in order to provide airway protection. Adapting masks intended for open circuit diving to fully closed does however present some difficulties. Most OC masks are designed either for surface respiratory protection (firefighting) applications and adapted for underwater use, or intended for use with large amounts of surface supplied gas. Common examples of these are the Interspiro AGA, and the Kirby Morgan Bandmask.

In the case of the Interspiro type of mask, the water seal comes from a tight fit, combined with a positive internal pressure, maintained through the use of spring loaded exhaust valve, and spring loaded diaphragm. Very little water enters the mask but any which does is expelled through the exhaust valve, located at the bottom of the mask. (In the case of the AGA, the exhaust valve is combined with the demand valve, but in many other surface use masks, the exhaust valve is a separate unit) Cool dry air entering the mask is directed across the faceplate, and then into the oralnasal mask built into the mask. Use of flapper valves separates the path of the inhale and exhale gases, reducing dead space, and keeping the faceplate from fogging up.

The problem with this is that the rebreather only provides for ambient pressure within the mask, and so leaks may leak in, not out of the mask. Leakage into the mask will have to follow the gas path back into the loop. In fully closed sets, or semiclosed sets that have high mounted dump valves (e.g. the Atlantis/Dolphin) this water will accumulate within the loop. Additionally, the gas coming out of the loop and being directed across the faceplate is warm and moist. When it contacts the cool faceplate, the gas will dump moisture, and the faceplate will tend to fog up. All this makes this type of mask unsuitable for use with a rebreather.

The Bandmask style of FFM is even worse. Intended for surface supply diving, little or no attempt is made to provide a gastight face seal. Rather it is intended to leak like a sieve, and the water that comes into the mask is expelled via the exhaust valve located in the bottom of the demand valve housing. Additionally a separate hand operated valve can be opened which provides a constant flow of gas into the mask, to defog the faceplate and blow water out of the mask. Gas usage can be prodigious, as can be water ingress. It's more than enough to make this type of mask useless for rebreathers.

There is however another type of mask. Here in Australia, they go by the generic term of "CressiSub", however they are produced not only by CressiSub, but also by TechniSub and ScubaPro. They are similar to a normal scuba half mask but with an extension. The lower skirt has 'grown' to cover the lower half of the face, and the skirt goes across the forehead, down the cheeks and under the chin, sealing the entire face off from the water. There is a simple oval hole, corresponding to the user's mouth, and the mouthbit of the rebreather is simply jammed through the hole. This provides the user with a normal mouthbit to grasp with the teeth and a dry space covering the rest of the face. There is normally no flow of gas within the mask, much like the situation in a normal scuba half mask, and the defogging techniques used with a half mask work as well with this type of FFM. A 'spider' or head harness retains the mask, and so most of the loads from the hoses that would normally be taken by the user's mouth, are transmitted to the whole head, greatly increasing comfort, and allowing the use of larger bore or more rugged breathing hoses. Any leakage will accumulate in the lower part of the mask, and by exhaling via the nose the user can clear the mask, just as with a half mask. In the event of loss of consciousness, the bit will usually be retained in the mouth, but in any case, the nose and mouth both remain dry, thereby preventing drowning. Users switching from a mouthbit to a CressiSub will find that the integrity of the rebreather/user interface is markedly improved and the quantity of water accumulating in the loop is quite minimal. Users of sets that have stiff loop over pressure valves (or no valves at all as is the case with some homebuilts) will find that they can dump gas from the loop nasally, just as they can with a half mask. Also it is possible to remove the mouthbit, while the mask is in place, allowing the diver to speak. This is useful during the prebreathe phase of the dive, where the diver can exchange information with others, without wasting oxygen by breathing atmospheric air and so requiring a second O2 flush.

The disadvantage of this type of FFM, inherent in all FFMs, is that in the event of a loop failure, the entire mask must be removed in order to change to bailout supplies. Apparently there is a new FFM on the market which provides some form of docking port, allowing you to swap supplies without disturbing the whole mask, but I've not had the opportunity to dive this type of mask and see if it will work in a CC application.

As with all FFMs, underwater vomiting will present major problems. I recently scrubbed a dive due to fear of seasickness causing me to vomit into my FFM. I don't have a shut off type mouthpiece, and my best guess for how to handle this would be to try to remove the mouthbit, and vomit into the mask, then replace the mouth bit, flood the mask with water, and then clear it. I have never tried this, and if anyone has met this situation with success perhaps you'd like to contribute to the next newsletter!

If using the spit 'n wash defogging technique, the user will also find that rinsing a mask that is attached to the set quite awkward. A small squeeze bottle of water will greatly facilitate this process!
Another slightly unexpected disadvantage I found when I made the switch to FFM on my rebreather was that I no longer liked to dive alone! The advantage of having a protected airway is that you won't drown, but this is of little use if you simply sink to the bottom to die. I now much prefer to dive in the company of others, who can swim over and push a few buttons on my set in the event of my passing out.

Return to Homepage

Return to index

Deep decompression stops... and rebreathers


The aim of the procedure is to keep gas in solution right up to the last step. When gas is kept in solution it diffuses more rapidly from tissues. If all bubbling is kept to a minimum until the last step, the bubble size will be least affected by pressure changes, compared to bubbling occurring during the traditional large pressure change upon ascent to the 1st, and deepest stop level.

Lastly, the bubbling we assume occurs with the short untested decompression times is kept to a minimum by using heliox for bottom mix and decompression*, the low solubility of helium reducing the volume of gas liberated by the final step to the surface.
*only feasible in open water with constant PP02 rebreathers.

The basics of the procedure is this:

*All deep dives are conducted with heliox rebreathers running PP02's of 0.7 to 1.0.
*The deep Pyle stops are modified to stops starting 2 atmospheres off the last level (e.g. if wreck is 75msw but last 5 minutes spent at 57msw, 1st stop would be at 40msw) provided the 1st table stop is not violated.
*All deco is on heliox , using the rebreathers to give us oxygen enriched heliox mixes at each stop.
*A partial flushthrough of oxygen is conducted at 10msw.
*On all dives requiring significant deco, nitrox 50 is carried for emergency decompression should the rebreather fail. (also doubles as drysuit inflation gas)
*For safety no open circuit second stage regulators are provided on oxygen cylinders, however oxygen should be staged at 6msw for emergency use.
*Heliox, nitrox 50 and oxygen tables are used as the basis for all decompression profiles.
These tables are then heavily modified. Notice how the 6msw stop is halved and this time is spent between 12msw and 30msw.
*For high workload dives, the time at 6msw is increased to approximately the standard length Buhlmann stop. The PP02 is kept at 1.4 max so hyperoxia is extremely unlikely.
*Should DCI occur post dive rebreathers and open circuit oxygen are available for in water treatment.

Standard table for heliox bottom mix and open circuit. Nitrox 50 and oxygen deco

75m for 20:00 (21) on Heliox
33m for 1:00 (23) on Heliox
30m for 2:00 (25) on Heliox
27m for 2:00 (27) on Heliox
24m for 3:00 (29) on Heliox
21m for 1:00 (30) on Nitrox 50
21m for 2:00 (31) on Nitrox 50
18m for 3:00 (34) on Nitrox 50
15m for 4:00 (38) on Nitrox 50
12m for 5:00 (43) on Nitrox 50
9m for 8:00 (51) on Nitrox 50
6m for 32:00 (83) on 100% Oxygen

Runtime 83 minutes

AARG Modified table, heliox is used for bottom mix and deco with a setpoint PP02 of 0.7 to1.0, a change to 1.3 to 1.4 at 10msw and up.
75m for 20 min
55m for 1 min
50m for 2 min
45m for 2 min
40m for 2 min
35m for 3 min
30m for 4 min
25m for 5 min
20m for 9 min
15m for 10 min
10m for 10 min
6m for 15min

Runtime approx. 85min

At this stage we have only used these procedures for approximately 40 exposures, however these early dives have had very encouraging outcomes.

Return to Homepage

Return to index

Dive Report
Rod and Suzies USA trip July 1998.
by Rod Nairne & Suzie Dudas

Our 1st dive was a checkout dive at Dutch Springs, a flooded quarry. Just in case we needed to acclimatize to cold water diving, and check out our trim for out upcoming dives off Montauk on the RMS Republic. Nothing special to note, the water temp was 10 C and we were in the water for 100 minutes.

Having use of the Dudas Diving Duds store made our gas mixing for the Republic trip easy; with the aid of the haskel gas booster we were able to get full pressure heliox 18 fills in our steel 72 cylinders, as well as boosting the pony bottles to 2500psi of oxygen. For the 2 of us then I mixed 2 72's of heliox 18, 2 72's of nitrox 50, and 2 pony bottles of pure oxygen. We also had a 3000psi 80cf tank of oxygen for top ups on the pony's, as well as for emergency in water deco on open circuit.

We had a 6 hour drive to the dock where we met up with the charter vessel Seeker. The Seeker has a long history in NE wreck diving and is run by it's second owners Dan Crowell and Jennifer Samulski. The second captain aboard for the trip was John Chatterton, crew was Pete Wohlleben, and Greg Mossfeldt, divers were John Yurga, Bonnie Merkel, Richie Kohler, William Cleary, Pat Rooney, Peter Hess, Joe Ferrali, Mike Trapani, and Jeff Schwartz.

We arrived at the Republic site later than expected due to strong winds the night before, so only 1 dive was made the 1st day. The Republic is a 570 foot long steam ship, a White Star Line Passenger vessel the victim of a collision in 1908. Of special note was than the Republic was the 1st shipwreck where radio was used to successfully summon assistance. The Smithsonian Institute allegedly offered prior salvors $250 000 for the telegraph key if recovered. Also, the ship is rumored to have been carrying $30 million dollars worth of gold eagles, 1908 value. For this reason the Republic site has been commercially salvaged several times, using saturation diving techniques. No gold and only a small fraction of the ships fittings/cargo has been recovered.
To our knowledge we were only the second or third group of recreational divers to visit the site; being only 7 miles from the popular wreck of the Andrea Doria, and a deeper dive, most divers are drawn to the larger Italian vessel where recovery of china a common place event.

We warmed up our scrubbers for about 10 minutes before the dive, and started out with a PO2 of around 0.4 to avoid any oxygen spikes on the descent to 250 feet with heliox 18. (Normally we run a diluent with 10% oxygen, and start the descent with close to 100% oxygen in the loop). Arriving on the bottom after we tied off our reel and started swimming along the wreckage. Unfortunately we ended up picked into the debris field off to the side of the wreck, a fact only evident when after reeling out about 150 feet of line, the huge ribs of the hull became visible.

Due to either a strong sense of survival or cowardice we decided not to swim around the wreck without the pace hindering reel. According to the saturation divers, the ship lies very flat with only 18 inches between decks, and massive hull plates obscure any hope of navigating along the sides of the vessel. The only artifacts we saw were a few portholes (we managed to recover 1 of these, as did 1 other diver). Suzie also spotted a few nice sized lobsters, one a possible 15 pounder.

That night we had to moor on the Andrea Doria wreck site, and the consensus of the divers was to stay in place the next day and dive the Doria. Although I had a strong urge to go back to the Republic, Suzies mother I knew would be pleased with us diving the Doria. As it turned out, we had a nice tourist dive on this wreck, we swam down from the side of the ship to the bottom in search of the bridge area, the vis at the 255 foot bottom being only about 20 feet. 20 feet is not much on a 700foot wreck! With no luck locating the bridge or the compass stand from which Suzies Dad recovered the compass in 1967, we ascended slowly to the promenade deck, and swam aft to approximately half way along the wreck, on the way passing Gimbals hole. We were lucky today with a good 50 feet of vis at the 180-200 foot level.

Many of the veteran Doria divers returned from 'Gimbals hole' with bags full of China, from a new hole located originally by Gary Gentile this season. Apparently Gary remained tight lipped about where he was recovering the 1st class coffee cups but John Yurga located the same spot, a crack in the central stairwell at 205 feet.

As it turned out, Suzie was the 1st female to dive the Doria on a rebreather, a nice one-two for mother and daughter… Evie Bartram Dudas was the 1st female to dive the Doria with her husband, John Dudas.

For the tech-heads:

We used 72's 1 heliox 18 for diluent for the whole dive (with the breather we do "oxygen enriched heliox" deco's, no gas switches). The purpose of the 18% oxygen is so we can emergency deco out on open circuit bottom mix to 70 feet, where we would switch to our other 72 filled with nitrox 50, in the event of a loop flood. The nitrox also is used for suit and BC inflation. We run a constant partial pressure of oxygen of 1 ata until 30 feet were we bump it up to about 1.3 to 1.4 ata. As it turned out, Suzie had to switch to open circuit nitrox 50 at 30 feet after her loop flooded. The cause of the flood was a pinched 0-ring on the canister.

Return to Homepage

Return to index

(c)Rod Nairne 1998, Silent Submersion 2000

An increasing number of experienced RB divers are discovering the advantages of using heliox diluent, including using the gas for decompression. At Silent Submersion we even use heliox for shallow 50 to 80 foot dives.

Here are a *few* points for helium usage, hopefully to stir up some discussion, and hopefully stop the practice of all the new RB divers which is to dive nitrox down to 50msw because they "don't have the ticket". SCARY!


*is easier to breathe at any depth
*is less soluble than nitrogen
*produces no narcosis
*is cheap to use in a breather and in fact breathers in some ways were developed for it's use
*results in no diver fatigue after a deep dive, in fact in our experience produces less fatigue at 250 than nitrox at 100! So at 100 it's got to be incredible.
*can be used for decompression with a fully closed unit
*Saturation divers use heliox deco right to the surface, they are the real pro's
*takes less heat energy out of the scrubber so theoretically extends scrubber life in cold water
*Requires less dive planning for a deep dives than trimix with several gas switches on ascent


*is a difficult gas remove from the tissues
*is dense, so is hard to breathe even at 100 feet (by comparison)
*produces longer deco for long dives (which is what breathers can be used for, right!)
*produces narcosis, something you don't want while operating a device nicknamed "death machine"
*Far more nitrogen can be absorbed in your body than helium due to high solubility
*May result in higher incidence of neurological DCI due to high soluability in fatty tissue such as spinal cord
*Due to density theoretically takes more heat energy out of scrubber, could result in premature scrubber failure in cold water
*is really only any good in Guinness


Dave Crockford from the UK has been diving his inspiration with heliox to 93m and reported that HPNS or Helium willies were not a problem for him. One of the team did suffer slightly with compression arthalgia but it soon passed off once stable at 93 metres. (He is in his 60's) It manifested as shakes, evident in the hands.

Return to Homepage

Return to index

Copyright Rod Nairne. 1998, Silent Submersion 2000.

The recent advent of closed circuit rebreathers for recreational diving has been accompanied by a change in the P02 setpoints commonly used in these devices. Historically setpoints from 0.5 to 0.7 were used however now it is rare to find divers running a setpoint below 1.2

It may be instructive to investigate some possible reasons why a 30 year tradition in military and commercial/scientific diving has been broken in such a short period. Have we learned something new? Has human physiology changed? Has there been a change in the operating reliability of the oxygen control systems of current models?

These and other questions will have to remain a mystery, I guess it all adds to the "mystique" surrounding closed circuit rebreathers, a mystique only increased by fatalities and near misses.

I have decided to put together some points to help in the choice of the best setpoint for your application. Some of these points will be controversial to many, especially recreational/technical divers, however I believe most of it is accurate and to the best of my knowledge reflects the cumulative and hard won experience and operating procedures of the professional diving community. I make no excuses for being biased, I run and recommend a setpoint between 0.7 and 1.0.

Note: Common practice is to always flush-through with pure oxygen at 6msw for decompression, to give a po2 between 1.3 and 1.6 depending upon efficiency of the flush.


Low partial pressures. (0.5 to 1.0)

*May lead to early hypoxia in the event of 02 add failure in closed position
*May cause problems on rapid ascents from depth due to falling PPO2
*According to theory will give longer decompressions (I say theory because I believe more trust is placed in the Buhlmann system that it deserves, based upon the number of DCI incidents I have observed)

High partial pressures. (above 1.0)

*In the event of solenoid failure in the open position gives less time to detect and shut off compared to a low setpoint, before dangerous PP02's are reached.
*Causes real operational problems on descent due to oxygen spiking,
*All the procedures used to reduce 02 spiking on descent, rate on the scale from inconvenient to dangerous, such as shutting electronics down, 02 valves off etc.
*depletes the body's natural defence (see Bill Mee's article this issue) to high PPO2's so that if a oxygen spike occurs late in the dive the chance of a convulsion is much higher (a fact borne out by combat swimmer experience)
*loads the body's oxygen clock during the high work load portion of the dive giving less of a safety factor for shallow water pure oxygen decompression.
*requires special procedures for sensor calibration/verification due to the setpoint being above 1 ata.
*leaves little margin for sensor inaccuracies on the high end of the scale, i.e. a 20-30% low reading will result in PPO2's exceeding 1.6
*there is limited data on the safety of high PP02's in closed circuit rebreathers, (what there is all bad) so qualifies as experimental diving, thereby increasing the risk where risk is already a significant operational factor.
*almost all reduced inert gas loading is lost at depth, the difference between a setpoint of 1.0 and 1.4 at 75msw is only 4.7% inert gas.
*high PP02's trade off the risk of hyperoxia against a perceived reduced risk of decompression illness, where hyperoxia will more likely lead to mortality, this is indeed mystical.

Additional Comments from Walter Starck:

Under normal circumstances some degree of spiking upon descent is unavoidable unless you choose to have 0.0 % O2 in your dilutent supply. Descent also often entails some vigorous swimming if a current is present. With a high setpoint you can easily enter the zone of possible oxygen toxicity. Falling PPO2 during rapid ascent is not normally a problem as it is automatically detected and O2 added by the control system.

Increase in PPO2 can take place quite rapidly. Decrease is limited by the rate of metabolic use and takes place slowly providing time to detect and take appropriate action. With low PPO2 remedial action only involves manually introducing more gas. High PPO2 requires shutting off the O2 supply, flushing the loop and introducing new gas.

The effect of frequent and prolonged exposure to PPO2 > 1.0 ATA is unknown but both theory and clinical evidence suggest the possibility of cumulative long term tissue damage.

Though running a high setpoint does result in less decompression as required by tables any decrease in risk of DCS is problematic. In other words the increase in inert gas absorption accompanying a lower setpoint is compensated for by longer decompression. The risk isn't increased so long as the required decompression is adhered to.

Modes of sensor malfunction tend to be in the direction of less sensitivity to O2 rather than more. This leads to a greater likelihood of PPO2 being higher than reported by the sensors rather than being lower.

The exposure/response pattern for oxygen convulsions is chaotic. The pattern for hypoxia is much more regular and consistent.

Everything considered, the upper limit of PPO2 is more dangerous than the lower. A setpoint of around 0.5 to 0.7 provides a comfortable buffer against hypoxia and stays well away from the multiple dangers toward the upper end. The only real advantage of a high setpoint is less decompression. The convenience is outweighed by the risks. I suspect the only real reason why higher set points have been chosen recently is that for the kind of dive profiles most users are doing they can save significantly on decompression and the divers involved are free to make their own choice. In the military and commercial worlds diving medics have a lot to say about such things and the divers themselves are required to follow the rules.

Return to Homepage

Return to index

Passive addition rebreathers

(c)Jason Rogers

We've all looked at one time or another at the equations that govern the movement of gases through active addition SCRs. They should now be well understood. There seems however, to be much confusion with respect to passive addition SCRs.

There are two main types of passive addition SCRs around, and they all share some common aspects.

Manual units
Automatic units

Manual units are those in which the diver dumps a volume of gas out of the loop and replaces it. This volume may be a full lung volume, or a part of the loop volume, or some type of flush through. A description of a dive using a unit like this can be found on the AARG website, the dive report of the DeeWhy and Meggol.

Automatic units dump some part of the loop volume with each breath. Generally some specific fraction of the diver's tidal volume. Automatic units can be further divided into three general categories.

Fixed ratio. These dump a set percentage of the divers tidal volume overboard (usually from 10-25%) with every breath. Oxygen fraction varies with depth, becoming closer to the drive gas as the diver descends and the effective dump becomes greater.

Depth compensating. These dump a percentage of the diver's tidal volume, which varies in inverse proportion to the absolute pressure. On the surface these units generally dump between 20 and 33% of each breath, but that decreases with depth, to keep the Oxygen fraction constant and reduce gas wastage. The fraction dumped falls so that for instance a unit that dumps 33% of each breath at the surface, dumps 3.3% of each breath at 90m (10 Bar absolute). These units will provide a fixed FO2 regardless of depth, when used with the same drive gas because the effective dump remains constant.

Partially depth compensating. These units are half way between the fixed ratio units, and the depth compensating units. They provide a high dump ratio near the surface, again 20-33% but the dump ratio is not fixed either with relation to a surface equivalent breath, or an actual breath at depth. They usually dump between 75% and 200% of a surface equivalent breath at 10 bar absolute, but of course this varies with design, and sometimes varies from individual unit to unit within one design.

Exactly how this dumpage is achieved varies from unit to unit, but some underlying principals can be discerned.

The factors which determine the composition of the inspired gas are, the Oxygen fraction of the drive gas, the oxygen consumption of the diver and the quantity of gas that is dumped from the loop.

Generally it is taken that the composition of the inspired gas is equal to the composition of the dumped gas. This is not always so, as some designs inject the makeup gas (make up gas is the drive gas, injected into the loop to "make up" for the gas that has been dumped or metabolised) just upstream from the diver. Hence the dumped gas, is equal not to the diver's inspired gas, but rather is equal to the gas that the diver exhales. However, it is bad practice to choose one's gas limits based on the minimum level of oxygen found in the exhaled gas, as there is no guarantee that fresh gas will be injected on every breathing cycle.

The dumped gas's composition is critical when it reaches the point where the O2 level is at it's lowest. This point can be determined by calculation.

The dump gas can be divided into two major components, the inert gas that is "left over" from the metabolism of oxygen by the diver, and an amount of drive gas that has passed through the system, unconsumed. This second component can be thought of as an unavoidable wastage. It's the part of the dump gas that contains the oxygen that keeps the loop above the minimum O2 level.

At this point, the only difference between the two types of rebreathers emerges. A manual system is calculated based on a time interval between dumps. This is based on the volume of Oxygen metabolised per unit time (VO2), or in other words, a rate of Oxygen metabolism. An automatic system is calculated with respect to individual breathing cycles, and depends on the ratio of tidal volume to the oxygen metabolised. Hence the convenient units are Litres and Tidal Volume Fraction, respectively.

So a couple of examples,

First for a manual system,

The drive gas is 50% O2, and the minimum dump gas O2 percentage is 20%. This allows the diver to surface at any time, and still have a breathable mixture. The O2 consumption is considered to be 3 litres per minute, which is normally a maximal VO2.

Now, the dump gas can be divided into two parts, for ease of analysis. One is the wasted gas. That's the O2 and it's associated inert gas that must never be removed from the loop gas. 20% of the dump gas must be Oxygen, and given that the drive gas is 50/50 mix, the inert gas associated with that will be the same volume, at 20%. Hence, adding those together, the wasted gas must represent 40% of the dump gas. The other 60% of the dump gas must then be the remaining inert gas that is left over after metabolism of the oxygen. If the diver is metabolising 3 litres/minute, then given that the drive gas is a 50/50 mix, the inert gas volume associated with 3 litres of Oxygen is also 3 litres. (Can you see why I picked 50 mix as the drive gas?). Now we can see that 3 litres is 60% of the total volume that must be dumped per minute.

?/3 =100/60
Multiplying both sides by 3
? = 3X(100/60)
? = 300/60
Dump volume equals 5.
or in other words, 5 litres of gas must be dumped from the loop per minute to keep the FO2 above 20%

I happen to have a tidal volume of 3 litres, so when diving 50% mix, I must dump twice per minute at the surface, and at 5/3rds of a bar (7m) I can dump once per minute. At 10/3rds of a bar (24m) I can dump once every other minute. This particular depth is rather awkwardly deeper than I'm really supposed to go on 50 mix, and in real life I usually use 60 mix and switch to a different drive gas below 18 metres. I also combine these procedures with the normal SCR safe practices such as a flush-through prior to commencing ascent and use of non narcotic mixtures.


In an automatic system the calculations are based on the fraction of each tidal cycle that is metabolised by the diver. There is a manufacturer at the moment who promotes a 26 to 1 ratio between tidal volume and volume metabolised. This equates to an Oxygen consumption of 3.85% of a tidal volume and is a normal human consumption rate. Divers however are notorious for maintaining exceptionally low ventilation rates, even with high workloads. You may be well advised to measure your own exhaled gas and compare the expired O2 level with inspired O2 level. I've done this experiment on myself, and found a worst case figure of a 10% difference. This was at the very end of an exhalation, and I would use 8% as my own worst case. My own open circuit gas consumption figures are not good however, and if you're a gas miser, you may wish to use a more conservative figure.

Calculations proceed in a similar manner to the manual system, except that we're not trying to find a litres per minute figure, but rather a fraction of tidal volume.

Using a 50/50 mix again, we can say that we will consume 0.08 of a tidal volume of oxygen, and the volume of inert gas associated with this will be the same (because of the fact that the drive gas has the same amount of oxygen and inert gas). Hence each dump must contain 0.08 of a tidal volume of inert gas. Again, we'll look at a minimum FO2 of 20%, so the dump gas must contain 20% Oxygen, and it's associated inert gas, another 20%. So 40% of the dump gas is wasted flushing gas. If 40% is flushing gas, then the other 60% gas must be the inert gas remaining after metabolism. You can arrange the figures in a similar way to the preceding calculation; 100/60 = ?/0.08 or multiplying both sides by 0.08 gives;
8/60 This is the fraction of each tidal volume that must be dumped overboard with each breathing cycle. To find the ratio of tidal to dumping, which is how these things are usually expressed, take the reciprocal, which is 60/8, or 7.5 to 1.

So for a 50/50 mixture, used from the surface, with a minimum FO2 of 20%, the minimum dump ratio is 7.5 to 1.

Now a fixed ratio rebreather will have the FO2 climb as the diver descends because the absolute amount of Oxygen metabolised remains constant while the quantity of gas that moves through the divers lungs and rebreather increases in proportion to absolute pressure. It is easy to calculate gas consumption ahead of time, simply calculate as for an OC dive, and divide by the dump ratio. Careful observers will note that this doesn't include the metabolised O2. A very rough correction for the diver's RMV can be made by adding the metabolised O2 to the dump volume, and recalculating the effective usage ratio. In the example above, the O2 consumption was 0.08. 0.08 plus 8/60 is 0.2133 and the reciprocal is 4.6875 at the surface. When the diver goes deeper, the oxygen consumption doesn't change with depth, so the advantage over OC increases. As the depth increases the advantage tends toward the dump ratio, but never quite reaches it. So for an automatic dumper, using 50/50 drive gas, the advantage over open circuit is between about four and seven times better.

A depth compensating rebreather will use the same gas at depth as it does on the surface, and the FO2 will remain constant. To calculate gas consumption, do the same thing, but there's no need to consider depth when calculating duration. The set will give an advantage over OC that is constant, and as though the diver remained at the surface. In this example the set would use the divers RMV divided by 4.6875 per minute. Of course none of this takes into account loop flushing and suit inflation, not insignificant factors with very small cylinders.

A partially depth compensating rebreather will fall part way between these two. Calculating gas consumption ahead of time is virtually impossible unless you know the exact dump ratio throughout the unit's operating range. Unfortunately construction of a truly depth compensating rebreather has proven a formidable challenge, and from the figures that get bandied about, the units which claim full depth compensation are in fact at best, partially compensated. The key to picking this out is claims that the unit will come closer to the drive gas FO2 as depth increases. This is of course counter productive for a single gas system, however, it has safety advantages, allowing a lower FO2 drive gas to be used at great depth.

In all, it can be seen that gas dumping rebreathers can offer many advantages over OC, SCR and CC units, however, they, like all units, have drawbacks as well. Like everything else, these units are not majic carpet rides and a good grasp of what is happening inside your unit is essential, both during the design phase and during the actual operation of the unit.

Return to Homepage

Return to index


(c) Halcyon

Closed circuit systems primary goal is ultimate gas efficiency.
Conventional semi-closed systems goal is gas efficiency and simplicity.
The Halcyon's goal is gas efficiency, simplicity, and diver safety

Halcyon Goals:

1) Maximize gas efficiency
2) minimize task loading, unit induced diver error, and hyperoxia/hypoxia
3) provide equipment warnings/alarms that cannot be ignored

Halcyon's goal was to produce a rebreather that was more efficient than
conventional semi closed units. Although a maximum gas efficiency was a
primary design criteria, Halcyon rejected the idea of developing a closed
circuit unit due to unacceptable task loading, potential for diver error,
and lack of intuitive alarm systems. Diver safety would not be sacrificed
for the last small fraction of gas savings.

Halcyon's design resulted in a semi-closed circuit rebreather that
approaches the efficiency of a closed-circuit unit but offers none of the
CCR's complications. It is capable 2 hours of operation at 300 feet on a
single 80cf aluminum cylinder. Between 60 and 200 feet, the Halcyon is 5 to
8 times as efficient as open circuit.


Unique among available rebreather designs, Halcyon's control system uses the physiological ratio between respiratory rate (RMV) and oxygen metabolism (VO2) to set the flow rate necessary for semi-closed operation. Workrate, respitory rate, and metabolic rates are linearly linked. Of each breath on the surface, 4% of the total gas inhaled is metabolized, which works out to a 26:1 volumetric ratio. The diver's metabolic oxygen requirements are not affected by depth, therefore the volumetric ratio changes in direct proportion to increase in pressure/depth, which is why the Halcyon utilizes a variable ratio. (see diagram)


When the diver exhales into the bellows, both the inner and main bellows
fill. The volume ratio of the inner to outer bellows sets the vent ratio, which in turn was determined by the 26:1 physiological ratio.
When the diver inhales, the inner bellows vents its contents to the water.
Fresh gas is injected through redundant regulators to replace the vented gas
on each breath.


The unique inner bellows also provides another distinct safety advantage. If
the regulators stop adding gas for any reason, the inner bellows will
continue to vent gas on each breath until the diver can't take a normal
breath. This intuitive alarm is the same as felt with open circuit
equipment, and the diver simply commences breathing from an alternate gas

By contrast, with conventional semi-closed rebreathers such as the Drager
Atlantis, the diver has little chance of recognizing that the gas flow
orifice is partially blocked during a dive. The result can be fatal: oxygen
will be depleted from the breathing gas until the onset of hypoxia. The
hypoxic diver will still be able to take a normal breath, but the
percentage of oxygen in the mix will continue to fall.

Similarly, should the oxygen addition system fail on a closed circuit
system, the diver has no reliable intuitive method of determining the
failure. The diver in this case is dependant on the electronic
warnings systems, which often fail at the same time as the oxygen addition
system. Even when all alarms are functional, many experienced, competent
divers have still become hypoxic due to "diver error," which really means
equipment-induced error due to lack of intuitive alarms.


At depth, the inner bellows of the Halcyon is mechanically squeezed by 2
levers, partially compensating for the change in pressure and maximizing gas
efficiency. The venting ratio varies between 4:1 on the surface and 10:1 at
300 feet. This variance allows the Halcyon to be efficient for diving to any
depth, both at the bottom depth and on deco, using the same Nitrox and
Trimix gases that would be used on open circuit.

Open circuit, closed circuit, and the Halcyon rebreather provide a
relatively constant gas mix independent of diver workload. Significantly,
fixed orifice conventional rebreathers such as the Drager Dolphin increase
the inert gas percentage delivered to the diver during elevated workloads,
resulting in a longer decompression obligation.

The Halcyon is different: If the diver works harder, his RMV will increase,
and the Halcyon will vent more gas proportionate to the diver's increase in
metabolic need. The Halcyon compensates for diver work rate and maintains a
decompression advantage over conventional semi-closed units.


The Halcyon utilizes a counterweighted bellows counterlung. The
ght eliminates the negative inhalation pressure normally
associated with back mounted counterlungs, such as on the MK15/16 military
rebreathers. In fact, the counterweighted bellows configuration has
outperformed all other types in US Navy work of breathing tests. Recent
testing by the US Naval Academy demonstrated that the Halcyon exceeded the
US Navy requirements for Resistive Breathing Effort (RBE) by 37% at an RMV
of 62.5 lpm.

Halcyon has engineered the optimal trade off between diver safety and gas
efficiency in rebreather design. Its ease of operation allows the diver to
concentrate on the dive rather than the equipment, and is the optimal choice
for those divers needing to concentrate on an underwater activity without
incurring significant task loading from their equipment. The Halcyon
rebreather has already proven that it has the "right stuff" for the
demanding work of underwater cave exploration; now the rest of the
technical, scientific, and exploration diving community are discovering the
benefits of an efficient, simple, and safe rebreather.

Return to Homepage

Return to index

Mandatory reading: Oxygen sensors


From the rebreather list (click here) by way of Walter Starck:


I have copied the recent sensor test thread to a friend, Bernhard Engl in
Munich, Germany. Bernhard is a senior electronics engineer with Siemens.
He is a specialist in electronic sensors and controls and an experienced rebreather
diver. He has also conducted extensive tests on oxygen sensors under
simulated diving conditions. He is not on the RB list. With his permission for
posting to the list, his comments follow:
Dear RB list:

Recently the issue of sensor testing and sensor storage
came up on the list, and most of it is wrong or misleading.
Some hints to get things right:

1. Galvanic cell type sensors come in a sealed container.
As soon as you open it, they come to life (some don't).
It takes a while for them to stabilize (=reliable readings).

2. Galvanic cell type sensors die for five main reasons:

a) The electrode material gets consumed by the
electrochemical process = normal end of life.

b) The electrolyte dries out.

c) The sensor gets poisoned: some contaminant enters
it and the electrodes get spoilt. They get covered
by unwanted by - products forming parasitic layers on the
electrodes. Most of the "dead on arrival" sensors from a)
belong to this group (contaminated during manufacture).

The remaining two are specific for hyperbaric environments:

d) The sealing system fails and the sensor leaks. Mainly
applies to sensors with gas spaces in the electrolyte
or with a pressure equalisation system not suited for
hyperbarics, such as intentional gas bubbles inside.

e) Gas bubbles build up in the electrolyte (DCS of the sensor)
and act to buffer the PPO2, or block parts of the electrode
surface or the membrane, yielding false readings, and may
lead to bizarre behaviour if the bubbles start moving.

3. The temperature compensation of most sensors is not suited
for rapid changes of temperature. Readings are only valid in
complete temperature equilibrium, which almost never happens
in RBs (lab or chamber experiments may be very misleading...)

4. Storage:

Keep the sensors in humid air and accept they won't live forever.

The humidity helps to keep them from drying out and the small
current from the air helps to keep the electrodes clean - by
consuming them, contaminants have less chance to build up
their destructive layers. The sensors stay "trained" and give
reliable readings instantly and not without the usual (and
unstable) wakeup period.

5. Pot Testing:

Have a computerized pressure pot which is able to do actual
dive profiles automatically. Decompression as for a diver
(faster would kill the diver, everything slower does not make any
sense at all). Record pressure, temperature, sensor output.
Have multiple sensors of the same type in the pot and compare.
Do hundreds of profiles from shallow to deep over a few months.
Have reasonable "surface intervals" for the sensors and observe
what happens during those, too. Have some humidity in the pot
and use stainless steel only. Analyze the data with tools like
MATLAB and assume calibration before every simulated dive.
Reject all sensors showing failures, gross errors, major changes
in step response, growth of nonlinearity, strange behaviour in
the surface interval, before reaching most of their rated lifetime
(this requires recording of the total oxygen dose normalized to
percent hours.)

6. Field Testing:

Have a modified RB running on open circuit so you know what
gas the sensors see. If they see fresh gas, you don't have
humidity and much lower gas temperature, if you have them
on the exhalation side, you have humidity (->condensation)
and higher temperature. If you have them on exhalation side
after a filled scrubber they even get hotter. O2 consumed by
metabolism can be treated as being constant per individual.
Record the data from your set of three sensors and analyze.
Readings still valid ? Sensors still track with each other ?
Reaction in thermoclines or when entering cold water from a
warm place or vice versa ? Recovery from flooding ?
Never attempt to use sensors on the real thing before you
are absolutely sure they will do.

7. Maintenance:

Note installation date on each sensor. After a third of the
service lifetime (which is a fraction of the rated lifetime)
replace the oldest one with a fresh one. This means you
always have a mix of sensor ages (or even manufacturing
runs) and this lessens the threat to run into a systematic
problem occuring at about the same timeframe. With this
scheme you don't waste more sensors than with replacing
all three at the same time after a full lifetime. The replaced
sensors still may be good for gas mixing before their full life
is over.

8. The big gotcha

Even if you have found sensors that appear to work for RB
(the Teledynes and their ripoffs, hint, hint) the manufacturer
might change their manufacturing process at any time and
unless they are specified for RB (which they aren't) he will
care a damn whether this engineering change kills you -
you are abusing the product, so who cares. You might be
better off if you use sensors which truly are specified and
made for RB. Beware of the garage outfits, though.

9. A related problem (although not about sensors):

It is nearly impossible to get electronic components specified
for use in life support equipment. Almost every component
manufacturer has a so called "life support policy" in the data-
books essentially saying they prohibit use of their products
in life support systems unless you get a written approval by
their CEO, which you most probably can't. The guy is no fool.
This does not mean those "industrial" grade chips won't work in
your RB, it's just another obstacle for those who want to sell
rebreathers and get rich by doing so.
You want to use cheaper "commercial grade" components ?
They are basically those who still limp along and are good for
disposable items such as electronic toys, TVs, etc., and can
suffer from dramatic performance loss at marginal supply
voltage or close to their rated temperature limits, so beware.

I hope this post helps somewhat to avoid the ongoing repetition
and duplication of common errors and mistakes as evidenced by
some previous posts on sensors and sensor testing.

Bernhard Engl


Return to Homepage

Return to index

 More: Oxygen sensors

Some recent posts on sensors on the RB list appear too enthusiastic to
me, so some words of caution may be in order. I have tested various
sensors for RBs during the last decade, partly for my own interest,
partly as a consultant for sensor manufacturers who were approached by
RB manufacturers and relied on judgement from a third party on whether
their products would be advisable for this type of application. I had
galvanic cell sensors, amperometric sensors, polarographic sensors,
paramagnetic sensors, etc., and ran tests on a computerized pressure
pot using real gas mixes and real diving profiles besides the stressful
spike exposures no human diver could survive.

One of the problems with galvanic cell type sensors is that with no
exception, all I tested did suffer some gas bubbles buildup in their
electrolyte, given sufficient number of pressure cycles, some types
more, others less. Sufficient amount of gas bubbles at the right
place on the electrodes and the sensor readings get position dependent
and this certainly is no feature you would bet your life on.

This is the reason why no galvanic cell sensor should be carried to
the end of its electrode life in a RB, and why the sensors should be
arranged and replaced such that the three never have about the same
accumulated prior exposure. Just note the installation date on each
of them and after a third of their "RB life", replace the oldest one.
The "RB life" is much less than the "normobaric life". Lacking official
manufacturer specifications, I propose this should be correlated to
total accumulated time under water and pressure pot experiments watching
out for signs of trouble under controlled conditions.

One specific issue with those gold-cadmium cells which happen
to have a second pressure equalizing membrane beneath the circuit
board. Watch out for those: Their cadmium electrode oxidizes to
reddish crystals. Now, the pressure equalizing membrane (usually
on the other side as the oxygen conducting diaphragm) is squeezed
against sharp cystallic protrusions when the unit is pressurized.
It is yet unclear how many cycles are required until this membrane
fails due to material fatigue i.e. puncturing. This failure
mechanism is not present in fresh units, so just a few pressure pot
cycles are absolutely meaningless and prove nothing.

Tests must be done over the full service lifetime over a significant
number of sensors and then each one must be carefully disassembled
and inspected before any conclusion on fitness for RB use is drawn.
The "RB life" figure also could be derived from such experiments.

A related failure mechanism: The smart guys who solder wires to
the pins and then fill the back side of the sensor with epoxy to
keep humidity out. This spoils the said pressure equalizing system,
and causes mechanical deformation and failure of the sensors under
hyperbaric conditions.

All the trouble with the second pressure equalizing membrane could be
avoided when its job is taken over by the front side oxygen conducting
membrane. Usually there is another spacer diaphragm or mesh between
the gold electrode and the cadmium electrode and enough distance from
the crystals on the cadmium electrode to be safe. This solution might
mean a larger diameter of the oxygen conducting membrane compared
to the industry standard form factor.

This solution of course does not solve the problem of gas bubble
buildup in the electrolyte. Anyone out there who can prove the contrary
on some sort of "miracle sensor" ? Just send me a handful of free
specimen, I will be glad to run them thru my torture cabinet. Notice:
the sensors will be destroyed in the process and won't be given back.

I can't tell whether these mechanisms apply to the Teledyne sensors, as
I did not test or disassemble any of those, and I certainly don't want
to scare anyone. But all of you out there who bet their life on sensors
should know that certain failure mechanisms exist, most sensors are not
made for hyperbaric applications, they are just misused for RBs, and a
few dozen daredevils who survived until now do not prove it is OK to
use ordinary sensors not explicitely specified, designed, manufactured,
and tested for use in RBs. Always ask yourself how many owners of modern
electronically controlled CCRs are already dead:

Two percent ? Five percent ? You still want to push sensor life ?

My advice:

Throw them out as soon as possible, put in a fresh one as often as you can
afford it. Just compare with the other expenses (adsorbent, gas, logistics,
travel) and don't try to save money by pushing sensor life. It's the sure
road to hell, regardless of the quality or lifetime of the same sensor
under normobaric conditions.

Bernhard Engl

Return to Homepage

Return to index

Open circuit

Have been thinking about something Rod Nairne posted recently:

> As for your opinion that you should only use a breather when it is really
> needed, I tend to agree with you, I have started to dive OC again and it is
> a lot less stressful. Especially here in the US where helium and nitrox
> fills are so readily available and cheap.

I have to agree. Once the novelty of RB diving wears off you can really begin to
appreciate how carefree and enjoyable OC really is. With RBs you also spend more time
dicking around with equipment than you do diving. Where you need them they are great but
it is even better when you don't.

Best regards,
Walter Starck

Return to Homepage

Return to index