Every CCR dive has a quiet, binary question running under it: do you still trust the loop? When the answer is no, the right move is to switch to open-circuit bailout immediately. The exact failure can be diagnosed on bailout while you climb the ascent plan with your team. The diver who first asks whether to bail out has usually already crossed a threshold that should have produced a bail.
This article is about that threshold. It covers the failure signatures that justify an immediate bail, the cell-voting and CO2 evidence that pushes the decision, the PO2 spike scenarios that force the call, and the procedure for getting off the loop without compounding the original problem. None of this replaces formal CCR training. It is a working framework that supplements what you already learned about loop failures and bailout planning, so the next time the loop hands you ambiguous information at depth, the call comes out faster.
What Counts as a Real Reason to Bail?
The mantra most instructors hand new CCR divers is straightforward: bail out now, diagnose later. The phrase exists because divers naturally want to debug the unit they spent years learning to trust. At thirty meters, with a confused PO2 reading and the first edge of a CO2 headache, the temptation is to keep working the problem. That instinct is exactly what gets people killed on rebreathers.
A real reason to bail falls into two categories. The first is the hard bail, which is any failure where the loop is no longer breathable or the electronics have lost the ability to manage gas. A flooded loop, a scrubber that has stopped working, a solenoid stuck open at depth, a cell disagreement you cannot resolve, or a verified high or low PO2 outside the working envelope are all hard bails. There is no diagnostic step before the switch. You bail, signal your team, and start the bailout ascent.
The second is the soft bail, which is the decision to bail when the loop is still technically functioning but something feels off. A working CO2 monitor that has thrown an early warning, a temperature stick moving sooner than the dive plan predicted, a manual flush that did not seat clean, a controller that briefly reboots and recovers. These are situations where the loop might be fine, but the diver’s confidence in it is not. Soft bails are a judgement call. Most experienced CCR divers will tell you they bail soft about as often as they bail hard, and they trend conservative: if the dive plan is replaceable but the diver is not, the cost of a precautionary bail is low.
How Do You Read a Cell Disagreement?
The oxygen partial pressure displayed on the handset is a voting average from your cells, typically three of them. As long as the three cells agree within a tight tolerance, the controller can trust the reading and the solenoid fires accurately. The disagreement that should worry you is not a one-millivolt noise floor difference. It is the cell that drifts visibly low while the other two hold their position, or the cell that pins itself at the same reading no matter what you do to the loop.
A divergence beyond roughly 0.05 atmospheres at depth is a clear early warning sign. A divergence beyond 0.1 is operationally critical and means the controller’s voting average is no longer reliable. At 0.2 or more, you are flying blind on the gas the unit is making. The average could land high enough to push you into toxicity range or low enough to put you near hypoxia, and you have no way to know from the display which scenario is unfolding.
The cell that stops responding to a manual oxygen addition or a diluent flush is telling you something specific. A current-limited cell will read the same value before and after the flush even though the actual PO2 changed. A voltage-pinned cell behaves the opposite way and saturates at its calibration ceiling. Both modes mean you do not have a working sensor in that slot. Field experience helps here, but the durable rule is that any cell which fails a manual flush response check is a bailout trigger, not a “let’s see how it looks at the next stop” decision. That is also why CCR divers stagger their cell rotations rather than replacing all three at once. The way oxygen sensors age between their stamped service intervals is the exact reason staggered rotation gives you a useful cross-check at sea level that you cannot get when every cell was installed the same week.
When Does CO2 Tell You to Bail?
Carbon dioxide buildup on a CCR is a different decision than cell disagreement because it does not show on the handset. The diver is the primary sensor. A bailout-triggering CO2 picture has a recognizable shape: a headache that came on quickly, an air-hunger sensation that goes away only when breathing rate climbs, and the specific tightness of breath that does not respond to a slow exhale.
The single most useful rule for this decision is plain: if you have asked yourself whether you might be hypercapnic, you already have enough evidence to bail. A clean scrubber and a healthy loop do not produce that question. By the time the question is forming, the answer is on the way to yes. Bail off the loop, signal your team, breathe open circuit for a few minutes while the symptoms wash out, and let the dive plan handle the rest.
A scrubber temperature stick that moves faster than your runtime prediction is a separate, mechanical warning. Cold water pushes the reaction front through Sofnolime faster than the manufacturer chart suggests. A leaky scrubber seal short-circuits the bed. A poorly packed canister channels at the wall. Any of these can cut the working duration in half without warning. Some divers bail at the first anomalous stick reading. Others continue under a tight watch with a teammate observing. The right answer depends on the dive’s depth, the team’s experience, and how recoverable the planned ascent is from the current depth.
Recognizing the earliest symptoms of CO2 buildup on the loop deserves the same priority as any other lock-out failure mode, because hypercapnia degrades judgement faster than nearly any other dive emergency. By the time you are mentally arguing about whether the headache is from cold or from CO2, the part of your brain that should be making the bailout call is already operating with less oxygen-rich blood than it had a minute ago.
How Does a PO2 Spike Force the Call?
The opposite end of the cell disagreement problem is a verified PO2 outside the working envelope. The handset displays a reading that the cells actually voted for, the cells are not disagreeing, and that reading is no longer where it should be. This is rarer than cell failure but more dangerous because the controller is doing exactly what you asked it to do.
A solenoid stuck open will push the loop PO2 upward as the unit injects oxygen the diver no longer needs. The display reads high, the cells vote correctly, and the diver experiences the early signs of central nervous system toxicity such as facial twitching, an unusual taste, or anxiety that does not match the dive. The right response is to manually close the solenoid by removing electrical power to the controller, flush the loop with diluent to bring PO2 down, and bail off if the manual control authority does not feel decisive within a few breaths.
The lower-PO2 side of the spike is more insidious. A diver who descends without switching from a surface diluent, or who runs the loop after a manual oxygen addition has lapsed, can find themselves breathing a hypoxic mix without symptoms until they lose consciousness. The bailout trigger here is the handset reading itself: any verified PO2 below the unit’s hypoxic warning threshold is a hard bail. A diluent flush is a diagnostic step, not a recovery. If the loop returns to a safe PO2 after the flush, you can think about staying. If it does not, you are already on bailout.
Tracking how oxygen toxicity builds across a CCR dive is the slower-burn counterpart to the acute spike scenarios described above. A diver who has been holding a high setpoint for a long deep dive can be close to a CNS limit before any single PO2 reading looks alarming, and the right call in that scenario is to drop the setpoint and rebalance the schedule rather than wait for a spike to force a bail.
How Should You Bail Without Adding Risk?
The act of bailing off the loop is the simplest part. The mouthpiece either has a bail-out valve that switches you to open-circuit gas without removing the mouthpiece, or it is a dive-switch valve that requires you to remove the loop mouthpiece and replace it with an open-circuit second stage. Practice this until it is automatic, because the first time you do it underwater should not feel like the first time.
The harder part is the team protocol. Your bailout has to be visible to your teammates within a few breaths. The standard signal is a hand-on-mouthpiece motion combined with whatever the team’s pre-agreed “switching to bailout” gesture is. Your team should immediately verify that you are stable, get close enough to share gas if your own bailout supply fails, and start the ascent with you. A solo bailout is a contingency for a team that has lost cohesion in current or visibility. It is not the plan.
The ascent itself should be deliberately disciplined. Whatever caused the bail has not stopped affecting you the moment you switch gas. CO2 takes minutes to wash out. CNS toxicity does not resolve on the way up; it requires depth reduction and time. Cell disagreement does not change the actual gas the body just absorbed, so the half-times still apply for as long as they apply. Move with the bailout schedule, not against it.
Knowing the math behind how much bailout gas you actually need is what determines whether the ascent you are about to take is safe at the depth where you made the call. A bailout that runs out of gas during the deepest decompression stop is no longer a bailout. It is a different emergency, and the work to avoid that scenario happens at the dive planning desk, not at depth with a confused controller in your hand.
How Does Silent Diving Help After a Bailout?
A precautionary bail and a hard bail both deserve a unit inspection before the next dive. The diver does not always know which failure mode caused the call, especially on a soft bail where the symptoms ambiguously resolved on open circuit. Routine maintenance does not always catch what caused a bailout, which is why Silent Diving’s authorized service team runs a dedicated post-incident inspection that covers cell rotation and verification, solenoid actuation testing, scrubber canister and Sofnolime inspection, controller firmware diagnostics, and a leak-down check on the loop seals.
Bring the unit in with the dive details documented while they are fresh. The bailout depth, the elapsed runtime, the symptoms you noticed, and the order in which the handset displays changed are all useful inputs for the technician. The goal is not to assign blame to a part. It is to make the next dive safer than the one you just bailed out of, and the conversation with the technician usually surfaces a small habit or torque value that prevents the same call from repeating.
Frequently Asked Questions
What is the difference between a hard bail and a soft bail on a CCR?
A hard bail is any failure where the loop is no longer breathable or the controller can no longer manage gas safely. A soft bail is the decision to leave a functioning loop because the diver’s confidence in it has been compromised by an ambiguous symptom or controller behavior. Both end the dive on open-circuit bailout. Hard bails leave no room for diagnosis. Soft bails are judgement calls that experienced CCR divers tend to make conservatively, because the cost of an unnecessary precautionary bail is much lower than the cost of staying on a unit that turned out to be failing.
Can you safely diagnose a CCR problem before bailing out?
No. Diagnosis happens on bailout while you ascend. Diagnosing on the loop assumes the loop is still trustworthy, which is exactly what is in question. The cells, scrubber, controller, and your own neurology are the same instruments you would have to use to evaluate whether the unit can be trusted, and a failure in any of them is what triggered the question in the first place. Move to bailout, then think about what happened.
How do you signal a bailout to your CCR team?
The standard signal is a clear hand-on-mouthpiece motion combined with the team’s pre-agreed bailout gesture. Within a few breaths the rest of the team should be at arm’s reach, verifying your stability, confirming your bailout gas, and starting the joint ascent. The signal is intentionally visible rather than subtle, because the worst time for nuance is a teammate who is two minutes away from losing consciousness.
What if you bail out and the problem turns out to be minor?
You finished the dive on bailout and inspected the unit afterward. That is the correct outcome. A precautionary bail that resolves to a non-event is cheaper than the alternative. CCR divers who routinely tell themselves “it was probably nothing” eventually meet a dive where it was not nothing, and the cost of that single dive overwhelms every precautionary bail they avoided over the prior decade.
Should you ever rebreathe a flooded CCR loop?
No. A flooded loop cannot produce predictable PO2 control because the cells are no longer seeing dry gas, and the wet scrubber reaction can liberate caustic gas that is itself a serious injury hazard. Bail off, complete the ascent on open circuit, and have the unit professionally serviced. A loop that has taken water at depth needs full disassembly and inspection, not a surface rinse.
How much extra decompression does a CCR bailout add?
It depends on the depth at which the bail happened, the bailout gas selection, and how much of the planned decompression the diver had already completed. A trimix bailout at depth followed by an open-circuit nitrox stop ladder typically adds modest decompression time compared to the original CCR profile. The dive computer needs to be configured to handle the gas switch cleanly, which is one of the reasons configuration on the wrist computer matters as much as configuration on the unit handset.
Do solo CCR divers use the same bailout criteria?
Yes for the criteria themselves. The solo CCR diver carries more redundancy, usually two complete bailout gases and a second oxygen control path, but the threshold for switching to bailout is at least as low as for a team CCR diver. Without a teammate to verify symptoms and share gas, the solo diver has even less margin for staying on a questionable loop. The right rule for the solo diver is the same as for the team diver, applied earlier.
Need help applying this to your own CCR setup?
Talk with Silent Diving before your next dive.
Get practical guidance on AP Diving products, rebreather service, parts, training, and planning support from the Silent Diving team.