Calibrating O2 for Altitude

[Jax]

According to this website, there is 84% less O2 at 5k ft than at sea level, even though the fractions would be the same.

So, to what value would calibrate their O2 analyzer for Altitude?

How much of a change in altitude would require a different value?

I've seen some more modern analyzers come with a table of values, but I don't have one. Does anyone have a formula?

 

[Jim Lapenta]

It does not require a different value. Dalton's Law is the prevailing one here. 21% is still 21%. The sum of the partial pressures of a gas remain the same as do those individual percentages. If you calibrate the analyzer with the surrounding air the percentages stay the same. All that changes is the NUMBER of molecules that make up that percentage. If you use the tank air to calibrate the analyzer the same still holds true. The altitude has no effect on the percentage of O2 in the mix. I analyze at my home around 500fsl and then at a mountain lake at over 3200fsl and the mix doesn't change.

 

[muddiver]

That is kind of what I was going to try and say.

 

[Jax]

It does not require a different value. Dalton's Law is the prevailing one here. 21% is still 21%. The sum of the partial pressures of a gas remain the same as do those individual percentages. If you calibrate the analyzer with the surrounding air the percentages stay the same. All that changes is the NUMBER of molecules that make up that percentage. If you use the tank air to calibrate the analyzer the same still holds true. The altitude has no effect on the percentage of O2 in the mix. I analyze at my home around 500fsl and then at a mountain lake at over 3200fsl and the mix doesn't change.

As I understand it, the analyzers do not actually measure the fraction of gas, but they measure the Partial Pressure of it. If the overall air pressure at 5000 ft is 84% of that at sea level, it leads one to believe the partial pressure is equally reduced, and therefore the user has to adjust for it.

 

[fdog]

Jax, the short answer is that you calibrate the analyzer ("zero" to 20.9%) under the same pressure conditions as used when analyzing.

As John rightly points out, the % doesn't vary.

Unfortunately, the oxygen analyzers do not detect percent. They detect ppO2. When you calibrate an analyzer, you are telling the machine "this ppO2 corresponds to 20.9%".

The fault is easily illustrated, at altitude, by calibrating an analyzer from a scuba tank of air to 20.9%; then, expose that analyzer to ambient air. The analyzer will promptly drop, for example, to 19.8% or so here in Reno (4500').

If you calibrate the analyzer by flowing gas, then, test the mix by flowing gas, all's good. The problem comes from calibrating to ambient air (not flowing gas), then flowing gas to test.


All the best, James

 

[Charlie99]

Fdog, the drop you see is also due to the effect of water vapor. Dry air from a tank has essentially no water vapor, while ambient air does. The water vapor pressure reduces both N2 and O2 partial pressures.

The effect you note about the sampling system presenting slightly higher pressure to the O2 sensor than when you just wave it around in the air will also be a problem at sea level.

 

[Charlie99]

As I understand it, the analyzers do not actually measure the fraction of gas, but they measure the Partial Pressure of it. If the overall air pressure at 5000 ft is 84% of that at sea level, it leads one to believe the partial pressure is equally reduced, and therefore the user has to adjust for it.

Normally your sampling system presents the tank air to the O2 sensor at very close to ambient pressure. That's what adjusts for the difference in altitude.

If you calibrate with a known tank of air, and use the same sampling system setup to test your nitrox, then the analyzer will read the correct % of O2.


An example at 18,000' or wherever the atmospheric pressure is 0.5 atm:
you calibrate with a known tank of air.
So the known air from the tank is presented to the O2 sensor at 0.5 atm. The ppO2 will be 0.209*.5 = .1045 atm.

Then you test a tank of EAN40. The ppO2 will be 0.5 atm pressure * 0.4 FO2 = 0.2 atm.
Your meter is calibrated to read 20.9% per 0.1045 atm.
0.2atm ppO2 will result in a reading of 0.2/0.1045 x 20.9% = 40%.

If your sampling system presents the tank air to the O2 sensor at slightly above ambient pressure, the calculations are still the same and you still get the right result (provided you keep the flow the same between calibration and test, and therefore the pressure is the same).

 

[Jax]

Normally your sampling system presents the tank air to the O2 sensor at very close to ambient pressure. That's what adjusts for the difference in altitude.

If you calibrate with a known tank of air, and use the same sampling system setup to test your nitrox, then the analyzer will read the correct % of O2.

Wow . . . I finally have a use for my 3cf?

It's been just too long since I've had the mechanics of liquids. . . .

 

[Jax]

I'm thinking that if PP_G = P_T * F_G,

and P_T@5000' = 84% * 1 ATA = .84, then

PP_O2 = .84 * .21 = .1764 . . . which - wow, algebra, .1764 = 84% of 21% . . .

Something is niggling in my skull about the mechanics of liquids that do not make that a linear equation.

 

[Charlie99]

Just remember, it's not the 1 or 2% errors in FO2 that kill you or bend you.

It's the gross mistakes such as the label being for a tank other than what it is attached to; or the analyzer calibrated to 21% with something that isn't really air.

My general rule of thumb is to do a precise cal using what I think is air from a tank, and then do a crude check using ambient room air.

The correction factor for 100% relative humidity and 86F is that FO2 is down to 20.0%. The pressure from sampling systems also makes the room air measurement read a bit low --- again a fraction of a percent.