Is the Mk VI / SE7EN really that dangerous

[Gareth J]

I have read the thread as it has developed with interest. It is nice to learn a little more about the unit from those who are instructing on the unit, and to understanding the changes over the last few years since it first came on to the market.

rainpilot, I apologise if you felt my post was inaccurate. I did try to be above board, by stating I had minimal knowledge.
Looking at the general feedback, I still feel I was reasonably accurate. I certainly stated that any statement that the Poseidon unit was dangerous was unfounded (at least no more so than any other CCR, possibly less so).

taimen, figures for units sold are probably classified as commercially sensitive.
Certainly in the UK, the AP units have the highest volume of sales, which is reinforced by the number of units on the secondhand market. AP, did have the advantage of being one of the first selling units into the recreational markets. KISS probably being the second notable supplier in the UK. The two units having marked differences, one being eccr and the other mccr. There are now so many suppliers that without research, it has become more confusing.

I am well aware that there are a number on here with significantly more experience with different units. My experience is limited to my AP unit, the Sentinel, which a couple of my friends dive, and the KISS which one of my older friends dived until he stopped diving. I do like the look of the JJ, if I had money to burn, I would possibly look at them as an alternate to the AP unit I currently use. I am also aware that the Sentinel has undergone some major modification since VMS took over the support, I think they where very brave to take on the Sentinel.
Apart from the reputation of the manufacturer, I think the second most important factor is local technical support. The picture in the USA is probably very different to the picture in the UK/Europe.
One of the big advantages for me when I first got the AP unit was that in theory, there where others using the same equipment making support, assistance and advice readily available locally. However, as it has turned out, in my group, I have for the most part the only person on an AP CCR, the one other user, has moved back to OC, due to a lack of diving and retirement.

 

[rjack321]

The picture in the USA is probably very different to the picture in the UK/Europe.

AP units have to some extent fallen out of favor locally (Seattle area) due to a variety of factors - support and an active local dealer is a one. They were more common 10yrs ago, but there were also fewer units on the market back then. APs are still around, just less common.
Nowadays you are most likely to see Revos, Megs, and JJs (esp amongst the GUE crowd). A few older kiss units still in circulation - since they were formerly located just to our north in British Columbia. A few Sf2s (<10 I would estimate). Not many X or defenders at all (despite subgravity being "close by" in Utah). Poseidon, Optima, Hammerhead are also rare or non-existent. I have never seen a hollis, borus or mk15 here - although I am sure some exist.

Basically what you might expect based on what training is available, dealerships, and parts.

 

[uwxplorer]

It only needs to go a bit above the setpoint to demonstrate continued linearity. 1.8+ppO2 is enough to illustrate that the cell is not compromised at ~1.3

I see. So it's not really a continuous calibration (as I thought I had understood), except for the range of pO2 where it is (which is kind of unpredictable)? Or specifically up to the setpoint?
I can't see how this guarantees that you don't have two current limited cells at great depth, say using 10/50 at 80 m (9 ata). Your diluent puff will get cell 2 to give an acceptable reading, cell 1 reads high but the exact value can't be relied on, and when everything is mixed, you have close to your setpoint on both cells, except that the actual pO2 is much higher...
I'd rather have a few more O2 cells to feed my brain cells with contradictory info...

 

[RainPilot]

All we care about is that it is not current limited below 1.6 , all cells are current limited somewhere. As long as the cell maxes out at, say, 3.0, and at 60m it still reads 3.0, that cell is fine wrt linearity. The dil will then be an overriding calibration. So at 80m on your 10/50, the dil cell should read 0.9 ie xx mV, the O2 cell reads the same mV (given the starting differences between them when first calibrated) and so the dil cell plus the theoretical cell will verify the correctness of the O2 cell and the ability of the cell to read high PO2 is validated as well.

I have an issue with too many cells, I know for a period that seemed to be the way everyone was going.

Lets say I have a conventional voting logic system with 3 cells, A, B, and C.

A and B fail in the same way (same cell batch, both get wet, potting issue, short circuit, whatever)

Now, with 3 cells, the good cell C will be voted out and may lead to a Hypo or Hyper scenario.

So, what's the answer? Add another cell! Ok, problem. AB read one value, CD read another, still no way to tell which ones have failed.

Add another cell to break the tie? Do some complex-ish mathematics (for me, most underwater maths is complex) regarding previous mV readings? Do a full dil flush and check the readings against expected? These are all strategies that work, some better than others, and there is NOTHING wrong with them.

The OTHER way of skinning the cat is to remove the need for those, by continually doing the checks that you would do if you suspected a bad cell, and reporting back if there's an issue.

Yes, it requires trust in the system but we are no longer in the days of dive computers failing all the time and a careful design avoids most any undetected fault. I am ok with a system failing, as long as it tells me its failing in some way.
(Thats why I am keen on SS O2 cells, since they either give good info or no info, the failure modes are not nearly as insidious as the galvanic.)

The Poseidon has 3 redundant computers. One is inside the head, one is inside the battery and another inside the paddle/M28. They all read the data and do the calculations, then compare. I would MUCH rather trust voting logic on software than physical sensors.

Now, I know the plural of anecdote is not data, but Richard Pyle and Brian Greene and co have been doing HUNDREDS of dives to the 500' mark , almost routinely, for YEARS now on the MK6, then the Se7en. Those are a lot of dives to do without any serious incident that I am aware of.

I imagine that the Meg, the JJ and maybe the AP(?) are being used regularly for these sorts of dives, in these sorts of numbers, I think the accident stats speak to the general safety of CCR even on these huge dives, and that the Poseidon is not, in any way, inherently unsafer than any other unit. A different mindset, different engineering choices, different paradigm, sure. Unsafer? Not the machine.

 

[taimen]

There has been almost complete silence regarding the solid state cell for past year or so.
Are they officially released for sale yet and are new units shipped with them?

 

[RainPilot]

There has been almost complete silence regarding the solid state cell for past year or so.
Are they officially released for sale yet and are new units shipped with them?

SSS is available to buy with a 3rd cell fitting. It is digital so the CPOD is required to power the SSS
CPOD incl. SSO2 and Cable

At TekDive or DEMA, a few weeks ago, they had a Se7en with SSS installed. They are not selling separate sensors yet, but there are quite a few divers using them out in the wild.

 

[JohnnyC]

The problem with the argument against 3 cell voting logic is that the assumption is being made that the diver implicitly trusts the electronics and makes no checks to validate the information being generated by the O2 cells (essentially diving the way Poseidon wants you to dive the unit). That's simply not the case. Yes, 2 bad cells throws voting logic out the window, but that ONLY means that the voting logic will incorrectly control the solenoid if everything else is left to their own devices and the diver does not intervene AT ALL. That is not the way divers are taught to dive the rebreather, and every MOD1 course I've taken takes explicit steps to outline the fallibility of 3 cell voting logic and the way to both validate cell readings, as well as how to dive the unit in the event of a failure of voting logic. Do some divers ignore this? I'm sure they do. There are also divers that dive 2 year old cells and they end up dead. Those divers will be just as dead regardless of whether or not they're using 3 cell voting logic or Poseidon's validation scheme. Those divers don't dive with bailout, dive to the limits of their scrubbers duration to save a couple sheckels, make "visual jumps" in overhead environments, etc. They are walking accident slopes regardless of the number of safeguards they have in place.

Divers running mCCR's with 3 cells are arguably even safer in that there is no voting logic that could possibly make a decision to inadvertently change the PO2 to an unsafe level based off bad information. None of the information is hard to deal with, and for those, a sticker with dil values and mV outputs at different depths is easy enough to reference. "Close enough" is close enough. We're not doing rocket surgery, we just want to make sure that at 10m dil is close to .42 and O2 is close to 2.0.

Two ways to skin a cat, but I would argue that 3 cell validation when used properly is safer if immediate bailout is not your single solution to a failure. Ideally they'd make a SSS with a battery and DAC that allows you to put it in any rebreather that currently takes galvanic O2 cells. But Poseidon is like HK, you suck and they hate you.

The next part drifts off topic but I think it needs to be pointed out, and is unit agnostic. If it pisses anyone off, sorry, I'm not directing this towards anyone in particular other than using the example of Richard Pyle and Brian Greene.

Here's the problem with using Richard Pyle and Brian Greene as examples. They are not GOOD rebreather divers, they are LUCKY rebreather divers. The Bishop Museum website is a great example of this. (I know it's a 20-year old website, but I'll address this later) The "Lessons Learned" is a substantial list of times when the good doctor should be dead and drifting on the sea bed. He is lucky, hands down, and while I'd rather be lucky than good any day of the week, I'd rather be good and not have to rely on luck at all. Now, to address the argument that all of this is old news.... They regularly dive in open water to 100m+ depths (easy to verify, I'm watching a video right now where they hit 150m in 2016, this is common for them) with a single AL80 of bailout butt-mounted, with other bailout staged elsewhere up the wall at shallower depths. This is open water, at depths in excess of 100m. The math clearly doesn't work out in their favor, even in a benign bailout scenario, let alone something like a CO2 hit where their SURFACE consumption could peg 65lpm. Now extrapolate that to 100m (let alone 150m) and that single AL80 lasts 4 minutes. 4 minutes to make a safe ascent to their staged bailout. Again, LUCKY, not GOOD. People have died because they did not have access to staged gas. Quite a few people in fact. 4 minutes goes fast when things are breaking. It's not a hardship to hedge your bets, carry enough gas.

If anything, they're examples of what NOT to do. And it has nothing to do with the machine. Any modern rebreather, as long as its functioning, will make those dives day-in and day-out without issue. It's fantastic that they have faith in the unit (as I do in mine), it's definitely a testament to the reliability of the rebreather (or any rebreather) when it works. The issue is when it doesn't work at 150m, again a potential with any modern rebreather. Failure on the boat is potentially a lost day of diving, failure at 150m is another animal and evidence points to them being ill-prepared for it.

To look at the experience argument another way, look at BASE jumping/wingsuit accidents. Jeb Corliss is the face of wingsuit flying (literally thousands of BASE jumps) and he damn near torched himself jumping off Table Mountain in South Africa despite being a base jumper for over 10 years. Dwain Weston killed himself in Colorado wingsuiting alongside Corliss. Over 1200 base jumps without so much as a broken bone, hit a bridge, severed one of his legs, and bled out on the rocks below after his parachute opened due to the force of the collision. A 2012 study by a doctor at the University of Colorado indicated that 72% of wingsuit fliers had witnesses a death or serious injury, 76% had experienced a "near miss" with about 24 fatalities a year. "The average active career of a proximity flier is 6 years, after which it is generally curtailed either by death, injury, or prudence." (Oxford comma because come at me bro) These are HIGHLY experienced people, and they're still dying. While there are a plethora of reasons, the commonality behind all of them is that the activity, whether deep rebreather/cave/whatever diving, BASE jumping, wingsuit flying, has a much higher risk profile, and those with experience are more prone to accept a normalization of deviance while undertaking those activities, without scaling the failure response to the activity. CFIT under VFR happens ALL THE TIME amongst experienced pilots because they've done that milk run to Denali a thousand times and know that it's a piece of cake. Until they park the tail of their Piper Super Cub through the cockpit into the side of a mountain that hasn't moved in 10 million years, give or take a couple million. (Seriously, Alaska is just the worst. And not just because it's cold and the animals want to eat you.) The guys in the Plurdalen that subsequently became the unfortunate stars of "Diving Into The Unknown" had a boatload of hours doing deep overhead penetration dives.

The point is, it's not a matter of the plural of anecdote not being data, it's that we shouldn't even be using them as anecdotes in the first place, at least not to prop up the infallibility of a piece of hardware. FWIW I don't do "Pyle" stops either. Because he's a fish scientist, not a decompression doctor, and the proof of the pudding is in the eating.

 

[uwxplorer]

All we care about is that it is not current limited below 1.6 , all cells are current limited somewhere. As long as the cell maxes out at, say, 3.0, and at 60m it still reads 3.0, that cell is fine wrt linearity. The dil will then be an overriding calibration. So at 80m on your 10/50, the dil cell should read 0.9 ie xx mV, the O2 cell reads the same mV (given the starting differences between them when first calibrated) and so the dil cell plus the theoretical cell will verify the correctness of the O2 cell and the ability of the cell to read high PO2 is validated as well.

I am sticking to pO2 to simplify, using the nominal conversion factor of each cell.
Dil puff will never check the linearity of that cell above 0.9. So when it will read 1.3 (or whatever the Poseidon setpoint is), I have no idea whether it is 1.3, 1.6 or anything else. I have to trust it is not limited.
Now, I am not sure whether there is any comparison made between the two cells (2 cell voting "logic"?), but if the O2 side cell reads something that is not 9 ata after the O2 puff (which is most likely the case), but is above 1.6, what do I learn? Nothing really. When the puffs are mixed up in the loop, both cells read 1.3: everything might be OK, but then things could also be terribly wrong, as I don't know whether my first cell is limited, and neither do I the second one...
I get that Poseidon's approach could be fine for recreational diving (<5 ata). Maybe. It would be easy to add a 3rd cell for tech diving (if I recall correctly that is an option?) and add an umpteenth error code, when classic voting logic fails. As stated by JohnnyC above, this doesn't mean that the controller can figure out the likely status of each cell, but it is an invitation for the diver to diluent flush the loop (only if it safe) to make a sanity check. Or bailout, in doubt.
Don't get me wrong, I was psyched by the marketing when I knew nothing about rebreather diving...which in Pylean fashion, I still believe is the case after 5 years diving a rEvo.

I am also puzzled by the response time of Poseidon cells. A <1 s puff of O2 or diluent at recreational depth will diffuse away from the cell pretty fast (not mentioning flow, due to both injection and breathing). How you can get a reliable reading is kind of counterintuitive for a sensor guy... but again, I am just wondering, not criticizing or anything.

 

[JohnnyC]

Response time is an interesting topic! Of course slow cell response time, even if ultimately accurate for PO2 is a potential indicator of an issue.

The validation scheme in the stock 2 cell Pelagian is to exhale loop volume, inhale sharply to trigger the ADV blowing any moisture off the face of the cell. More volume is exhaled and then a deep inhale that triggers the ADV will blow dil across the face of the cells, albeit slower and in larger volume than a single puff. The electronics on the unit are designed to be as fast as possible, showing immediate changes in cell output. Apparently this is quite a bit faster than other rebreather electronics, but allows dil validation with very minimal gas required.

Depending on the response time of the Poseidon, I could see it going either way. If a large volume of dil is required to validate, that would play hell with buoyancy and loop volume and PO2. I would lean to it being pretty responsive though. I don’t anyone would design a rebreather that’s purposely a pain to dive.....

It seems as though you could never truly isolate your dil without having to shut down the valve, as the unit will continuously be utilizing it for validation. I’m sure it’s not a problem in terms of volume, but in the event of having to shut down the dil side for some reason, you lose the ability to validate the cells accurately, and you’re trusting the single cell again. You’re back to one real cell and the “theoretical cell” in the brain. That doesn’t mean the other cell isn’t providing accurate PO2 data, just that you’ve lost the ability to validate cell health. I think that’s what that means, again, not a Poseidon expert, just my interpretation of their documentation and the explanations here.

 

[Superlyte27]

So, what's the answer? Add another cell! Ok, problem. AB read one value, CD read another, still no way to tell which ones have failed.

That's not accurate.