REFLECTIONS ON CALIBRATION
MULTIPLE SENSORS,
AND DESIGN CONSIDERATIONS
"THIS IS YOUR LIFE"
This is indeed your life you are taking into your hands. As I sit in my Gazabo with my fast internet line to my laptop, looking out on a beautiful spring day, I think about life and how much I love it, how much I love my family, my flowers, my fish, and diving. Every one of these is impacted by my designs to increase my safety.
When I was first asked to do this project it had one goalpost. Make a heads up display so I the diver would have a heads up to take another look down at my gauges even more often than I do. I check gauges constantly as a habit.
The beginings of the project was posted on the rebreather list and a long involved thread developed. I also received numerous off line suggestions, questions, and comments. Since I believe that all discussion is good as long as it is on a non emotional level, I enjoyed the discussions but realized that there were so many views and equipment configurations, that I could not please everyone so in the words of Ricky Nelson in the song "Garden Party", I pleased myself. The following is some information on the concepts and limitations of calibration, dual sensors, and other similar topics associated with the LEDHUD project.
The Right HUD Signal
There are as many opinions about how much information the HUD should present as people having those opinions. I chose a 5 level situation but not use the device to replace my regular displays. It is your display---it is your life. Make the conditional branches do what you want them to do. Other concerns are to either use two diodes or a dual color diode. That again is your choice. Another choice is to either run a set of wires to the LED at your mask or other place or to run a fiber optic to carry the signal without worry about flooding. Have fun and let me know what you do.
Voltage dividers and sensors
Look closely at the top of the schematic.......the output of the sensor is across the resistance of the
10,000 ohm pot. That gives the proper resistance to the sensor....the circuit is set up as a voltage divider so the output is proportional to the resistance of that portion of the pot. This way I can adjust the output of the sensor to a value below its regular output. If you use two sensors, you can make their outputs to the op amps match or even better yet make the outputs of the op amps match .
Calibration
Calibration is defined in metronics (the study and application of measurement) as setting the zero and slope of a curve that defines the relationship between the actual measurement and what the measurement should have been. If you refer back to the two graphs on the main page, you will see that I checked the output of the op amp against the measured input. It was linear. Thus I could predict the output from a known input or calculate the input from the output. I did the same with the output of the ADC through the stamp with the same results.
Actual dynamic calibration uses not a simulator but a known gas standard. It is going to be unit specific and situation specific. Many oxygen sensor systems are just calibrated in air expecting an answer of 20.9 %. Others use 100% oxygen. Lets look at the benefits and drawbacks to each.
Are you concerned more with accuracy at the 20.9 percent level or the 100 percent level. After you make that decision you are on your way to deciding how to calibrate. If you want the accuracy at the 20.9 percent level preventing Anoxia, consider using air to calibrate the system. If the op amp and ADC are linear, the percent of error will be the same across the range. For example, if you are getting a .208 that is an error of .001 over .208 or approximately .5%. At a theoretical PPO of 1.00 you would be off by .5% giving a reading of .995. How important is that to you? On the other hand if you calibrate at 100% oxygen and and get a .999 PPO a theoretical PPO of 20.9
would be off by only .1%. and the 20.9 percent oxygen would read .2089 PPO. The question is did you gain accuracy by expanding the range. The answer is....An Adult Diaper...depends....is the sensor linear? What is the precision? Precision is the reproducibility of a test. There are lots of things to consider. How long are you going to trust the system? Sensors die during dives according to many experienced RB divers.
Water Vapor Effects
At room temperature the air can contain up to about 3% moisture when the relative humidity is 100%. Calibration gas may not contain that much water or not contain water at all. Our rebreathers are wet environments. I would be concerned if I introduced pure oxygen into a wet environment and got an output representing 100%. Depending (there comes that word again) upon the level of saturation, it should read between 97 and 100 percent oxygen. The percentages hold up when you go down to the bottom of the scale so a 20.9 should read below .209 PPO but above .203. How important is that to you? You must draw the line, it is your life. After rereading the above consider this example....You pull the sensor out of your rebreather at the dock and calibrate. You set the final output to 20.9...thats what air is right....WRONG....it depends upon the moisture of the air at the dock. If the humidity is 100% there is at least 3% of the air taken up by moisture so the real oxygen content is 20.9 x .97 or 20.27. Do you care? Maybe not, but the devil is in the details. That is why I set all my oxygen display systems up with a output of .2 in ambient air. I am personally far more concerned with Anoxia than having too much oxygen since I stay way away from the 1.6 level and like it just fine at a PPO of 1. Hey, it is your life.
When Do You Calibrate?
Do you check things out at the beginning of each dive? At the beginning of each trip? The beginning of each day? What level of checking do you perform. In the project I put a reed relay in to act as a switch. The button command in code allows a branch statement that would put the hud in a calibrate mode. A comparison between the value received and the expected value can cause the green led to blink slowly if the value is too low, too fast if the value is too high, and a "Goldlilocks" steady blink if it is just right. Adjusting the pot makes the output of the sensor send the correct signal for the calibration gas. You must decide the limits you will accept. I cannot do that for you...It is your life.
How many Sensors?
As requested on the list, I designed the system for one sensor. It can be upgraded to 2 with the addition of a couple of lines of code and the addition of a couple of components. If you place another sensor in the system, the output should be across a 10,000 pot just as the single is. The positive output goes to an unused section of the quad op amp. The negative goes to the common ground. Install the two resistors on the op amp just as done in the single example and the output of that op amp goes to the unused channel of the ADC. The lines of code toggle the output of the ADC from one channel to another. Then you can compare, average, or indicate as you desire. It is your display and your life. You can add an additional ACD, gaining two more inputs. Check out the Stamp Manual for how to do this. It is there.....Now you can poll, average, etc making it as complicated or simple as you want. It is your display and your life.
How to house it.
That is going to depend upon a lot of the above and the rebreather you are using. I have about a month to make up my mind. It will certainly go in different locations on different rebreathers.