Oxygen sensors — formulas to check linearity with an ambient air and an oxygen reading

[LFMarm]

This is cool QC, but also a pretty clear demonstration of what normal cells should look like? I thought they all (or most) curved off just like like this, rather than being completely linear?

Since we are diving in target ranges between 0.7 to 1.6 ppO2, and mostly in the 1.0 to 1.3 range, what might matter more than linearity to down to FO2 0.21 is what the mVs and curve look like in more relevant ranges.

Having valid, expected, *responsive* mVs at 0.21 & 1.00 (within spec) and approximate linearity still matters, but if we do get acceptable mV's at FO2 ~1.00, we are not going to be critically wrong at 0.7 or 1.3--even if the curve from FO2 0.21 is noticeably non-linear.

I wouldn't begin to worry unless the mV's were out of spec, or a sensor was old, inconsistent or slow. ?

According to Analytical Industries, the sensor should be within 2% linearity in the range pO2 = 0.21-2.0. Agree that for our use it only matters 0.5-1.3 but I would like to use this check to be able to tell when a sensor needs to be replaced because diverging from linearity too much.

 

[inquis]

More maths would probably show that even with a substantial curve-off from 0.21, the error from 1.0 to 1.3 is going to be acceptably small.

Much depends on where it stops being nonlinear and by how much.

For example, an assumed exponential deviation started at 0.8 bar (no error) with a 5% error at a PO2 of 1.0. Seem innocuous enough? Calibrating to that loop content results in displayed/actual PO2s of 1.0/1.0, 1.2/1.3, 1.4/1.7, 1.6/2.2 bar.

An even worse example is if the exponential deviation started at 1.0 bar (again, zero error there) with a mere 3% error at 1.1 bar. After calibration, displayed/actual PO2s are 1.0/1.0, 1.2/1.3, 1.4/1.8, 1.6/2.6 bar.

The first illustrates why it's a horrible idea to calibrate before every dive. Both illustrate why you should verify BEYOND your intended PO2 range. Yes, we run at least 3 sensors for good reasons, but history shows that's no panacea.

 

[LFMarm]

I would add that sensor behavior can change during the course of a long dive because of humidity. When I am doing dives with significant deco obligations (>60 mins) , I always do sensors checks on the way up from the bottom with dil flushes to make sure I am not ending up overshooting when pushing pO2 above 1.3 for accelerated deco.

 

[justinthedeeps]

Certainly if the curve becomes too flat, the y~x error could multiply dramatically in zones of 1.4+ atm. But it is rarely tested, except during less-than-clinical in-water ~6m O2 flushes.

I guess this is why Divesoft (and others?) offer the (up to) 3 atm sensor testing apparatus. Just do the experiment and see if 2.0 atm checks out? I saw it happening for 'ok'(?) sensors during COVID with the shortages. I'm not sure I would want to drive my sensors up to that level of chemical reactivity and millivolts though? Does it reduce sensor life? Another 'what if'