Complacency or Confidence?: Follow Up

[Jim Wyatt]

Sensor accuracy is +/- 1%. So, a 32% mix could read anywhere from 31-33%. When it is hotter & more humid outside more error is introduced. I have a 13 ft3 bottle filled with air and I calibrate my sensor to this known gas before each usage.

There is already enough error inherent so I try to minimize the error with this procedure.

[jimdiverman]

I’m interested to see what level of accuracy yall think is acceptable. I think I remember from my nitrox class a decade or so ago that there was some rounding that took place and the mix should be within 1%. i.e. + or - 1%. I’ve been fooling around (no not like that guys) with my analyzer and cannot seem to get consistent readings. Mine reads to 1/10th of a percent i.e. 32.x%. Is there an analyzer that is "lab quality" on the market that might read to say 1/100th of a %? (that’s not bench size and costs 1K or more) seems to me that with an inconsistent instrument and slight differences in mix could result in a cumulative error of at lease a % or 2. 32 vs. 34. For me that’s bordering on unacceptable. anyone else agree. Am I way off the mark here? Id just like to really know what’s actually in my tanks.

thanks

--jm

I have not personally seen an analyzer that has more than one digit to the right of the decimal point. I agree with you that it would be more accurate to have the extra digit. I usually round up when entering a whole number into my computer to limit my MOD, but rounding down would give me a more conservative pO2 credit. Its never been a problem, but it is certainly not an exact science.

[jimdiverman]

Sensor accuracy is +/- 1%. So, a 32% mix could read anywhere from 31-33%. When it is hotter & more humid outside more error is introduced. I have a 13 ft3 bottle filled with air and I calibrate my sensor to this known gas before each usage.

There is already enough error inherent so I try to minimize the error with this procedure.

Jim,

That's a good idea about the spare air bottle. I will have to try that.

I have a Nuvair analyzer, and the manufacturer's instructions talks about adjusting the starting PO2 on a sliding scale based on temperature and relative humidity (i.e. at 20.8 or 20.7 instead of 20.9). They publish that table with the instructions.

So to do this according to Hoyle, are you also supposed to take temperature and relative humidity readings with your sling psychrometer prior to testing your gas? Further, should you actually test the air at the intake where the compressor that is feeding the air/nitrox system is pumping from?

Jim

[chimie007]

For me that’s bordering on unacceptable. anyone else agree. Am I way off the mark here? Id just like to really know what’s actually in my tanks.

My take. People are worrying way too much about the exact O2 content. The only time it really makes a difference is for deep dives like trimix where the O2 is so low that a 1-2 % up/down affects deco quite a bit.

In cave country 32% is the norm and is the gas used in probably 90%+ dives. It could be 30 or 34% and it won't make a difference. Deco and MOD are not an exact science. IMHO Hydration, working hard at depth, CO2 buildup are more important deco/tox factor than 2% difference in O2 content. People would be better off paying attention to these than being anal about the O2 content they have.

That being said the issue I care about is : I'm getting the gas I asked for since an error could be deadly. I always analyze my mix (while I have to confess that I recently have been borderline complacent in cave country. This thread should bring me back in line .

As for the accuracy of an analyzer, well it really depends on the way it is calibrated especially in high humidity environment. Calibrating in ambient air could lead to 1-2% difference if not corrected. The only real option is to calibrate from a scuba tank of known concentration.

[chimie007]

Here is table that corrects for humidity and temperature. Assuming summer weather in Florida, your analyzer should read about 20% in ambient air.

http://web.archive.org/web/200205151...s/cla-f_u3.htm

[jimdiverman]

Here is table that corrects for humidity and temperature. Assuming summer weather in Florida, your analyzer should read about 20% in ambient air.

http://web.archive.org/web/200205151...s/cla-f_u3.htm

Interesting chart. According to this chart, above 70 degrees, you never calibrate to 20.9, and even at 60+ degrees, when is the RH ever 10% in Florida? This could account for my readings being a few percent higher than the mix I am getting from Edd.

I would be interested to know what mixes people are getting when they analyze their gases at from say... Cave Excursions in High Springs, and Cave Adventurers in Marianna when you ask for 32%.

That would be an interesting poll.

-Jim

[jmaddox]

I would be interested to know what mixes people are getting when they analyze their gases at from say... Cave Excursions in High Springs, and Cave Adventurers in Marianna when you ask for 32%.

That would be an interesting poll.

-Jim

Interesting eh. Useful no. There are, in my opinion, way to many variables to generate anything slightly approaching useful valid data. Now if it was the same analyzer, same day, same time, same temp, same humidity, same shoes, and holding your mouth the same way, OK maybe the analytical results could be comparable. Has anyone thought about a calibration gas that is a known and certified O2 concentration? That’s the way we do our instruments at work. It wouldn’t be that difficult to make a calibration box that holds temp and humidity consistent. Just a thought. After all the concentration of 02 in your tank is not influenced by temp and humidity just the sensor right?

Thanks for the link to the table. that table brings to light the exact reason I want an analitical grade sensor. I need another significant digit to calibrate my sensor to 1/10th of a %.

I’ve tried several times to "close the loop" with my analyzer. Calibrate to air, analyze tanks, then let the sensor rest and see what it settles back to. Compare to original calibration #. Rarely is it the same. Usually it’s off by a 1/10% or 3. I don’t really know what this means other than the sensor is not accurate enough to come back the same conclusion. I guess this is not that big of a deal if accounted for by normalizing the concentration by adding additional time to your deco schedule? Say a couple of minutes at the deep stop and 10-foot stop. Hell I like to do this anyway just for funzies.

--jm

[chimie007]

Has anyone thought about a calibration gas that is a known and certified O2 concentration?

It's call a tank of air.

Humidity is only a factor if calibrating ambient air. If you use dry air from a scuba tank, calibrate to 20.9% then analyze your tank, there is no need to use correction factor. You should get within 0.5%.

Temperature also affects calibration. Gas coming out of a tank is cooler which is why if you use an air tank to calibrate you end up holding both temp and humidity constant between the calibration gas and your unknown.

Variables are variables... there could be a 100 but as long as they are kept constant they aren't an issue...

I’ve tried several times to "close the loop" with my analyzer. Calibrate to air, analyze tanks, then let the sensor rest and see what it settles back to. Compare to original calibration #. Rarely is it the same.

Try the same with air tank. I quite sure you will get good behavior.

I guess this is not that big of a deal if accounted for by normalizing the concentration by adding additional time to your deco schedule? Say a couple of minutes at the deep stop and 10-foot stop. Hell I like to do this anyway just for funzies.

IMHO If you think that you should adjust deco because of a 0.5 % O2 difference, you are really way pass the accuracy of deco models. Heck a different table/computer/desktop software will give you a larger range of numbers than 1 % in O2 content. It would be same thing as using seconds to calculate deco stops....

[karst.window]

In addition to analyzing everything everytime. I would also mention not to forget to breath your gas too. It should be tasteless and odorless. Unfortunately so is carbon monoxide... Mistakes can be made with regards to filling that can introduce contaminants into the breathing mix.

This becomes something you could possibly forget to do when staging or diving with bail-outs. You set up your regs, check the pressure but never breath the gas... Could have dire consequences.

There are few things worse than coming off the loop to a bailout that tastes like deisel fuel.

[Gary]

I’ve tried several times to "close the loop" with my analyzer. Calibrate to air, analyze tanks, then let the sensor rest and see what it settles back to. Compare to original calibration #. Rarely is it the same. Usually it’s off by a 1/10% or 3.

I find my analyzer coming off a hibernation peroid does this too. I think if you use it a few times - then calibrate, analyze, check calabration against original source - you may close the loop to within 0.1% or less.

I think tables assume +/- 1% on O2 though to those few extra fractions of a percent probably don't mean that much.