Diluent flush and current limited cells

[sheck]

Matt, Mark, These discussions are invaluable for people like me who are just getting into CCR as they tackle very serious aspects for diving our machines safely. I'm literally printing everything out to read it and compare it to my course notes.
There is so much you can learn in 7-8 days...
Thanks for all the good advice

Sheck

 

[sheck]

Does this look like reasonable and/or safe approach to you?

After all the pre-dive checks, of course.

Going down:
At 20feet/6m or slightly deeper, I check for current limited cells and see if all 3 cells read 1.6 or a bit higher.
Going down my set point is 0,7. Solenoid is run manual. The PP02 will increase, anyways, as I go down.
My ADV is on.

Once at the target depth:
I keep my set point at 0.7 and do a dil. flush to verify that my cells read the correct ppo2 at that depth.
I think I keep the set point at 0.7 and add 02 manually to maintain a 1.2-1.3 ppo2. Occasionally, I may spike it a bit to 1.4 to see that cells read higher that my 1.2-1.3 set point.
Also, once I reach the target depth and for the remaining of the dive, I shut off the ADV (flow stopper) and add dil. manually only if needed. I'm assuming that most of the loop volume that's gone (02 metabolism) is replaced with the 02 that I'm adding manually. Unless I have to clear my mask or something like that.

Going up:
My ADV is still shut off as I'll be injecting 02 on my way up to maintain a desirable ppo2. Set point is still at 0.7 (as a parachute) for the solenoid to kick in automatically if needed.
I get to 20feet/6m, after all the deep and deco stops, and flush the unit with pure 02 to finish my dive. It also allows me to check current limited cells again at the end of the dive.
I climb on the boat still on 02, if it was a demanding dive and go home.

Sheck

 

[Mark Chase]

Does this look like reasonable and/or safe approach to you?

After all the pre-dive checks, of course.

Going down:
At 20feet/6m or slightly deeper, I check for current limited cells and see if all 3 cells read 1.6 or a bit higher.
Going down my set point is 0,7. Solenoid is run manual. The PP02 will increase, anyways, as I go down.
My ADV is on.

Once at the target depth:
I keep my set point at 0.7 and do a dil. flush to verify that my cells read the correct ppo2 at that depth.
I think I keep the set point at 0.7 and add 02 manually to maintain a 1.2-1.3 ppo2. Occasionally, I may spike it a bit to 1.4 to see that cells read higher that my 1.2-1.3 set point.
Also, once I reach the target depth and for the remaining of the dive, I shut off the ADV (flow stopper) and add dil. manually only if needed. I'm assuming that most of the loop volume that's gone (02 metabolism) is replaced with the 02 that I'm adding manually. Unless I have to clear my mask or something like that.

Going up:
My ADV is still shut off as I'll be injecting 02 on my way up to maintain a desirable ppo2. Set point is still at 0.7 (as a parachute) for the solenoid to kick in automatically if needed.
I get to 20feet/6m, after all the deep and deco stops, and flush the unit with pure 02 to finish my dive. It also allows me to check current limited cells again at the end of the dive.
I climb on the boat still on 02, if it was a demanding dive and go home.

Sheck

Its prety much what i do

Ill never be on the surface with anything less than 100% 02 in my CCR so I use low set point and manualy 02 flush on the boat befroe I stand up, big squirt on the lift before I jump and more 02 if its anything like a swim to the down line.

Doing this I have no need to know my PP02 read my HUD or look at my display. All things that are often dificult in sunlight and too demanding when your focus is not missing the shot line.

SO I hit the shot and decend imediatly on 100% 02 decending to about 8m where ill ensure my PP02 is reading well above set point hopfulley 1.8 ish.

Then ill vent and flush dill do my bubble checks and decent sucking in dill on low set point all the way down.

Once on the bottom you will 99 times out of 100 be on pure dill and can quickly check your PP02.

Then ill switch to high set oint and watch the cells rise to 1.3. If stable ill go do the dive.

Ocasionaly ill check hand set and whak in a little 02 just to see the readings go above set point. Any fast or slow cell reaction will also be made apparent by this action.

Thats it foe me till I 02 flush at the ascent from my 9m stop and check cells can go above 1.6 again

But if when i check my handset or 02 inject during the dive i dont feel comfortable with what i see on my display? ill stop and do a low set point dill flush confirmation all over again.

Sounds like a lot of work but it realy isnt

 

[sheck]

Its prety much what i do

Ill never be on the surface with anything less than 100% 02 in my CCR so I use low set point and manualy 02 flush on the boat befroe I stand up, big squirt on the lift before I jump and more 02 if its anything like a swim to the down line.

Doing this I have no need to know my PP02 read my HUD or look at my display. All things that are often dificult in sunlight and too demanding when your focus is not missing the shot line.

SO I hit the shot and decend imediatly on 100% 02 decending to about 8m where ill ensure my PP02 is reading well above set point hopfulley 1.8 ish.

Then ill vent and flush dill do my bubble checks and decent sucking in dill on low set point all the way down.

Once on the bottom you will 99 times out of 100 be on pure dill and can quickly check your PP02.

Then ill switch to high set oint and watch the cells rise to 1.3. If stable ill go do the dive.

Ocasionaly ill check hand set and whak in a little 02 just to see the readings go above set point. Any fast or slow cell reaction will also be made apparent by this action.

Thats it foe me till I 02 flush at the ascent from my 9m stop and check cells can go above 1.6 again

But if when i check my handset or 02 inject during the dive i dont feel comfortable with what i see on my display? ill stop and do a low set point dill flush confirmation all over again.

Sounds like a lot of work but it realy isnt

It sounds like a safe way of doing it.
Thanks again!
Sheck

 

[uwxplorer]

This. And the reason that this is not always easy is that you need to be deeper than your low SP to do it properly. [i.e. shallower than 25m for an air-dil]. If you try doing it at high-SP it really doesn't tell you much at all.



Try this: https://cognitasresearch.files.word...the-inquest-into-the-death-of-philip-gray.pdf


Matt.

Thanks for providing a link to this report.
I find the attempt to account for the psychology of Evolution divers (and their frustration with cells, and in fact distrust of AP cells) quite fascinating.
Point 121 however sounds bizarre, since there should be no qualm using up a lot of diluent once at depth (unless they expected to be in for a rough ride in the shallows).

The last recommendation suggests to go one step further. Namely, whenever a situation occurs where:

- considering the depth and diluent and latest calibration
- if 2 cells read too low compared to the setpoint or the diluent's pO2
- AND the other cell reads high (way above setpoint and diluent's pO2)
- THEN voting logic should be abandoned.
The rationale is that this situation is likely to be one of 2 current limited cells, and one good one (of course if all 3 are bad, there is even less reason to rely on voting logic).

It would be interesting to hear what Shearwater (or other "voters") thinks about this suggestion (which probably isn't new).

 

[matt]

Thanks for providing a link to this report.
I find the attempt to account for the psychology of Evolution divers (and their frustration with cells, and in fact distrust of AP cells) quite fascinating.
Point 121 however sounds bizarre, since there should be no qualm using up a lot of diluent once at depth (unless they expected to be in for a rough ride in the shallows).

The last recommendation suggests to go one step further. Namely, whenever a situation occurs where:

- considering the depth and diluent and latest calibration
- if 2 cells read too low compared to the setpoint or the diluent's pO2
- AND the other cell reads high (way above setpoint and diluent's pO2)
- THEN voting logic should be abandoned.
The rationale is that this situation is likely to be one of 2 current limited cells, and one good one (of course if all 3 are bad, there is even less reason to rely on voting logic).

It would be interesting to hear what Shearwater (or other "voters") thinks about this suggestion (which probably isn't new).

121:I think the point is badly made. At 55m the PO2 could be 6.5. Bringing 6.5 down to <2.55 is going to take a fair amount of diluent. I doubt that Philip considered excessive use of diluent in his thinking process at all and find it unlikely that diluent supply played any part in the incident at all. That said, anyone who relies on a "quick flush" to check cells is on a hiding to nothing with PO2 excessively beyond 2.55.

Agree it would be useful to see some manufacturer views on the voting-logic. Abandoned: I think the unit could give a strong warning that the solenoid is open and that the PO2 does not seem to be rising on 2 cells [need not be cell, could be gas issue, but if 1 cell is rising...]. I'm not particularly in favour of automation, however, as it really is down to the diver to make the choices underwater. Quite a number of duff decisions made for this particular dive.

Matt.

 

[uwxplorer]

121:I think the point is badly made. At 55m the PO2 could be 6.5. Bringing 6.5 down to <2.55 is going to take a fair amount of diluent.

Matt.

I'd say it will take exactly as much diluent to flush the loop whether its pO2 is 1, 2, 10 or 100 ata to begin with...

 

[matt]

I'd say it will take exactly as much diluent to flush the loop whether its pO2 is 1, 2, 10 or 100 ata to begin with...

Agreed for a full-flush. I think that's the point that is being made.

 

[uwxplorer]

Is the Evolution equipped with a BOV plumbed into the diluent cylinder? What about the drysuit?
I would understand that if the BOV depends on the diluent cylinder's content, you may want to think twice before depleting that gas source by a full flush at 55 m. It could be argued that such a configuration is not adapted to tech diving though...

One other thing I was surprised by is the brief mention of Martin Parker being quoted as suggesting that the victim did not calibrate just prior to the fateful dive to avoid being prevented from diving by the computer (if calibration had failed). If correctly quoted, this is an example of a manufacturer confessing that too much computer control can have the exact opposite effect than what it is intended for (protecting the user)...

 

[matt]

Is the Evolution equipped with a BOV plumbed into the diluent cylinder? What about the drysuit?

I don't know where his BOV was connected. It's equally likely that it was connected to the Dil as it was to an off-board; both configurations are popular.

I would understand that if the BOV depends on the diluent cylinder's content, you may want to think twice before depleting that gas source by a full flush at 55 m. It could be argued that such a configuration is not adapted to tech diving though...

I think the point Fock is accepting (he's not putting it forwards) is that a full dil flush at 55m would use a lot of gas.

One other thing I was surprised by is the brief mention of Alan Parker being quoted as suggesting that the victim did not calibrate just prior to the fateful dive to avoid being prevented from diving by the computer (if calibration had failed). If correctly quoted, this is an example of a manufacturer confessing that too much computer control can have the exact opposite effect than what it is intended for (protecting the user)...

Martin said:

Another possibility - suggested by Mr Parker of AP Diving - is that Mr Gray deliberatelydid not calibrate at that time because he did not wish that to result in an unsuccessfulcalibration. Without positive evidence supporting it, I cannot accept that hypothesis. MrGray had attempted calibration on the evening of 7 February 2013. The unit calibratedsuccessfully and quickly. He had no reason to expect or fear that a further calibrationwould be unsuccessful.

This is true, if you suspect a calibration cannot be completed due to errors then you need to get the unit on and skip the calibration. Even a MUST can be ignored. You can even jump in with it turned off and it will come on with a START ERROR you can OK. If a diver wants to dive a unit then the diver will dive the unit.

Matt.

 

[uwxplorer]

So you CAN dive the unit in manual mode (with potentially no valid cell reading)?
That's interesting (and different from the Poseidon unit, as I recall, since it has no manual mode)... and scary if the inferred behavior is common among Evolution divers.

 

[Mark Chase]

On the volume of gas to flush issue

A: plumb in off board if required or use BOV to flush off board bailout gas of a known PP02. You have your BOV plumbed to off board gas right?

B: Not enough gas to flush and confirm cell function? Bail out and go home.

 

[Brad_Horn]

The last recommendation suggests to go one step further. Namely, whenever a situation occurs where:

- considering the depth and diluent and latest calibration
- if 2 cells read too low compared to the setpoint or the diluent's pO2
- AND the other cell reads high (way above setpoint and diluent's pO2)
- THEN voting logic should be abandoned.
The rationale is that this situation is likely to be one of 2 current limited cells, and one good one (of course if all 3 are bad, there is even less reason to rely on voting logic).

Quite a bit of research and development has gone into this problem, see http://www.deeplife.co.uk/or_files/Test_Data_for_Sensor_Fusion_Algorithms_RV_130325.pdf

Alternatively there is the 2007 Poseidon white paper https://zenodo.org/record/17661/files/MK6_White_Paper_v2.pdf covering their solution to it, through innovative use of just the 2 cells & now the additional option of a solid state sensor to cover the Hypoxic risk in more depth.

 

[uwxplorer]

None are "voters" though...

 

[rjack]

Which rises another question for me... My unit comes with a predator controller and a fisher cable to add something else. I'm still deciding what to do. a HUD? a EXT pretel or a satand alone petrel?
For the reason you described above, would I be better off by getting a standalone petrel with no cell info but just set point and correspondent deco info.

Thanks

Sheck

Setpoint doesn't actually affect your deco (esp. in the ocean) as much as you'd think. In this case it would be conservative anyway (2 cells low @1.1, one "true" @1.3, the shearwater votes out the 1.3 and calculates based on the 1.1) and you can just do some extra time. Considering you are running it manually with a bit more bouncing around of the fO2 this isn't really a big deal on modest ocean deco dives.

 

[Mark Chase]

Thanks for providing a link to this report.
I find the attempt to account for the psychology of Evolution divers (and their frustration with cells, and in fact distrust of AP cells) quite fascinating.
Point 121 however sounds bizarre, since there should be no qualm using up a lot of diluent once at depth (unless they expected to be in for a rough ride in the shallows).

The last recommendation suggests to go one step further. Namely, whenever a situation occurs where:

- considering the depth and diluent and latest calibration
- if 2 cells read too low compared to the setpoint or the diluent's pO2
- AND the other cell reads high (way above setpoint and diluent's pO2)
- THEN voting logic should be abandoned.
The rationale is that this situation is likely to be one of 2 current limited cells, and one good one (of course if all 3 are bad, there is even less reason to rely on voting logic).

It would be interesting to hear what Shearwater (or other "voters") thinks about this suggestion (which probably isn't new).

Quite so I have used ALL my dill at great depth due to unit failures and still finished normal deco of over 2 hours

On the wether or not to go manual issue?

KISS go manual on low set point every time

Its the safe option and avoids any possabuility of bad decisions made under stressful conditions.

As of course does bailing out but as we all know bailing out for CCR divers is sadly right up there with poking your self in the eye with a sharp stick

ATB

 

[sheck]

Hi Mark,

on a much older thread (2013) but related to my original question on this thread, you explained the following:

"
Ok i am at 60m and i am diving 16/50 diluient


I see my hand sets saying 1.3 2.2 1.3


The very first thing you have to do as per CCR MOD1 is dill flush or bailout.


So I switch to low set point and triple dill flush.


I am now expecting to see 1.12 1.12 1.12


If i see 1.12 2.2 1.12 I know exactly whats going on.


If i see 0.5 1.12 0.5 I know cell 2 is the only working cell


If i see 1.12 1.12 1.12 I know cells 1 and 3 are current limited and cell 2 is reading corectly.


Id leave the unit on low set point and end the dive imediatly.


Or if in a cave or other situation where this is not possable Id turn the dive on 0.7 set point and try and stay calm and relaxed to avoid the tox trigers.


I have no idea if i have the balls to do that, but id have a go because what other choice do you have.


ATB

Mark[/I][/I][/I] "

"If i see 1.12 2.2 1.12 I know exactly whats going on." it means cell #2 is way off and not working, right?

"If i see 0.5 1.12 0.5 I know cell 2 is the only working cell" This is obvious, yes.

"If i see 1.12 1.12 1.12 I know cells 1 and 3 are current limited and cell 2 is reading corectly." I'm having trouble understanding this one. You bring down all cells to the known ppo2 of 1.2 for the 60m 16/50 diluent. How do you get to the conclusion of cells 1 and 3 being current limited only with a dil flush? I mean, we know they all can read the 1.2. After the dil flush, wouldn't you need to spike it with 02 and see if cells 1 and 3 get stuck at 1.3 and 2 goes up to let's say 1.6?

Thank you

Sheck

 

[uwxplorer]

Yes, he probably would want to spike all cells with some O2, to verify his suspicion that 2 cells are current limited and that he was breathing a high pO2, as indicated (correctly or not) by the 3rd cell. The likelihood that 2 is the bogus one (for instance jumped to 2) is low.
But it is 1.12, not 1.2 (7×0.16). The 1.3 they read initially were most likely bogus (I mean in this hypothetical scenario).

 

[emmbee]

The last recommendation suggests to go one step further. Namely, whenever a situation occurs where:

- considering the depth and diluent and latest calibration
- if 2 cells read too low compared to the setpoint or the diluent's pO2
- AND the other cell reads high (way above setpoint and diluent's pO2)
- THEN voting logic should be abandoned.
The rationale is that this situation is likely to be one of 2 current limited cells, and one good one (of course if all 3 are bad, there is even less reason to rely on voting logic).

It would be interesting to hear what Shearwater (or other "voters") thinks about this suggestion (which probably isn't new).

Hi,

The problem of making sense of 3 sensors isn't new. It's the same in aviation, and the solution there is the same: voting logic (or use more sensors ;-) ).

I kinda like your idea, but there's a few problems I can think of:
- depth and dil, that's two new failure modes (faulty depth sensor, wrong dil) - let's ignore that and just use the setpoint.
- what you're proposing is basically to allow the highest reading sensor to veto the other two. That's extremely vulnerable to a single sensor reading high, possibly consistently high. Others have pointed out that the deeper you, the less ppO2 matters with regards to deco, that's true of course, but that's not true at all shallow. Then there's hypoxia. I'd be careful about unintended consequences.
- the devil is in the detail - what's "way above"?
- one virtue of voting is that it's simple to figure out what the computer is doing. This makes it much easier to spot and diagnose issues. Your idea might seem simple right now behind my computer, but I'm sure it'll be much less so in the water, never mind in the water with a problem to deal with.

IMHO, the problem with current limited cells begins with calling this a failure. It's not. Current limiting is what happens when a cell is running out of lead. Anyone calling an empty cylinder a failed cylinder? A car that ran out of gas a failed tank? No? Well then. Same thing. That matters for two reasons:
- it allows people to shift responsibility to the manufacturers: they should make better cells that last longer. They can't. Not electrochemical cells, that is. It's up to YOU, the user, to make sure they don't run out.
- people tend to categorise things in what's reliable and what isn't. Good cells and bad cells. Why not use the good cells, then? Because the "good" cells are the one that are going to get current limited. The "bad" ones died long before that. The "perfect" cell made of perfect materials will still become current limited.

Since we're proposing ideas, I got one too I'm only saying this because the coroner's report linked to 2 pages ago mentions that the two closest cells are averaged...

An alternative would be to simply take the median of all 3. It's just as simple to grasp (easier, actually). It changes nothing when 3 cells, or just 2, agree.

So what the difference?

Imagine two current limited cells, but not quite the same. They read, respectively, 1.22 and 1.15 at 1.3, 1.3 and 1.215 at 1.5, 1.38 and 1.22 at 1.7 (in case you're wondering I sort of picked that off Paul Raymaekers's paper on page 1 of this thread, I'm not completely making this up ). Assume the third cell is correct. The average system will inject until the ppO2 is 1.7: (1.38+1.22)/2=1.3. The median will only inject to 1.5: median of 1.5, 1.3 and 1.215 is 1.3. I know people dive 1.5. 1.7, I've never heard anyone say it's a good idea. Not a big help, but some help nonetheless.

Maybe I should add that median value is the default process for triple redundancy in aviation (but sometimes they average, too).

Any thoughts?


Cheers,

Matthieu

 

[Randy Thornton]

IMHO, the problem with current limited cells begins with calling this a failure. It's not. Current limiting is what happens when a cell is running out of lead. Anyone calling an empty cylinder a failed cylinder? A car that ran out of gas a failed tank? No? Well then. Same thing. That matters for two reasons:
- it allows people to shift responsibility to the manufacturers: they should make better cells that last longer. They can't. Not electrochemical cells, that is. It's up to YOU, the user, to make sure they don't run out.
- people tend to categorise things in what's reliable and what isn't. Good cells and bad cells. Why not use the good cells, then? Because the "good" cells are the one that are going to get current limited. The "bad" ones died long before that. The "perfect" cell made of with perfect materials will still become current limited.

Well said Matthieu! I agree 100%!