Comfort Zone - What range of Gas Analysis?

[The Chairman]

Then I would guess you have a much wider swing in what your analyzer reports before you question its output?

You're missing the point. Once you calibrated your tank, there was no 'swing'. Your technique, not the analyser was flawed. It doesn't matter if you calibrate off of air or an air tank: calibrate. If it calibrates and is stable, there is no worry about a 'swing'. However, there are many other factors, including hydration, fatigue, how close you push the tables, ascent rates, respiration rates, work rates and the list goes on, that will affect your chances of getting hurt.

One of the best carpenters i have ever known had the most rickety table saw I've ever seen. However, he knew his equipment's limits and prduced amazing work. It's not that he eliminated the error (slop) in his saw. Rather, he designed and worked in a way that ameliorated the effect of the slop. Tables and PDSs are simply tools. You should know how and when they will fail you and dive accordingly.

 

[Jax]

Actually, my point is not the analysis, it is, "At what point do you question your analyzer?"

 

[Storker]

"At what point do you question your analyzer?"

When the two-point calibration fails to read within the instrument's margin of error at both points, after several calibration adjustments.

The other thing you could try is to do a 2 points calibration: basically you calibrate the analyzer with 2 known gasses (like air and O2). In this way the Analyzer can build a more precise reference line (from point A to point B)

A one-point calibration isn't considered sufficient in science unless your analysis value is quite close to the calibration value. If you want to measure at some distance from your single-point calibration value, the potential error is too big to trust the instrument's numbers. That's when you need two-point calibration.

 

[The Chairman]

Pergamentum init, exit pergamentum! You claim to be an engineer so look at the failed process: You did not calibrate. Before you question your analyser, you should question your technique and your understanding.

 

[Akimbo]

I don't know what analyzer you are using or how good the manual is but this might give you some insights:
http://www.teledyne-ai.com/manuals/man_320series.pdf

 

[Storker]

The problem - from an analyst's point of view - with the O2 analyzers I've looked at, is that they only have one calibration adjustment. I guess that's the offset, not the slope. For a proper two-point calibration, you need to be able to adjust both offset and slope. That way, you know you can trust your analyzer between the two calibration points (provided the response is linear). If you can adjust only the offset, the two-point calibration only serves to confirm (or refute) the assumption that your instrument's slope is correct.

If your two-point calibration doesn't work, you need to either adjust the slope, or change the sensor. Alternatively, you can calibrate close to your measuring value. That way, a slope error will be negligible.

 

[Akimbo]

Barring unusual problems with the sensor, galvanic fuel cells are repeatable enough for what divers use them for. Ideally, you want to calibrate against a known source that is in the ball-park of the gas you are testing. Fortunately air is close enough for Nitrox divers.

We purchased industrial cylinders of cal gas (about 230 Ft³) on saturation diving systems because the deep mix was HeO2 usually in the 1-6% range. The sensors did surprising well even when comparing the readings at sea level to sensors in the chambers. For example, the sea level reading would be 2% and the inside instrument would read 30% (0.3 ATA really). That is a pretty wide spread.

 

[Jax]

I don't know what analyzer you are using or how good the manual is but this might give you some insights:
http://www.teledyne-ai.com/manuals/man_320series.pdf

This why I am asking about your comfort zone -- not about the technologies.

I know that my VTI analyzer is +/- .5% equipment, and the sensor is a +/- 2%.


I am wondering when you - diver - raise your eyebrows at the output, and question the equipment.

I have used mine enough, and second checked with the store's result, that I believe my VTI is PD accurate.

When the sensor fails, the reading doesn't stabilize.

When the battery fails, there is a warning light to replace it - but the equipment doesn't act wonky.


When my VTI says my reading is off the shop's reading by more than 1%, I'll ask the store to show me their calibration. The VTI has always been right.

When it is off by .5%, I'll double check everything and analyze the risk to my dive (normally negligible).


I have always been taught, it doesn't matter how precise the measurement, how many calculations you have the computer run, it is YOUR project and YOUR responsibility for the final outcome. Learn to recognize "it doesn't quite fit" and find out why before you continue. That is why a professional engineer costs so much.

For some reason, I thought others may have an "eyebrow lift" range at which point they question the values the device gives them.

. . . So, why don't you?

 

[sea_ledford]

This why I am asking about your comfort zone -- not about the technologies.

I know that my VTI analyzer is +/- .5% equipment, and the sensor is a +/- 2%.


I am wondering when you - diver - raise your eyebrows at the output, and question the equipment.

I have used mine enough, and second checked with the store's result, that I believe my VTI is PD accurate.

When the sensor fails, the reading doesn't stabilize.

When the battery fails, there is a warning light to replace it - but the equipment doesn't act wonky.


When my VTI says my reading is off the shop's reading by more than 1%, I'll ask the store to show me their calibration. The VTI has always been right.

When it is off by .5%, I'll double check everything and analyze the risk to my dive (normally negligible).


I have always been taught, it doesn't matter how precise the measurement, how many calculations you have the computer run, it is YOUR project and YOUR responsibility for the final outcome. Learn to recognize "it doesn't quite fit" and find out why before you continue. That is why a professional engineer costs so much.

For some reason, I thought others may have an "eyebrow lift" range at which point they question the values the device gives them.

. . . So, why don't you?

There are two things that cause me to question the analysis: first, and most common is that I keep having to reset the value during calibration. I know that under ambient pressure, if I set my analyzer to 20.6, it will read 20.9 when reading from a tanks of air. So at least I don't have to hook it up to an air tank every time I test things. If I find I'm having to reset it back to 20.6 on a regular basis, I figure the sensor is going bad.

Second reason is if the result is off my more that my expected flub factor. I get that this is the number you are looking for, but for me it depends on the situation. Almost all of my filling and blending is done on a boat with banked 32% and air. For the diving we are doing, it is typically 130-150' for the first dive and 130-110' for the second. For the 150's it pretty much straight air, for the 130's its half air, half 32%. Depending on where the tank starts, reading should be in the 25-28% range. If it is outside of that, I start questioning the analysis. For the 110's we're topping up what is left from the first dive with 32%, so again depending on what was in the tank to start with, I expect readings in the 29-31% range. Because I am responsible for a dive program, rather than just myself, if we are getting weird readings, everything stops and we work through the problem before anyone gets in the water. Fortunately, there are usually at least 3 analyzers from 2 different brands with different aged cells, so that is a pretty easy one to check.

But my comfort zone for reading variability from what I expect is pretty wide. Then again, I'm not starting with an empty tank, adding 433.2 psi of O2 at 15 cuft/min, cooling, then topping with hyperfiltered air and expecting to get a precise measurement.

-Chris

 

[Doppler]

Pretty sure you already use this to begin your calibration... but just in case.