Aqualung Legend LX First Stage Failure at depth

[databob]

Sorry for so many posts. Here is a DAN incident report I just came across - "Air supply cuts off when regulator failed"

Divers Alert Network, Air supply cuts off when regulator failed

In comments - "This particular first stage has an Automatic Closer Device to seal the first stage when it is taken off the tank, to prevent water getting into it."

3 people, all with automatic closure devices, lose all air suddenly at depth. I don't know the specific cause and don't really care what it is, but something is not right with this. In the DAN incident report this diver could not get to his buddy and went to the surface from 53 feet with no air. Also note in the comments - his air slowly returns later and he has about 1/2 to 1/3 tank - sounds familiar.

Including me - 3 exact incidents of no air at depth. That just can't be good.

 

[lexvil]

While I understand the concept behind the ACD, I just don’t want to dive anything with something like this in my air circuit. The failure rate seems to be statistically very low, given how many AL Regs I see in use it’s to many for my taste, once a regulator gives me pause on its performance I stop using it and if I have the same problem with more than one in the same model line I move on.

 

[alvinsuper]

Sorry for so many posts. Here is a DAN incident report I just came across - "Air supply cuts off when regulator failed"

Divers Alert Network, Air supply cuts off when regulator failed

In comments - "This particular first stage has an Automatic Closer Device to seal the first stage when it is taken off the tank, to prevent water getting into it."

3 people, all with automatic closure devices, lose all air suddenly at depth. I don't know the specific cause and don't really care what it is, but something is not right with this. In the DAN incident report this diver could not get to his buddy and went to the surface from 53 feet with no air. Also note in the comments - his air slowly returns later and he has about 1/2 to 1/3 tank - sounds familiar.

Including me - 3 exact incidents of no air at depth. That just can't be good.

Mares has its own version of the ACD called the AST. I didn't know Aqua Lung had that as well until I stumbled on this thread. When I learned about Mares AST, I remember being diffident. Now, the evidence is piling up, apparently. I think I'll stick to the old thumb-on-the-inlet secret technique to save my first stage from water intrusion.

Glad you're fine.

P.S. This is the Aqua Lung ACD explained


And this is Mares AST

 

[rsingler]

I have watched this thread for two years, and have commented elsewhere about the ACD.
@Diving Dubai has some strong feelings about the ACD, and lots of opinions have been rendered about the cause of this sudden gas shutoff posted by @databob:

I want to thank @databob for his persistence in posting relevant details about this incident. Let no one be mistaken, this sudden shutoff was real. All the posts saying it doesn't make sense, or didn't happen, are head-in-the-sand rejections of a decidedly uncommon, but clearly documented (thank you, @databob !) event.
The pdf he attached was my "lightbulb moment". I had been following Aqualung's safety notice about the shutter valve torque, and I have concluded that we were all handed a red herring. It has nothing to do with shutter valve torque.

(EDIT: The entire discussion of mine that follows is wrong! The discussion is interesting, but my initial conclusion is incorrect. It took several weeks and many more posts, and some discussion, but a back channel communication from Aqualung revealed the truth that @LuisH figured out on his own mute than a year ago in this thread. It's a block of a 6 mm insert on a tank valve. Sorry for the spoiler. Now read on and enjoy! Way to go, Luis!)

I think I have the explanation, if you'll bear with a long-winded description.
This is a problem ONLY with the yoke version of the ACD, as the design of the DIN version is different. So all the DIN comments are essentially irrelevant.
First, let's review the part names, so we're all on the same page:

As @Diving Dubai has pointed out, we still need good o-ring seals for gas flow to happen before the failure. That led me to conclude that it couldn't be a loose shutter valve, because the lower oring sealing the Shutter Valve to the Yoke retainer would rise out of its sealing position.
In this diagram, the red arrow points to that o-ring. A loose shutter valve would result in that part being higher, and potentially above the sealing land of the Yoke Retainer, resulting in an audible leak.

Additionally, if the shutter valve were loose (and unscrewing it places it higher in the Yoke Retainer), then the sealing point of the ACD (green arrow) would also lift up, opening the ACD. Hence, no blockage to air flow. My conclusion? The ACD "fail-closed" problem is NOT a function of shutter valve torque or loosening. My opinion only.

Next consideration: yoke retainer torque or loosening.
A lot of discussion revolved around the fact that if you grab a first stage that's pressurized, and twist it, you can unscrew the yoke retainer from the body of the first stage, because the retainer is fixed by the tank pressure, but the regulator body is subject to unscrewing if inadequately torqued on servicing.
That's a fact. But again, the seal of the yoke retainer to the body of the regulator is a face-sealing o-ring that will leak like crazy if unscrewed. While that would cause an airflow loss to the first stage, it would also be completely obvious. As @databob has noted, the problem was reproduced on the boat with NO LEAKS. So it's not a loose yoke retainer. Evaluation of the engineering of the ACD also confirms that unscrewing the yoke retainer from the reg body would have NO effect on the ACD.

Next consideration, and the one I think is key to what may have happened:
Look at the three pictures harvested from @databob and the video posted by @alvinsuper above (thank you!):
Picture #1: Loose yoke; fully extended shutter crown spring; CLOSED ACD!

Picture #2: Tight yoke; fully compressed shutter crown spring; OPEN ACD!

Picture #3: Transition state: Yoke being tightened, shutter crown spring JUST coming under tension, but ACD STILL CLOSED!

Now since this last picture was captured from the video, and has a little stutter in it, it's easier to look at this state using a previous picture rotated to the same position:

Here we see a CLOSED ACD, and an incompletely compressed Shutter Crown.

This is where I want you to bear with me:
IF we could get a tank sealed to the Shutter Crown in this position, the tank wouldn't leak, but the ACD would be closed.
If we added just a HAIR more compression to the Shutter Crown, the ACD would open slightly, and everything would be okay. Or would it???

Now I need you to think about your various tank valves.
Some have a "land" (the carved out area where the valve o-ring sits) that is very shallow. Tank valve o-rings for those just seem to sit on the surface of the valve, and are the designs where the o-ring falls out easily. We all know them. Getting a seal requires a definitive tightening of the yoke knob onto the tank valve. Those AREN'T the valves I imagine are the problem.
Better valves have a deep land, with the o-ring deep in the land. You can occasionally hear the o-ring seal with this type of valve: there's a very brief "pffft!" and a tiny "pop!" and the o-ring seals. How? Instead of a face seal requiring significant pressure on the top surface of the valve o-ring, the o-ring instead seals up against the junction of the shutter crown's flat ring face, and the side of the oring land in the tank valve. O-ring distortion makes the seal, which is why duro 90 is specified for these high pressure seals: the o-ring is resistant to extrusion after distortion.

We are all taught in our initial Open Water training to NOT overtighten the yoke knob. Why? Screw it down tight, add 3000 psi on a knob screw that's green with verdigris corrosion, and it's VERY difficult to get your reg off the tank after you depressurize. Instead, we're taught to use "just three fingers" to tighten the knob, and depend upon the o-ring seal to make the connection. And indeed that's what happens. The tank seals to the reg, and after it's pressurized, it's very hard to move anything, no matter how lightly you tightened the knob.

So far, so good.
Now I want you to think about "sliding parts" and sea water and a little verdigris corrosion. Where is that obvious? At the sleeve of the LPI connection between your inflator hose and your bcd. Sometimes that lightly spring loaded sleeve just won't move smoothly. It "catches", due to a grain of sand, or corrosion, or previous abuse.
Well, the Shutter Crown inside the Yoke Retainer is a similarly lightly spring loaded sliding connection.

Imagine that you're screwing down your yoke knob, and you don't want it to be too tight, and it seems to firm up and you stop screwing in.
But imagine that it "firmed up" because the shutter crown caught just a little against the inside bore of the yoke retainer. Maybe a salt crystal, maybe sand, maybe a ding inside the yoke retainer from a previous attempt to tighten the knob when the reg wasn't centered on the face of the valve and it pushed asymmetrically until the problem was discovered and fixed.
So now the tank valve o-ring is resting just lightly against the face of the Shutter Crown. There's JUST ENOUGH compression of the light spring under the Shutter Crown to open the Shutter valve a hair. And due to the excellent design of the tank valve o-ring land, when you pressurize, the valve o-ring seals up nicely against the face, and there's no leak there. The Shutter Crown seals against the tank valve not due to the pressure of that light spring, but due to the 3000 psi inside the compartment of the Yoke Retainer/Shutter Crown assembly, which forces both o-rings against their lands.

So you've got air flow, and all seems okay. But the Shutter Valve is just BARELY open. If you'd done an SPG check during a high-flow inhalation, you might have seen a dip, or you might not.

Now you're diving, and there are maybe 15 pounds of pressure against the tank valve o-ring (3000 psi x 0.05 sq in oring face area). Fifteen pounds pressing on that rubber for 10...20...30 minutes.
And at some point, that o-ring flattens just a hair. The Shutter Crown settles a hair deeper into the o-ring land of the tank valve. And what happens at that point?

As the Shutter Crown rises imperceptibly, the Shutter Valve closes that last little bit, and your air flow stops.

 

[rsingler]

Part II, continued:

When you bolt to the surface, and check the problem out, there's indeed no flow, because the Shutter Valve is closed. But it's only barely closed, so over the next 15 minutes, the tank air slips imperceptibly by the valve, the SPG pressure rises again. But when you take one or two breaths, it's just like the exercise where your tank valve is barely cracked open, and you suck the pressure down immediately.

Yeah, when it goes to the shop, maybe the Shutter Valve was a little loose, maybe the Yoke Retainer was undertorqued. But as I have suggested, I believe both those findings were red herrings.
Just once, the reg was attached lightly to the tank, and it seemed okay because the knob started to tighten up as the Shutter Crown caught on the inside of the Yoke Retainer bore.
But I'd be willing to bet that if you reattached the reg to the tank, ensuring full depression of the Shutter Crown into the Retainer, you would be unable to duplicate the problem.

So IF this is a correct analysis of the issue, what are the preventative steps?
1) Rinse your first stage religiously. Salt water penetrates behind the Shutter Crown and may just fester there. Indeed, this may be a reason to NEVER use the plug/cap between the knob and the face of the Shutter Crown, because that will compress the crown, making it harder to diffuse salt out of the tiny compartment where the spring lies. Instead, maybe depend upon the ACD to keep water out of the internals, and soak your first stage in fresh water overnight after a trip, so that fresh water percolates down into the Shutter Crown spring area. I dunno.
2) When you attach your reg to a tank, watch for the subtle depression of the shutter crown as it mates to the tank valve. Get used to how far it should depress. Practice with the cap while watching the shutter crown. It should move about 2 mm, if I recall correctly from servicing them. If it doesn't, screw and unscrew the cap to see if the crown "catches" as it is depressed. If so, have the reg serviced. And when you dive, make sure you have fully depressed the Shutter Crown with the knob, before you give up on screwing the knob in. It shouldn't be hard, and it shouldn't take more than "three fingers" force to screw it down, but make sure it moves.
3) Finally, do the old fashioned inhalation check when pressurized. It's hard with AI or electronic SPG's, but with a dial face you should not see ANY movement of the tank pressure needle during a vigorous inhale. Just as we demo'ed with a barely cracked tank valve, a barely open Shutter Valve will impede flow, and the SPG needle should drop perceptibly if the Shutter Valve is only barely open, with a fully open tank valve.

In summary, the ACD is a good theoretical design (though I would submit, an unnecessary trinket to add). While I would never characterize these incidents as "operator error", the implementation of this design, plus salt or sand or verdigris corrosion, plus tight tolerances, make the possibility that the Shutter Crown might hang up and not be noticed a real design flaw. Especially if, when you take your reg set into the LDS (after almost dying at depth due to a sudden gas shut-off), they say "Everything's tight!", and "We can't duplicate your problem", and (not said out loud) "Does this guy even know what he's doing? There's nothing wrong here."

There are my thoughts.
Whaddya think?

 

[Diving Dubai]

@rsingler

Great post.

I wouldn't' say I have strong feelings, for the ACD, although it's true I have (now) 4 ACD regs sets(I can't let a bargain go) I also have 3 other Apex sets without ACD. But as a mechanical engineer I don't like wishy washy theories and blame without proof.

I'll need to re read your post. but I think I get it (yoke is bad Din is good )

My one query (which might become apparent on re-read) is that the Shutter crown has a max movement of 0.040" /1mm (On DIN - I'm presuming Yoke is similar) so we're dealing with a very tiny deviation in the tightness of the yoke

Are we saying that the spring pressure of the Shutter crown is enough to form a gas seal? When its fully back it's a hard mechanical force, whereas as you hypothesize if the shutter crown is barely sealing, surely gas pressure could force it back, and because the yoke isn't tight you get a gas leak around the crown because there's not enough compression on the O ring to form a gas seal?

Not disputing your well thought out post, its just a quick question.

If I've misunderstood, sorry - it might become apparent after sleep

 

[rsingler]

@rsingler

Great post.

I wouldn't' say I have strong feelings, for the ACD, although it's true I have (now) 4 ACD regs sets(I can't let a bargain go) I also have 3 other Apex sets without ACD. But as a mechanical engineer I don't like wishy washy theories and blame without proof.

I'll need to re read your post. but I think I get it (yoke is bad Din is good )

My one query (which might become apparent on re-read) is that the Shutter crown has a max movement of 0.040" /1mm (On DIN - I'm presuming Yoke is similar) so we're dealing with a very tiny deviation in the tightness of the yoke

Are we saying that the spring pressure of the Shutter crown is enough to form a gas seal? When its fully back it's a hard mechanical force, whereas as you hypothesize if the shutter crown is barely sealing, surely gas pressure could force it back, and because the yoke isn't tight you get a gas leak around the crown because there's not enough compression on the O ring to form a gas seal?

Not disputing your well thought out post, its just a quick question.

If I've misunderstood, sorry - it might become apparent after sleep

@Diving Dubai , I wasn't picking on you. You've been very supportive of the ACD for your uses, and if it works for you (as it well might in your diving conditions), then I have no argument. It helps that you're a DIN diver.

Thank you for that total excursion data for the DIN ACD. It's my recollection that the yoke version moves a bit more, but I may be mistaken. In any case, you're right: the total opening movement is very small. That just goes to show how little tolerance there is for open vs. closed in that Shutter Valve.

As for the sealing force, I don't think the tank seal is more than minimally helped by the ACD spring. Looking at the diagram, although tank pressure is trying to "blow the Shuttle Crown off the tank" like a bottle cap, the tolerances are very small, and the Crown can't retract enough to allow extrusion of the tank o-ring with many tank valve designs. Additionally, if the Shuttle Crown "hangs up", the tank valve o-ring land is definitely deep enough to contain the o-ring as it jams itself into the corner formed by the face of the Shuttle Crown and the side of the valve land. If the Crown has indeed "hung up" with the Shuttle Valve barely open, then it's mechanically the same as though the Crown had fully retracted.
The part that's speculative is how the Crown might nestle deeper into the land for that final closing movement. It's not as though the Crown is like a balanced poppet. I'm guessing on that one, but if it can't be Shuttle Valve position, then it has to be some shift in the Crown position to close the opening in the Shuttle Valve. Conversely, a loose Shuttle Valve would improve opening, not worsen it. It might cause the ACD to not seal as designed, but not the opposite.
So if a first stage is loosely attached, without full depression of the Shuttle Crown into the Yoke Retainer, it won't initially leak if the side lands of the tank valve are high enough to contain the tank valve o-ring. The first stage is kept in position by the yoke, but can back away further if the first stage "rocks" back and forth during the dive. What's "floating" is the Crown, and if 1) the tank valve lands are high enough to contain the oring, and 2) the knob gets effectively "unscrewed" another 1/2mm by this rocking (which might be easier at low tank pressures than with a full tank), then the Shuttle Crown might creep up that final amount to almost completely close off the Shuttle Valve.
In fact, this might be the most likely cause, as I think about it. The problem is that it doesn't explain the DAN report cited above. In that case, there was supposedly a consistent failure below 900 psi tank pressure, despite switching tanks. If the Shuttle Crown isn't hanging up consistently in that case, I have no explanation. I just don't think the ACD spring is strong enough to work against all but the lowest tank pressures. Still...?

In any case, the prevention is the same: ensure that the Shuttle Crown is fully depressed into the Yoke Retainer when you attach your first stage to the tank.

 

[Diving Dubai]

@rsingler
Certainly your explanation has some plausibility. Given we have rejected other causes, and we know there was an incident there is a high probability that you could be onto something.

I've just dug out the Yoke parts from when I had mine converted - I was surprised to find them given that was 3 house moves ago.

On the Yoke variant the ACD is one complete "sealed" unit which the yoke nut is part of (see Pic)

I measured the shutter movement - it's the same at approx 1mm

I guess it might be possible for this part to become unscrewed (how much before you get a leak I don't know) But I would expect to feel that looseness in the Yoke.


In the Mid 80's - late 90's I was involved with Air Accident Investigations, my speciality being structures and identifying structural defects. Many weeks/months were spent trying to piece together all the evidence, disproving all sorts of wild a$$ theories put forward (in the media) and identifying which parts of the puzzle were red herrings

There will always be a chain of events leading to the final incident, if one link is missing the incident probably doesn't happen.

We don't know whether the cylinder used for the dive in question had been used before (on that rig) or it it were a new cylinder. It could just be a case of ill matched tolerances, damage or as you mentioned, the yoke not fully attached or located.

Having seen the filter (I don't rebuild my regs) - it is quite dense, can we be sure that it wasn't' a filter blockage (asking out of interest)

You've been very supportive of the ACD for your uses, and if it works for you (as it well might in your diving conditions), then I have no argument. It helps that you're a DIN diver.

I won't bore you with examples of why the ACD suits me (in addition of piece on mind and the ability to be lazy when washing my gear) I appreciate others not wanting such a device. That said I have no need for the filter on my 1st stage nor the debris tube in my cylinders, because my gas comes from an impeccable source and my cylinders are scrupulously 100% O2 clean - but I wouldn't dream of taking them out.

 

[Couv]

Whaddya think?

I think you are on to something. I don't have any regulators with an ACD nor have I ever seen one; so I'm going off of my sketchy understanding of it's operation. Here are a few additional "what ifs" to insert into the mix. The mix being a stuck/binding shuttle valve, loose yoke screw, and tank valve o-ring land design.

1. It's not unheard of that tank pressure gets behind the tank valve o-ring and exerts force on the back side of the seal. Combined with a slightly loosened yoke screw, could the pressure behind that seal push the ACD away from the valve causing a partial closure?

2. Note the seal in this picture with the red arrow pointing to it: (yes I stole the pic from the good doctor):

What if the seal or the sealing surfaces were temporarily compromised just long enough to trap pressure augmenting the spring? Again, combined with a loose yoke screw and/or a soft tank valve o-ring, the pressure would tend to force the ACD toward the closed position. Add a pressure differential due to lower supply pressure midway through the dive, and you've duplicated the issue the OP experienced.

 

[rsingler]

On the Yoke variant the ACD is one complete "sealed" unit which the yoke nut is part of (see Pic)

I measured the shutter movement - it's the same at approx 1mm

Thanks for checking this!
Any chance you can quantify the force necessary to start movement of the Crown into the Retainer?
Knowing this force and knowing the area of the Crown face, we could estimate the net "unsealing" forces provided by various tank pressures.