Benefits of redundant continuous pO2 monitoring?

[stuartv]

That's what I had originally thought, but I'm now struggling to see how having two O2 monitoring computers is beneficial, as if both computers are deriving their information from the same O2 cells, then surely they'll both just say the same thing?

Therefore under what circumstances is having a secondary O2 monitoring computer beneficial?

I think you now have the answer to this question. But, to put it succinctly, the benefit of a secondary O2 monitoring computer is for the case when the controller itself dies. E.g. the battery dies.

That said, I doin't know if that is a benefit on a Poseidon VII or not. Can you fly one of those manually, if the controller dies? I have a feeling not. I just spent a week diving with a buddy that was on one of those and listening to him describe all the different ways that it sucks. But, maybe you CAN fly it manually - in which case the secondary with O2 monitoring would be handy.

Also, earlier, you were talking about monitoring the O2 cells but not the ppO2. Those are really the same thing. If the Petrel EXT is readying the O2 sensors, then it knows (and will display) the ppO2 for each sensor.

The shearwater is intended to have 3 cells connected not 1. You can use one is SCR mode I think

The Shearwaters with Fischer cable will monitor 1, 2, or 3 cells. You do a calibrate on it and it will detect how many sensors you have connected to it and set itself accordingly.

This is actually a source of frustration for me and I have emailed Shearwater an enhancement request. I want my NERD to monitor 3 cells, but only use 1 slot on the Center Row of the display. I want it to display the single ppO2 value that it is using for deco calculations. I presume that it does some kind of averaging of the sensor values to come up with that single ppO2 value. I want to display just that, with some yellow or red color coding to alert me if a sensor value is out of range or has been voted out.

As it is, if it is calibrated to 3 sensors, it forces you to have the entire Center Row of the display devoted to showing the individual values. On my Monitor (versus the Controller), I would like to use a couple of the spots on the Center Row for other data.

 

[JohnnyC]

This is actually a source of frustration for me and I have emailed Shearwater an enhancement request. I want my NERD to monitor 3 cells, but only use 1 slot on the Center Row of the display. I want it to display the single ppO2 value that it is using for deco calculations. I presume that it does some kind of averaging of the sensor values to come up with that single ppO2 value. I want to display just that, with some yellow or red color coding to alert me if a sensor value is out of range or has been voted out.

As it is, if it is calibrated to 3 sensors, it forces you to have the entire Center Row of the display devoted to showing the individual values. On my Monitor (versus the Controller), I would like to use a couple of the spots on the Center Row for other data.

The problem with displaying only an average PO2 on the monitor is that in the event of a failure of the primary controller, you need to be able to do the voting logic in your own head. You can't do that unless you can see all 3 PO2 values. Tapping around to get to a 3-cell readout while dealing with a controller failure would be annoying as hell. The alternative is to bail out. Either way, you need easy access to all of the information. Simply telling you that a sensor has been voted out means nothing unless you can actually validate what the sensor is telling you, which requires you to monitor all 3 cells. Or just bail out and deal with it later.

Realistically, if we could ensure that the computer could detect a cell failure state with 100% accuracy, we wouldn't even need to monitor our PO2 at all, we'd just tell it the setpoint we wanted and ignore it. But unfortunately, considering the fallibility of galvanic O2 sensors and small-scale manufacturing, having access to only partial information is of little benefit, we really need access to all of the information.

Now Shearwater could easily display avg PO2 only in one column on the row and have a button press display all 3 cells, but that sort of breaks their interface scheme if they make it a primary operation, and of lesser value if it takes several presses to get to your full cell readout.

I don't disagree that it would be easier with one number, I just don't think they could make it a slick operation and still have it be functionally safe when a failure occurs.

 

[stuartv]

@JohnnyC I realize that all of this could just be my major inexperience talking...

I agree that they would need to add a way to quickly switch it to 3 cell monitoring. But, it seems to me that a controller failure would not be an emergency that would not tolerate 2 or 3 extra button presses. Again, maybe that's just my inexperience talking.

For a quick way to change, I would foresee left button presses to scroll just past the options for CC -> BO and Setpoint 1.3 -> 0.7. If you're in Avg ppO2 mode, you get an option that appears there, Avg ppO2 -> 3-cell. So, 2 or 3 left button presses and 1 right button press and now you have switched to seeing all 3 cells.

And, of course, this is all something that only is relevant if the user chooses in System Setup to display Avg ppO2 on the Center Row.

So, nobody has to use it unless they want to.

 

[NAND]

...adding to what JohnnyC already explained: There is another hidden benefit.

Knowing the exact sensor that is current limited while performing UW oxy-flush. I'd like to emphasize checking your cells during ending a dive as a regular procedure.

(And I know it sounds funny: You may like to use a permanant marker to identify the cell-connectors with numbers. No Idea why Paul never fixed any labels...)

 

[stuartv]

I do a sensor current-limiting test at the end of the last dive of each day (per my training). If my monitor had the option we're talking about, I would switch it to 3-cell mode at the time of doing that test.

And, yeah, I did use a Sharpie to number all my sensor connectors.

 

[JohnnyC]

For sure I don't disagree that it would be nice to be able to display a wider variety of info on the center row, although for me I like having all 3 cells. It's more of a "I want all the options, let me decide" sort of thing. And I don't disagree that just a couple button presses would be easy. The question is how that effects the rest of the menu structure and if it effects it enough that it screws with other things.

Ultimately it comes down to information management, what type of info you need and when you need it. Having the options is good. That being said, it's not something I would use. Doing cell validation takes some amount of focus, having to mess about trying to get to the right screen as well is just annoying to me, especially in shallow water. Less so in the middle of a dive doing a linearity check since it's much easier at depth.

 

[KevinNM]

The only potential issue is a full loss of the rebreather electronics. Most attempt to solve this by dual batteries running each controller/dive computer separately. It does still leave you with the issue of cell failure, hence the 3 cell voting logic approach.

I'm pretty sure people have had 2 cells fail on a dive, which votes the working, non-current limited cell out of the P02 system.

This approach has failed in other cases. For example the Airbus A320 uses 3 AoA sensors on the aircraft in a voting configuration. On 27 Nov 2008, 2 out 3 froze on flight 888T, and the airplane crashed into the Mediterranean Sea when it voted the still working sensor out of the flight control system.

It's hard to figure out a better non-absurdly expensive approach, but you can't just trust the 2 out of 3 system blindly.

 

[Gareth J]

I'm pretty sure people have had 2 cells fail on a dive, which votes the working, non-current limited cell out of the P02 system.

This approach has failed in other cases. For example the Airbus A320 uses 3 AoA sensors on the aircraft in a voting configuration. On 27 Nov 2008, 2 out 3 froze on flight 888T, and the airplane crashed into the Mediterranean Sea when it voted the still working sensor out of the flight control system.

It's hard to figure out a better non-absurdly expensive approach, but you can't just trust the 2 out of 3 system blindly.

Kevin

I agree there is the possibility, and it has happened that two cells voltage limit.
As you also correctly state, how much redundancy and double checking do you add.

I think there are two issues. One is what solutions can you implement in hardware/software (i.e. equipment based solutions). And dive practice or diver behaviour solutions.

Number one you are supposed to monitor your PO2.
If you have deviation between the readings you are supposed to take proactive action to confirm the issue. A loop flush, and a manual O2 injection give you linearity and a feel for the cells. (Many do an O2 flush prior to descent at 6m which confirms the cells are not voltage limiting at the start of the dive.)
You can turn down the PO2 setpoint, so even with voltage limiting cells, you should be able to move the operating window to a point that the cells can cope with. At worst you close the O2, and manually fly the unit on a single cell. This may involve closing off the O2 cylinder or just moving the setpoint to low and manually injecting O2 to bring the loop PO2 up.

From an equipment point of view, and I acknowledge that different units have different alarms.
On mine, deviation between the cells is only tolerated within a narrow window - so you get a warning that you have a potential cell issue.
There is a high / low alarm if you deviate to far from the setpoint.
There is a high PO2 alarm, if any of the cells reads too high then you get a high PO2 alarm.
So loosing two cells (and the voting logic working from the two bad cells), should mean that the third cell is still triggering the alarms.

Alarms and electronics are part of an argument which approach is best;

• an electronic unit - where you can become lazy, relying on the electronics to keep the setpoint steady and warn you of faults.
• or a manual unit that does not actively control the PO2, but gives you a PO2 reading, which you then react to, adjusting the PO2 manually, or realising that you have a cell issue.

These would help you manage two cells failing. If you lost all three then you are off the loop. Similarly, if you loose the PO2 readings on he handset or hud.

The advantage of he independent cell, plugged into a dive computer, is that it is feasible to fly the unit totally manually with no active operational controls on the rebreather. Sharing cells with your Shearwater (or other computer) however, does not solve the issue of failing cells.

Gareth

 

[NAND]

I'm pretty sure people have had 2 cells fail on a dive, which votes the working, non-current limited cell out of the P02 system.

Not quite. The two failing cells usually do not fail equally. (Meaning those cells would not display equal values) Means in such case all three cells would be voted incorrect.
In terms of the Shearwater SW it would simply tell 'voting logic failed' and continue working as fixed po2 divecomputer.

So either u have 2 independent po2 monitors in addition (eg. Revo exp) and stay on loop mCCR after DIL verification flush or you hail out.

 

[stuartv]

Not quite. The two failing cells usually do not fail equally. (Meaning those cells would not display equal values) Means in such case all three cells would be voted incorrect.
In terms of the Shearwater SW it would simply tell 'voting logic failed' and continue working as fixed po2 divecomputer.

So either u have 2 independent po2 monitors in addition (eg. Revo exp) and stay on loop mCCR after DIL verification flush or you hail out.

According to what I have read elsewhere here on SB, it HAS happened that someone died because they had 2 cells that both became current-limited (to the same limit) at the same time. Their computer voted out the one good cell and then kept pumping in O2, trying to get the ppO2 up, until they toxed.