Rudimentry sensor mV monitoring

Gilles

Attached is a short essay about the above topic.

In the hopes it generates some interest and contributes something useful.

G

Attached Files

Sensor Calibration basic princpls.doc (177.5 KB, 159 views)

__________________
Gilles
http://www.dirrebreather.com

Gibbon

Gilles,

Excellent - I hope this is the start to some wider study and better understanding for all.

If a gathering of data was to be done, your data logging sheet might be more usefull if people were to include the approximate water temperature that they were in.....

I think that one of my cell failiures would have been the temperature adjustment circuitry which caused the voltage to spike before disappearing almost entirely, I noted that Alex's ORR project has gone with cells that do not contain this circuit?????

Cheers,

Jon

Peter Piemonte

Thanks Gilles for an informative post. I dive a Classic KISS. Like Sutty, I have noticed a lower reading on cell 2 after dives that result in moisture being deposited on the cell faces. I always wonder why cell 2 seems the most affected. Under these circumstances, recalibrating at 100% O2 will usually result in cell 2 reading 3-4% low. The others seem unaffected.
Leaving the loop flushed with Dil overnight seems to help dry out the sensors on multi dive trips. I usually check the cell calibration just prior to diving under these circumstances. After recalibration, I normally see the sensors very close during the dive, and when flushing the loop with O2 at 20' I see ~1.55-1.6 on all cells. After drying the cells, I usually have to recalibrate cell two. Starting a cell MV log, seems a great idea. Peter

Danil

I think that the vapor issue may be worst with Teledyne cells, currently have 3 different brands of cells in my meg and the teledyne one always reads lower than the other 2 which always (more or less) are spot on after a 60 min dive...

__________________
--------------------
Dan Nilsson, Meg 67

silent running

Hello sabagia, in my experince with the Prism reading in air higher than .21 means you don't have 100% O2, maybe more like 97%. 100% O2 is sometimes hard to get and fill whips which have not been purged of air, can bring the % down further. Which of course makes me wonder what's in that other 3%...

And another thing, aren't O2 sensors for use in the high humidity, high PO2 environment of a CCR supposed to be able to cope with this? I have not found any difference in my sensor readings before a dive-dry-, during or after, with high humidity in the loop. -Andy











.

AD_ward9

If I may explain the reason why, as it may help on this thread.

As part of a safety case, we did a 5 year study of O2 sensors. We identified 14 failure modes. The test plan is attached so you see the rigour applied when it came to testing the sensors themselves. We liaised closely with several O2 sensor manufacturers to ensure we understood the sensor and its failures very well.

The only failure modes that give rise to a higher voltage than the correct voltage, in a sensor of appropriate design, is if the temperature compensation circuit fails. By eliminating that failure point, and chosing the right sensor construction, we know that when an O2 sensor fails, the output is lower than it should be. This means we can take the highest reading O2 value as being the correct one.

Failures in temperature compensation circuits can be sudden, or from the outset. None of the Teledyne sensors we tested had temperature compensation working to spec. The error of a completely failed temperature compensation circuit can be as high as 25%.

All galvanic O2 sensor produce a charge, which turns into a voltage when it is loaded with a resistor, usually about 270 Ohms (so do not waste time debating whether a 10K load or 100K load is best). The load should be as small as possible, from the point of view of the sensor manufacturer.
The smaller the load, the better, but the smaller the output voltage.

In the DL Open Revolution submission, we use cells that have 100 Ohm (1% tolerance) load internally, produce 4.5mV in air, and have a SMB male connector on them. We calibrate in air. This is an AII sensor: their quality is head and shoulders above all others we tested.

We check if the source is 100 Ohms by putting a small current via a 100K resistor onto it, from a waveform generated by a DAC. We also check the voltage level. We chose 100 Ohms and the low voltage to prevent someone plugging in compensated sensor from an Evolution by mistake, which has the same connector.

The unit has lots of very accurate temperature sensors, so the temperature compensation is performed digitally.

Cheers,

Alex

Attached Images

TP_O2_Sensors_060801.pdf (102.7 KB, 66 views)

Janos

Question: Have you never found it to be more than 25%? Eg, if I have a ppO2 reading of 1.2, can I be sure that worse case is all three sensors have failed high and it should be 1.5? (1.2 * 125%)

Janos

__________________
You can lead a horse to water but you can't climb a ladder with a large bell in both hands - Vic Reeves
www.hellfins.com/shed

AD_ward9

In fact, the output voltage can go up indefinitely if the load fails. That is another reason we check the load inside the cell is there.

If the temperature compensation circuit fails such that there is no load, the output voltage from the cell will rise and rise, because the charge generated by the sensor will have nowhere to go. That is, until it causes the input protection circuitry on the ADC to fail due to the overvoltage.

Thanks for asking about that point. I should have been clearer in my original email.

On the maximum that it can be if the load is there: it does depend on temperature. Assuming calibration is at 20C, and the maximum temperature 50C (the cells are not specified to work above 50C by manufacturers), then the worst you should see is 25%, based on our empirical results.

I should add, we did quite a lot of testing on big sheets of the membrane used in O2 cells. It is Zitex A-105 or G-104 (numbers from memory, so hope I have not swapped them). There are a few other membranes also used which we tested, such from those from GE, but all are fairly similar. The response time of the sensor depends on the thickness of the membrane. These membranes are not impervious to water: one can force water through them with quite low pressure. The KOH electrolyte inside the sensor is a very strong alkaline, and if diluted, the output from the cell will fall in value. The arrangements to keep water off the face are important factors in deciding how much water vapour gets through. It is very hard to breathe through these membrane sheets: they resist gas, but resist water about 50 times more. You can look up the water and air resistance figures on the web by Googling "Zitex A-105".

Alex

Gilles

I'm curious to find out if cell 2 is the only cell you notice moisture on? or do all 3 get moist, but only #2 is affected?


[/quote]Leaving the loop flushed with Dil overnight seems to help dry out the sensors on multi dive trips. I usually check the cell calibration just prior to diving under these circumstances. After recalibration, I normally see the sensors very close during the dive, and when flushing the loop with O2 at 20' I see ~1.55-1.6 on all cells. After drying the cells, I usually have to recalibrate cell two. Starting a cell MV log, seems a great idea. Peter[/quote]

I don't understand the purpose of this. Is it not preferable to calibrate the cells in the state you expect them to operate in (i.e. in a moist state)? Note that this is ok providing they are still behaving linearly.

G

__________________
Gilles
http://www.dirrebreather.com

Gilles

Thanks for the excellent feedback.

Would you have more info about where to acquire the above mentioned AII cells? I think I want some.

Do you know if the Meg has a problem calibrating with a 4-25 mV range. I don't think so but I'm unsure.

__________________
Gilles
http://www.dirrebreather.com