You have 35 seconds - starting now

Do you dive using manifolded twin cylinders with an isolation valve If so, how fast can you isolate and shut down a cylinder in an emergency And would you bet your life on doing it in the time asks John Liddiard, who has been conducting some experiments

TWIN CYLINDERS ARE PRETTY MUCH STANDARD kit once we get on to anything more than trivial amounts of decompression, the sort of dive that many divers do on 30m-plus wrecks every weekend.
     A decompression ceiling, like a wreck-penetration or cave-dive, should be treated as no clear surface. It is not quite as severe - you can always be recompressed, but you cant be undrowned - but the redundancy provided by two cylinders, two regulators and good gas-planning should enable a wide range of incidents to be survived without decompression illness or drowning.
     There are three basic ways to configure twin tanks: independent cylinders with independent regulators; manifolded cylinders; and manifolded cylinders with an isolation valve.
     With independent cylinders, a diver needs to swap regulators to balance between them the gas used. At any point in the dive either cylinder should have enough gas to get back to the surface should the regulator on the other one fail.
     On a dive planned by the rule of thirds, this amounts to never letting one cylinder get more than a third ahead of its counterpart. In practice, most divers seem to keep a little extra margin, perhaps aiming not to let one cylinder get more than 50 bar ahead of the other.
     With manifolded cylinders, a diver breathes off a single regulator for the entire dive, except for a few breaths on the back-up regulator to check that it is working near the start of the dive. There is no need to swap regulators, as the two cylinders are interconnected.

Many technical divers frown on manifolds without isolation valves, on the grounds that a failed joint in the manifold would drain both cylinders. For that reason I wont consider them further, other than to advise that targets for shutdowns can still be estimated by a similar process to the one I am about to describe.
     Even if the manifold has an isolation valve, following a regulator failure with gas escaping, both cylinders will lose gas until the isolator is closed; then the cylinder with the failed regulator will continue to lose gas until the cylinder tap is closed. Once everything is stable, the isolator tap can be re-opened to enable the working regulator to draw gas from both cylinders.
     The worst case for an independent twin-set should be that at any point in a dive the remaining cylinder will have just enough gas for the diver to get back to the surface. So the target for a manifolded set should be to be able to isolate and shut down, at any point in the dive, before half the remaining gas is lost.
     If the isolate and shutdown can be made faster, the diver with a manifolded set has a better safety margin. And if the isolate and shutdown is slower, there might not be enough gas to return to the surface. So a diver with an isolation manifold needs to practise isolate and shutdown drills to make sure that in a real situation he gains rather than loses.
     With that in mind, just how fast does an isolate and shutdown need to be I set out to take some measurements. Or rather, I co-opted Steve Chaplin to help me take some measurements.

I couldnt do it without help because I have always used independent cylinders. Its a personal thing. I like being able to break gear down and reconfigure it easily. Also, Im not very flexible physically - I am not sufficiently proficient at an isolate and shutdown drill to bet my life on it.
     On the opposite side of the fence, Steve has a set of 12 litre cylinders with an isolation manifold, probably the most common set-up among those who dive with manifolds. He is also quite flexible and proficient at gas shutdowns.
     Our experiments began fairly simply in the car park outside our clubs compressor room. Steves rig consists of 12.2 litre 232 bar Faber cylinders, Scubapro cylinder valves and isolation manifold equipped with Sherwood rubber knobs for easy shutdowns, and DIN-fitting Scubapro Mk 20 demand valves, one with a long hose, one with a pressure gauge.
     The only modification we made was to add a second pressure gauge so that we could measure each cylinder pressure independently.

Steve rigged his twin-set with regulators. We doctored it to simulate a range of gas-losing failures, and measured how much gas was lost from each cylinder over 15 and 30 second periods.
     The scenarios were:

  • Uncontrolled free flow (by holding the purge on a second stage)
  • Low-pressure hose failure, on both the long hose and the short hose (by disconnecting a second stage from the hose)
  • High-pressure hose failure (by disconnecting the pressure gauge)
  • High-pressure O-ring failure (by cutting a piece out of an O-ring).

We discovered that the worst-case scenario was a burst lp hose, and that it didnt matter whether it was a long or short one. It could actually blow a tank down faster than just opening the cylinder valve to the air with no regulator attached!
     Rounding to the nearest decimal, the loss rate varied from 32.5 litres/sec for cylinders two-thirds full to 20.3 litres/sec for cylinders at 60 bar. In the middle at 100 bar the loss averaged 24.4 litres/sec.

Our other initial finding was that the manifold did not restrict flow between cylinders at all. Even when the isolation valve was cracked only a half-turn open, both cylinders lost gas equally until the isolator was closed. Only just cracking the isolator might knock a few seconds off your shutdown time, but it will not significantly restrict the loss of gas across the manifold.

Knowing that the loss rate varied with cylinder pressure, would it vary with depth We conducted a further experiment to check this, carrying a spare cylinder and regulator on a dive and letting it blow at 30m. The rate of gas loss was unchanged.
     I put the data together in a spreadsheet to calculate target shutdown times for various cylinder sizes and pressures. The results are displayed as bar charts for cylinder sizes from 7 to 15 litres, with TI being the time taken to close the isolation valve and each column showing T2, the additional time allowable for shutting down the relevant cylinder valve.
     The first chart (see overleaf) is for cylinders at 140 bar at the time of failure. The 1/3 in point is on rule of thirds, where half the remaining gas is needed to get out safely.
     The second chart is for cylinders at 100 bar, selected as typical of the point at which an open-water dive would be planned to begin ascending if decompression stops were to be made on the main gas, without stage cylinders or separate deco gas.
     So for a diver with twin 12.2 litre cylinders, if the time to isolate (T1) was 15 seconds, the diver would have an additional 35 seconds to close the relevant cylinder valve (T2) before crossing the point at which he would have been better off without the manifold. Or, in the worse case, an additional 35 seconds before he condemned himself to missing stops or drowning.
     That doesnt sound too arduous. After all, you practise gas shutdowns regularly and have no difficult achieving these times. But incidents arent like practice exercises. They hit you by surprise at the most awkward time.

You are at 35m, slightly fuzzed with nitrox. You are engrossed in whatever you are doing - laying a line, measuring a porthole, teasing a lobster out of its home, taking a photograph, swimming through a wreck, sending up a delayed SMB. Suddenly there are bubbles all over the place.
     No matter how experienced you are, your first reaction will be one of surprise and shock. It will take a few seconds to gather your wits enough to realise that you need to drop whatever youre doing and shut down the isolation valve. You then need to fumble around and close the isolator. All the time bubbles are gurgling around all over the place. You hit the silt and visibility drops to zero.
     Even so, all that practice pays off and you close the isolator in just 5 seconds longer than your usual practice time.
     So the isolator is now closed, but bubbles are still streaming out from somewhere. You need to decide which cylinder tap is controlling the offending burst hose and close it down. In the process you might need to swap regulators.
     And all the while the stress is building.
     Despite all these distractions you get the cylinder tap closed, just about in time. Congratulations, youll live through the incident.

But suppose it takes a little longer Suppose in all the confusion you cant work out which cylinder valve to close and have to work by trial and error Suppose an lp hose failure occurs when you have less gas to start with
     I dont claim that our experiments or analysis is definitive. The numbers might come out slightly differently for other makes of manifold or demand valve, so you might wish to allow an additional margin, or take measurements of your own, at least enough to confirm that you can trust our results. I would aim for an additional 10 second margin on closing the isolation valve and 15 seconds more to close the cylinder valve.
     If you have a twin 7 litre set with an isolation manifold, perhaps you should re-think your dive strategy to keep the manifold closed. Treat the set as independent singles, swap regulators during the dive, and save the isolator for opening after the cylinder valve for an offending regulator has been shut down.
     If you have a twin 10 litre set, you need to be very proficient to make the manifold an advantage rather than a disadvantage. Perhaps you should be keeping it closed as I have suggested for twin 7s.
     If you have a twin 12 litre set, and if you are more flexible and more practised than I am, you might feel happy enough with your shutdown times. Its worth noting that although the various training agencies differ in their practice exercises and targets for isolate and shutdown drills, they all come inside the critical time shown for twin 12s on the charts.
     With a twin 15 litre set or larger, the margins dont look so tight. But dont become complacent.

One way to speed up shutdown times is with a remote-control knob on the isolator valve, so that the manifold can be isolated without having to reach over your head. For the inflexible diver, such a slob-knob certainly makes the first step of isolating cylinders easier, but you still have to reach back to close down the offending cylinder tap.
     Of course, you could also fit a slob-knob to each cylinder tap, but the routeing of all these knobs could then end up over-complicating things in a different way.
     You could also try mounting cylinders upside-down, as the police and military divers do. A few minor problems such as protecting valves while putting the rig on and routeing hoses need to be sorted out, but there are many tried and tested solutions. Some divers claim they can reach the taps more easily this way. Others actually find it more difficult.
     Whatever configuration you adopt, practise in controlled conditions. To benefit from a manifold, you need to be able to complete an isolate and shutdown drill fast enough to save sufficient gas to make it back to the surface. Overall, you must decide on which margins you are happy to bet your life.

BSAC, Jack Ingle: There is no hard performance requirement in the BSAC extended-range course. Personally, I aim to be able to isolate and shut down in 10 seconds.

PADI/DSAT, Mark Caney: Equipment requirement is for a minimum of twin 12 litre cylinders. The performance requirement is for a combined exercise where the diver fully closes and re-opens the isolator valve and both cylinder valves within 45 seconds, switching regulators as necessary.

IANTD, Kevin Gurr: My target is 15-20 seconds for a complete shutdown drill. That is: close the centre, close number one, switch to number two, open number one, switch back to number one, close number two, open number two, open centre, check number one and two.

TDI, Dave Crockford: A training skills circuit of closing and re-opening all three valves, switching and checking regulators as needed, should be completed within two minutes. Simple division gives 20 seconds to close each valve, though I would expect divers to be faster than this.

Divernet Divernet
1 Reaching for the isolator valve in the centre of the manifold
2 First close that isolator valve...
3 ...then close the cylinder valve feeding the failed regulator
4 Letting the hose blow for 15 and 30sec intervals. Note the foot holding the hose down safely
5 Simulating a high-pressure O-ring failure by cutting a quarter away

Start a Forum discussion on this topic