Nitrox: How to calibrate/analyze

[Web Monkey]

Web Monkey wrote:
"The shop will typically have a calibration air cylinder next to the o2 checker just for this purpose."

. . . and who certifies that this cylinder is of correct mixture? Is it the responsibility for the LDS to supply a "calibration tank" every time a tank is provided?

I won't use a tank I didn't test, so from my perspective, that would be a "yes".

Terry

 

[The Kraken]

I guess my point is what agency is the controling agency for standardizations?
In the gauging world all gauges must be calibrated to standards that can be traced to the NIST.

If a gas mixer is checking his analyzer against the gasses that he mixes, then his gasses will always conform to specifications because he is checking the gas with his gauge. IE . . . if his analyzer is already 2% off and he checks the gas in a recently filled tank, then the tank will be accepted as correct but will be 2% off.
Who checks the "calibration tank", the NOAA, DOT ????? Who checks the accuracy of the LDS's gauge???

 

[Web Monkey]

I guess my point is what agency is the controling agency for standardizations?

I have no idea.

Terry

 

[The Kraken]

Man, that's kinda scary!!! I suppose one could purchase a bottle of medical grade oxygen from a supplier of such things and use that as a benchmark.

Oh, well . . . on to the next thing . . .

Safe dives,
safer ascents . . .
D

 

[MXGratefulDiver]

this method is flawed because changes in atmospheric pressure alter your readings.. At altitude you will get a VERY wrong reading....

Well now, after reading this I walked out to my garage and tried a little experiment. Here's how it went ...

I turned on my Oxycheq analyzer and calibrated it to ambient air - the display flickered between 20.9 and 21.0 for a few seconds then settled in at 20.9

Then I connected it (through the drysuit inflator on my own reg) to a cylinder filled to 3300 psi - the display initially went up to 21.5, then settled back down to 21.0.

Thanks for the advice, I understand the principle ... and the fact that the major agencies teach you to calibrate using an air cylinder ... but I'll go with empirical data.

... Bob (Grateful Diver)

 

[Rick Murchison]

Originally Posted by DA Aquamaster
This can be a problem if you don't have a known tank of 21% to calibrate with. However, at a given altitude with a given analyser the difference caused by the pressure difference inside the sensor is fairly constant. So if you calibrate with a known 21% sample at a given altitude and adjust the reading on the analyser to 20.9% and then, without changing the calibration, expose the sensor to open air for a few minutes, you will get a slightly lower reading. With my analyzer at 4000 ft I will normally get a reading of approx 20.5. So in a pinch, as long as the barometric pressure is the same, I can calibrate in open air and adjust the analyzer to 20.5 rather than 20.9 and expect to get a fairly accurate result. The particular difference in readings will vary depending on on barometirc pressure and the characteristics of your particular analyzer.

This is surprising.. at 4000ft your value if its at ambient should be somewhere around 17.9.. I'm guessing that the cell is not truely equalized to ambient... this can happen if the back side of the cell is sealed and there is a pressure imbalance between the front and rear of the cell..
I cant think of another way without having some stagnant other than having some stangant pressurized gas in the chamber..
But the error on the higher mix is close to expected..

Whew!
Lots of inferences leading to not-quite-accurate conclusions here...
(1) Oxygen analysers measure partial pressure rather than percentage of oxygen. Well, yeah, sorta, but... Since the analyser is measuring the oxidation of the sensor, the rate of oxidation is determined by the number of oxygen molecules that react with the sensor, which is pretty much dependent on the density of the oxygen and the flow rate, and yes, the density is a function of ambient pressure (altitude) and FO2. But what you're missing is that the sensor itself is always at ambient pressure; once past the flow restrictor, the molecules in the gas being analysed spread out to occupy the same space as the molecules in the air, whatever the altitude, and so a sensor that is accurately calibrated to the percentage of oxygen in the ambient air or the air in a scuba cylinder will accurately measure the percentage of oxygen in the mix being analysed. In other words, making some kind of altitude correction will give you bad readings, since all the gases are being analysed at the same pressure. The oxygen sensor is not in a pressurized chamber.
(2) In DA Aquamaster's post he notes a decrease in percentage when going from scuba air to atmospheric air of some 0.4%. This is not unusual at all, because the percentage of oxygen in dry air - the air in a scuba tank - and normally humidified room air, is about that, but can be as high as a whole percent (in Alabama in the summertime - "it ain't the heat, it's the humidity"). Using ambient air is fine so long as you (a) know the temperature and relative humidity, (b) have a chart to show you what your calibration value should be (varies from about 19.9% to 20.9%) and use that value, and (c) have a flow rate over the sensor for calibration that's the same as the flow rate you're going to use for analysis. I have found that gently waving the sensor back and forth does the trick (verified many times using a calibration gas and the same flow restrictor as for the gas being analysed).
Y'all be safe now, y'hear...
Rick

 

[padiscubapro]

Whew!
Lots of inferences leading to not-quite-accurate conclusions here...
(1) Oxygen analysers measure partial pressure rather than percentage of oxygen. Well, yeah, sorta, but... Since the analyser is measuring the oxidation of the sensor, the rate of oxidation is determined by the number of oxygen molecules that react with the sensor, which is pretty much dependent on the density of the oxygen and the flow rate, and yes, the density is a function of ambient pressure (altitude) and FO2. But what you're missing is that the sensor itself is always at ambient pressure; once past the flow restrictor, the molecules in the gas being analysed spread out to occupy the same space as the molecules in the air, whatever the altitude, and so a sensor that is accurately calibrated to the percentage of oxygen in the ambient air or the air in a scuba cylinder will accurately measure the percentage of oxygen in the mix being analysed. In other words, making some kind of altitude correction will give you bad readings, since all the gases are being analysed at the same pressure. The oxygen sensor is not in a pressurized chamber.

actually not entirely true..

it really depnds on the sampling chamber.. the sensor chamber is always slightly above ambient (otherwise there would be zero flow).. if the gas flows freely enough (and flow rate low enough) the pressure increase is small and the adjusted value will be much closer to actual near the calibration point, but the slope (of the output) is definately effected just a few tenths near the calibration point will generate a few % error at the top end.

I rather error on the side of caution.. The results will vary based on analyzer design (mostly chamber output), a 32% mix isn't going to be that far off from actual, but by the time you are checking an 80% mix, on most desings I would expect a few percent error.. The more restricted a chamber is or the higher the flow rate is the bigger the error will be..

if you keep a flow around 2lpm the pressure increase is very small and the error near the calibration point will be very small..

I did some tests a while ago ( for a ccr controller) to see how much of a pressure increase You could get by playing with flow rates in restricted spaces.. I was able to easily get changes of more than 3 psi (thats over 6k ft of elevation), it definately makes a difference..

 

[MikeFerrara]

Rick's right and in fact what he says is even consistant with what the owners manual for may analyzer says. LOL

As far as the relationship between pressure and flow...it depends on the geometry.

 

[novadiver]

actually not entirely true..

it really depnds on the sampling chamber.. the sensor chamber is always slightly above ambient (otherwise there would be zero flow).. if the gas flows freely enough (and flow rate low enough) the pressure increase is small and the adjusted value will be much closer to actual near the calibration point, but the slope (of the output) is definately effected just a few tenths near the calibration point will generate a few % error at the top end.

I rather error on the side of caution.. The results will vary based on analyzer design (mostly chamber output), a 32% mix isn't going to be that far off from actual, but by the time you are checking an 80% mix, on most desings I would expect a few percent error.. The more restricted a chamber is or the higher the flow rate is the bigger the error will be..

if you keep a flow around 2lpm the pressure increase is very small and the error near the calibration point will be very small..

I did some tests a while ago ( for a ccr controller) to see how much of a pressure increase You could get by playing with flow rates in restricted spaces.. I was able to easily get changes of more than 3 psi (thats over 6k ft of elevation), it definately makes a difference..

When calibrating an analizer for 80% mix, I alway thought using 100% to calibrate the analizer instead or 20.9% gave a safer and more reliable reading.

 

[Upwelling]

I heard the analyzers don't last long and have to be replaced--doesn't this turn into a huge expense over time?