2 point calibration. Why?

mortenkjerulff

Hi all.

I am not a Meg owner (yet ), but could anyone explain me why the Meg is doing a 2 point calibration?
If sensor are linear it should be sufficient to do one calibration i ~100% oxygen. Or?

Morten

Sue

I guess it's a way of re-checking the sensors.

I did this on the Kiss as well at each calibration.

Sue

JonnyB

Well how do you cheak the linearity of a sensor if you dont have two points?

One way to do it is to take the factory specification and then check it at air, which will be to close to know if it is correct, so you would have to check it at ~100%O2, which is what you do on the meg, the air calibration is just a click to tell it that its in ambient air and it saves the values to check with the O2 one.

/Jonny

__________________
A quote from Crazyduck - In remembering our own Rob Davies.
"Outbound flight 777 heavy you are cleared for flight
Due west into that warm red Texas sunset
You have angels on your wings and divers memories on your six."

Rest In Peace

http://www.divetekcyprus.com
http://www.diveccr.com

PeteS

One possibility is that they are allowing for a zero error on the cells. Calibration on one point on the curve assumes that the other point on the curve is 0,0. If the other point is not 0,0 then the slope of the curve assumed by single point calibration is wrong.

If you calibrate at 0.209 bar PPO2 and assume 0,0 then the errors at 1.3 bar are relatively large. If you Calibrate at 1.00 bar PPO2 and 0,0 then the error on the slope of the curve is smaller.

And then there are linearity issues, because you're running the cells above where you tested them, and then..... and then......

Of course, in practice, the zero error is so small as to make bugger all difference............ unless you have some IT cells like the ones I tried a few years back that definately had a zero error of a few mV.

And I could be talking bollocks, which my wife says I'm fluent at. I think I'm definately better at languages after a few beers.

__________________
Pete S.

http://www.steggle.co.nz

Every silver lining has a cloud

Gilles

Two point calibration, along with the mV data that your calibrating with provides you with evidence on the health of your cells. If your O2 is 98%, then your theoretical mV in O2 should be 0.98/0.209 x mV in Air. If your O2 mV (taking care not to add pressure in the head during calibration) is some 2-3 mV below the theoretical value, then you have evidence of sensor non-linearity (PO2 can increase without a proportional mV response from your sensor(s). Not good .

Single point calibration and/or lack of the respective mV data cannot provide you with this indication.

__________________
Gilles
http://www.dirrebreather.com

mortenkjerulff

Thanks all.

So the procedure is really just to calibrate at 100% oxygen and then check the sensor readings in air ?

I had the impression that the electronics would fit a straight line between the readings in 21% and in 100% (or something like this). Which I find rather dangerous...

Morten

Gilles

Actually what is more dangerous is a single point calibration (with PO2 readings only) that provides no indication of cell linearity. A 2 point may provide an indication of cell failure (see attached).

True a straight line prediction is used, but this is valid whilst the cells are behaving linearly. A calibration that pressures the cells to a PO2 of about 2.0 would be ideal. A way of achieving this is to go off loop, reverse flush with O2 @10m, and read what the sensors say.

G

Attached Files

Good_Bad sensors.zip (11.6 KB, 46 views)

__________________
Gilles
http://www.dirrebreather.com

rdmmdr

most pic controled units will have in the sensor loop
first the sensor
second possibly the splitter for both heads incorporating the loaqd resistor
third the op amp
forth the a/d converter
and finally the pic
plus several connections along the way

if the hand set does not allow you to access the raw data from the a/d converter you dont know the health of the entire loop. good software should fail the unit if the a/d converter is out of range durring cal.

for example using a shitty op amp like the lm 324 could give you a zero offset of 65 mv on one channel and 0 mv on another. while i have found this to be constant and your pic could be hard calibrated to allow for this, it is a risk that i would not take.

using known voltage limits durring your cal confirms all parts of the sensor loop.

besides it does not take but one minute longer to do zero span calibration then juat span with a assumed zero, the key word here is assumed.

feel lucky that you are not doing real instrument cals. we do five points up and five points down twice to verify a sensor calibration. this allows for zero span offset and hysterisis.

rick

zèècalcoene

Hi all,

been reading some interesting threads lately..
However from what I've learned the last couple off months on this forum,
is that I check and calibrate my 3 sensors on 3 points;
1) calibrate VR3 sensors and other 2 sensors (rEvodream with HUD) in pure
oxygen.I made a setup (which I learned by J.Radomski on this forum)
in which I flow about 1l/m oxygen over the sensors without pressurising
them (above 1ATA).For this I use my KiSS style valve
2) I then close a valve fitted on a T-piece and allow pressure to increase
to 1.6 bar AND watch all 3 sensors..as u all know that VR3 sensor is
moving up slowier due to sampling rate but in a couple off seconds all 3
sensors should be stable and read 1.6 bar PO2..
3) I then flush the sensors with air and watch reading to drop to .20 PO2

Cheers

Ivan

Attached Images

	Kiss style O² injection.jpg (117.3 KB, 233 views)
	O2 calibratie close-up-small1.JPG (15.9 KB, 233 views)
	O2 calibratie setup-small1.JPG (58.5 KB, 231 views)

__________________
mCCR Dolphin Diver

RMalahowski

Just something I've run into over the years.

Sometimes you cannot get 100% oxygen depending on where you dive on this wonderful planet. It is better to calibrate to air than to calibrate to a fill that is not pure oxygen.

Also, just my opinion, 3 sensors are better than one to determine if you are getting good oxygen. In other words, the oxygen analyzers are usually one sensor devices. I generally use my three sensors to analyze the gas by pushing it through the loop rather than trusting the single sensor analyzers. Hence, calibrate to air and see what happens when you put the gas in your loop. You have to trust your electronics. Mine haven't let me down yet - and I know what they generally do.

__________________
Bob M.