Calibration question

[doctormike]

I find most of my cells are nonlinear from 0.21 to 1.0 to a certain degree. That knowledge can't be imparted to the computer reading those values. Also I don't own a cell checker, but I might buy the wiring harness for my cells and bug @doctormike

Mi cell checker es su cell checker.

 

[DA Aquamaster]

[QUOTE="joshk,

Post 1 of this thread is where Bruce P is quoted (from RBW) as specifically not calibrating so you can "see" if the cells have drifted. I disagree that this is in anyway a good idea. Not based on O2 sensors in particular, just from the entire premise of what calibrating is trying to accomplish which is to most accurately measure ppO2 on the upcoming dive(s).

I cant get over 1.5 or 1.55 at 20ft at the end of the dive because I have to breathe lol. And the cells are getting soggy. And if I am midwater trying to do multiple complete O2 flushes is mucking with my buoyancy big time. I can usually spike over 1.6 at ~24ft. I don't really do it very often as its only the coarsest of checks.

My experience has been the same. 1.5 to 1.54 is pretty normal at the end of a dive after thorough O2 flushes at 20 ft.

About a year or so into our CCR experience, I got a cell checker and started not just testing cells before trips, but also checking cells pre-dive and post dive. I know this will shock some people but cells that have been moistened in the head for a 3-4 hour dive just don't respond as well as cells that are dry.

As a result, I'll calibrate on the first diving day after a break with dry cells because it's what's available. However, I'll also re-calibrate immediately post dive with normal moist cells. On the second day I find that all three cells will read very consistently with each other - far better than on day 1 with the dry calibration - and with much closer alignment with what is actually in the loop under real world conditions.

A cell checker also highlighted that true linearity is a bit of a myth, or at least is something that is either taken out of context or not actually found in the wild. If I calibrate at 1.0, I get acceptably accurate readings in the range that matters from .8 to 1.2 with errors in the 1% to 2% range (at an actual PPO2 of 1.2, a 2% error will give you a reading of 1.17 - no big deal). However while I may have .05% less error at a PPO2 of 1.0, and around 1%-2% at 1.2, the deviation from the theoretical maximum will always continue to increase as I get further away from that calibration point. It's not uncommon to see a 5% or 6% error at a PPO2 of 1.6, which means at an actual PP02 of 1.6 I'm only going to see a reading of 1.54 to 1.52 on the computer. Until I figured that out, I blamed the low reading at 20 ft on un-swept volume in the unit and/or on less than perfect O2 flushing.

 

[rjack321]

A cell checker also highlighted that true linearity is a bit of a myth, or at least is something that is either taken out of context or not actually found in the wild. If I calibrate at 1.0, I get acceptably accurate readings in the range that matters from .8 to 1.2 with errors in the 1% to 2% range (at an actual PPO2 of 1.2, a 2% error will give you a reading of 1.17 - no big deal). However while I may have .05% less error at a PPO2 of 1.0, and around 1%-2% at 1.2, the deviation from the theoretical maximum will always continue to increase as I get further away from that calibration point. It's not uncommon to see a 5% or 6% error at a PPO2 of 1.6, which means at an actual PP02 of 1.6 I'm only going to see a reading of 1.54 to 1.52 on the computer. Until I figured that out, I blamed the low reading at 20 ft on un-swept volume in the unit and/or on less than perfect O2 flushing.

This is an unappreciated 'flaw' in how we calibrate sensors. In every other environmental monitor or instrument I work with the calibration range, regardless if its a one, two or multipoint calibration, extends from below the expected measurements to above the expected measurements. When a measurement unexpectantly falls above the calibration range the measurement is either 1) diluted and remeasured 2) redone with a properly calibrated meter or 3) flagged as an estimated value.

Unless you dive at a SP of 1.0, we are almost always diving outside the calibrated range of the cells. The higher the SP, the more small calibration deviations and changes at 1.0 effect the slope of the line further above. Which is a good reason for calibrating as close to the conditions of your dive as possible. Then the error from true of our estimations is minimized. And this is for properly functioning cells.

Personally I calibrate at the start of a dive series (like a week vacation) and anytime the unit has sat in my garage more than a week or so. I don't worry about day to day or dry vs wet differences although I think its important to keep them in mind when you think about "what is accurate" and what kind of precision you should expect at a 20ft flush or if your SP is at 1.45 after 3 hrs in the water.

 

[doctormike]

John from Narked@90 responded to my question about the quality of my first run data:

"There are some very critical things to getting good readings.
- Using 100% oxygen or putting the correct gas% in the spread sheet
- Flushing the chamber fully and completely
- Slowly raise the pressure while giving a little tap to the gauge while you look at it directly and not off at an angle.

Looking at your reading and working things back…I would say:
1, The wobbles in the line is you not getting the pressures nailed, this may be because your not looking square on to the gauge.
2, if I change the % of o2 used I get about 93% o2 before the line looks good apart from your wobbles.
So either you did not flush the chamber correctly and well or your o2 supply is not 100% or both"


I'll keep practicing!

 

[rjack321]

I find I have to flush the mini checker A LOT to get 100% in there. There are dead spots in that space that trap ambient air.

 

[tbone1004]

I find I have to flush the mini checker A LOT to get 100% in there. There are dead spots in that space that trap ambient air.

that's why I was happy to get a deal on the De-Ox since I can pressurize it to 10bar with air so I don't have to worry about purging

 

[doctormike]

I find I have to flush the mini checker A LOT to get 100% in there. There are dead spots in that space that trap ambient air.

OK, will try again this weekend. Wonder if shaking or rotating it while filling would help?

Also, gotta be careful and keep the vent button pushed down hard during the flush. I must have accidentally let up on it a bit and I wasn't watching the gauge, so the pressure ran way up, and the head blew right off! It must have gone over the burst pressure. Amazingly, it seemed undamaged and I was able to screw it back in and run the test.

I don't know if that damaged the gauge. It went back to 0 and seemed to work fine for the next run, but I have an email in to John to see if that is going to be a problem going forward.

 

[rjack321]

Not sure about gauge damage. I would suggest not flushing by pressing both buttons at once since at least on mine it doesn't quite vent as fast as it fills.

I fill to 1.5 - 1.8 bar and vent.
fill and vent over and over about 6 to 8 times.

 

[doctormike]

Not sure about gauge damage. I would suggest not flushing by pressing both buttons at once since at least on mine it doesn't quite vent as fast as it fills.

I fill to 1.5 - 1.8 bar and vent.
fill and vent over and over about 6 to 8 times.

OK, maybe that's safer. In any event, I now know to look at the gauge while doing this. I figured that constant gas flow through the pot would be better at filling it with 100% O2.

The manual does recommend the two button method:

"Gently press the O2 INJECT and the PRESSURE RELEASE buttons the same time and watch the pressure gauge slowly rise until it stabilises and rises no further, then look at the mV readings and when they have stopped rising as well, the pressure pot is fully flushed with the supply gas. Now release the O2 INJECT button only and let the pressure drop back to zero and the mV stabilise."

 

[rjack321]

I haven't read the manual in years, plus I'm a guy lol

I don't think I could track all the pressure and mV readings simultaneously while pushing 2 buttons. Plus at least on my Meg handset the mV reading screen will time out before the chamber is flushed. So I need even more fingers to toggle back to that. I just flush a whole ton of chamber volumes one at a time. Seems to work, but I honestly don't test the cells unless there's a reason to suspect them. The Teledyne/vandergraphs from ISC last me a year at my dive rate and I replace them. I flooded one set and replaced them and had one fail in FL which was replaced under their 1yr warranty. I have not run into a significant deviation from linearity but I admittedly don't go looking for one unless I have suspicions. I dried out and tested the previously flooded ones, they were ok at 1.0 but by 1.8 true (according to the pot) they read 1.71 Which in an odd way is a bit confidence inspiring that even cells which are that trashed still basically function.