OK, FMTYEWTK (Far More Than You Ever Wanted To Know) about making a calibration curve:
Well, like I said, the graphs that the company provided are just for illustration purposes, i.e., (1) the "data points" on them are drawn too large on too small a graph to use them as real data points, and (2) your individual sensor will not have the exact same response as another of the same kind, IIRC, the company said that the sensor responses are about +/- 15%, IIRC.
For the first part, you can use LabVIEW or do it manually, depending on your preference.
So, what you have to do is rig up a vessel in which you can place your sensor (with some kind of electrical feedthrough) and vary the pressure over the range you are interested in and measure it with a known "good" pressure sensor (you might even use the regulator from a gas tank for this, depending on your requirements). Then fill your container with gas at pressure A, take your measurement with your sensor, repeat for pressure B, pressure C, etc., until you have exceeded your anticipated range or the sensor shows no further response with increasing pressure. For your setup, this should be pretty quick, so take a lot of points all along the range. One thing you do know from the company literature is that there is likely to be a curved region and a straight region (maybe a curved region at the end).
At this point, the easiest thing to do would probably be to gather your data and forget about LabVIEW for a moment. We'll come back to it later.
Then, using Excel (or another spreadsheet, or graph paper), graph response (y) vs. pressure (x) and look at the graph. This last step is essential (look at the graph) and one which many students fail to do. You should be able to see where the points on the graph are in a straight line, and where they obviously curve up or down. If your graph is either one or the other, great. If not, determine approximately where the points change from one region to another.
If you use Excel or another spreadsheet, you can then fit each region to a mathematical model (in Excel, you do this by right-clicking on a data point and selecting "Add Trendline"). By trial and error, find a model which visually fits the data best. Then go to the next region (if more than one) and repeat.
Now, armed with that information, go to LabVIEW. If you didn't take your original measurements with LabVIEW, set it up to take them now. For each region determined in the last step, get response vs. pressure data as before (or use the measurements you saved from doing it earlier) and plug them into the appropriate subVI in the Mathematics->Fitting palette. You then have your coefficients for your calibration curve.
In order to determine the pressure from a sensor response reading, just plug in your regression equation with these coefficients (I wouldn't use scripting, just all the little math triangles).
If your sensors are steady over time, you won't have to do another calibration curve for a while, but you should keep your apparatus on hand and do it again soon to see just how stable your sensor is. Once the established calibration does not fit the data well, you have to recalibrate. By this time, you will know LabVIEW so well that you can do everything there (even doing your graphing in LV for the visual fit part) and it will go even quicker than the first time.
If that wasn't enough for you, this goes into the theory, although it deals mainly with linear calibrations. The "residuals" part starting on page 6 talks about getting the right model for each region, mathematically.
Have fun with it,
Cameron
To err is human, but to really foul it up requires a computer.
The optimist believes we are in the best of all possible worlds - the pessimist fears this is true.
Profanity is the one language all programmers know best.
An expert is someone who has made all the possible mistakes.
To learn something about LabVIEW at no extra cost, work the online LabVIEW tutorial(s):
LabVIEW Unit 1 - Getting Started</ a>
Learn to Use LabVIEW with MyDAQ</ a>
.
