Oxygen sensor failure

[Storker]

Two possibilities. Cell is current limited at 0.21 or your knob-twiddle-range is limited to a multiplier less than that necessary to accommodate your cell

Given that I can twiddle the knob way beyond 21% with a fresh cell installed, I kinda doubt the latter

And, minor nitpick: I think the former should be "Cell is current limited below 0.21"

 

[doctormike]

Given that I can twiddle the knob way beyond 21% with a fresh cell installed, I kinda doubt the latter

Minor nitpick, but that doesn't follow. The knob gives you enough of a multiplier with a fresh cell to show far greater than .21. But the knob can't give you a large enough multiplier for that failing cell to show .21.

Unless the failing cell is completely dead (zero output, which it isn't), SOME multiplier will let you show 0.21 for that cell, that's just math. But the knob can't give you a large enough multiplier. So the failing cell is either current limited, generating a flat mV output lower than (0.21 / MaxMultiplier), no matter what the PO2, or it's generating a continually variable mV output over the usable range, but of such low output that (MaxMultiplier x mV in air) < .21. In the latter case, it may or may not be non-linear. But a linear response where the mV output is so low that the multiplier knob can't calibrate still means an unusable analyzer.

 

[Storker]

You know, I just became aware of another failure scenario which could be rather unpleasant to a simple rec OC diver: the cell is current limited below 0.21, but the multiplier circuit can bring the voltage up to showing 0.21. But then the output would be flat from below 0.21 and beyond my max fO2, and my tank of properly mixed EAN36 would be identified as a tank of air. Could be nasty if I'm planning to go to 40m...

Fortunately, most of the unpleasant failure modes usually will be detected by using two different analyzers. And if they don't agree within the expected variance of ~1 percentage point, don't breathe the tank until it's been analyzed with a third unit which agrees with one of the others.

 

[doctormike]

You know, I just became aware of another failure scenario which could be rather unpleasant to a simple rec OC diver: the cell is current limited below 0.21, but the multiplier circuit can bring the voltage up to showing 0.21. But then the output would be flat from below 0.21 and beyond my max fO2...

Right, which gets back to my original question. Is that a thing? Do we ever see in actual use a cell that is current limited at a PO2 below 0.36 ? I don't know the answer to that question.

A recreational analyzer is simple by design, but there is something nice about being able to see actual mV outputs for each of your cells and to be able to track them over time. But my suspicion is that this is less of an issue in the < 0.36 range.

 

[northernone]

Right, which gets back to my original question. Is that a thing? Do we ever see in actual use a cell that is current limited at a PO2 below 0.36 ? I don't know the answer to that question.

I have one currently on my workbench that reads 16% to 25% accurately and afterwards drifts off to a current limiting maximum of 30%. Been that way a month now.

As compared against 2 other cells in the same gas bag (rebreather loop)

Cameron

 

[Storker]

I have one currently on my workbench that reads 16% to 25% accurately and afterwards drifts off to a current limiting maximum of 30%. Been that way a month now.

Thanks for that. To me, this means that the rec diver relevant failure mode which worries me the most is a real risk.

 

[doctormike]

I have one currently on my workbench that reads 16% to 25% accurately and afterwards drifts off to a current limiting maximum of 30%. Been that way a month now.

As compared against 2 other cells in the same gas bag (rebreather loop)

Cameron

That answers my question. Thanks!

 

[boulderjohn]

I am no scientist, and much of this discussion is beyond me, but I do have some experiences with aging cells that might be interesting.

I have an Analox ATA Pro trimix analyzer that was having problems calibrating and was way off at the far end of the scale with both the O2 and the helium sensors. Analox told me that if I was getting readings from both, I probably just needed to do a recalibration within the sensor itself, a process that involves a screwdriver. I opened up the box and decided that as long as I was in there, perhaps I should put in a fresh battery. Lo and behold, both sensors immediately calibrated perfectly, and they had no trouble measuring accurately throughout the full range.

A few weeks ago, I was with a group of divers, and we analyzed all our tanks in the morning before diving. The temperatures each morning were in the 30s F. I had a lot of trouble with both cells, especially the O2, the same ones mentioned in the paragraph above, that incident having occurred a bit more than a year before. Another diver had similar problems with his different brand analyzer, although his was more with the He sensor. We ended up using both analyzers on our tanks. A few days later I used my analyzer again, this time with temperatures in the mid 50s F. They both worked fine, although I had to turn the dial for both cells quite a bit for initial calibration.

In talking with a tech at the shop form which I planned to buy replacement cells, I learned that the O2 cell should last 3-5 years, and the He cell should last 10 years. The O2 cell was at its life expectancy, but the He was not. I replaced the battery again (it was only a year old) and, Bingo, the He cell was back to perfect, but the O2 cell was still calibrating at the far end of the dial. I have ordered a new O2 cell.

 

[northernone]

There are sensor troubles and also analyzer component troubles.

For example a corroding wiring harness can increase resistance and give a faulty or inconsistent reading on the display even if the oxygen powered electrical generator is in perfect condition. The same goes for the rest of the device.

Basic design is:
O2 cell - potentiometer (sanddollar stunt double for the photo) - resistor - volt meter

I can twiddle a dial on the potentiometer and get a nearly burnt out o2 cell to read output voltage (o2% on the display) I want.

Provided it is outputting a slight current (milivolts) I can run through the trimpot and connect to a volt meter.

What worries me is the failures are varied. Reads too high, too low, current limited, inconsistent, non linear failure etc. For example I had a cell that was matching the other 2 between 11% and 100% but when I got to depth the others read "120%" while it read "160%".

Could theoretically do the same behaviour between 21% and 26% but display 32 % as 38%.

Which I think it might bite someone eventually if there's only a single point calibration on a single analyzer between the initial fill and the end user. If the shop blends and analyzes followed by you analyzing with the same defective analyzer, the tool will fail us twice and we're no safer having checked twice because there's nothing to check the measuring rod against.

Cameron

 

[rjack321]

You guys are forgetting that cells have an established mV starting point, for most analyzers its 9-13 mV. You can't just start with "any" voltage reading and somehow play with the dial enough to get it to calibrate. Sure mathematically you could put a 1mV cell in there are turn up the potentiometer to the max and get it to read 0.21 and the cell wouldn't be capable of reading past 0.25 and that would be "dangerous". But the reality is that doesn't happen because while a 8mV in air cell could calibrate to 0.21 and read (perhaps) up to 0.95 once that cell falls to 5 or 6 mV in air its not going to read 100% at all. Instead of thinking your deco mix is off you suspect the cell, in part because you cranked up the potentiometer just to get it to read 0.21. By the time a cell won't read 0.32 or 0.36 its so far gone that you can't even get it close to 0.21 at the start, so even with a "one point" calibration you notice the issue. The reality is that OC divers aren't being killed by failed analyzers, its when they don't use an analyzer at all.