Purpose of this cross-examination:

To educate the Court about the calibration sequence run on the approved instrument at the factory.

To obtain an admission from the Crown expert that without the proper auto-calibration AND creation of a unique calibration curve, the instrument is not reliable.

Every individual instrument has a unique calibration curve.

Every individual re-calibration at the factory or at the factory-authorized Canadian service centre produces a unique calibration curve.

Q. And – anyway. So, the signal passes through the sample chamber and when the signal comes through the sample chamber

there’s an electrical signal that’s produced and so during calibration the person who is doing the calibration is entering commands into the Intoxilyzer 8000 or 8000C which then, the software, during an autocalibration sequence in the instrument then generates a calibration curve.

From the Exhibit. These are the steps taken by a factory or authorized service centre technician to conduct an original or a re-calibration of an 8000C.

A. Essentially, yes. The specific procedure, again, I don’t know what is required, but that essentially is it, yes. Q. Okay. Now the indication of analytical theory in the document that I’ve just provided with you [sic], is that the curve is something, first of all, that has to be fit to the values. A. Yes. Q. And can you just explain to us, if you have a whole lot of data, and you’re trying to – I suppose one way to think of it would be to do it manually on paper. You have a whole lot of data that you’re hoping for some kind of relationship between – well, relationship between concentration of alcohol and the indication coming off the instrument, that’s what you are looking for, and if the data – if you can draw a line among the data that fits the data, then that’s one old fashioned way of trying to deal with that relationship but these days, we use electronic means to create a calibration curve.

A. Yes.

Q. All right. That’s what we mean by fit.

[See the university textbook referred to below dealing with how to calculate "fit".] A. Yes. So when the response isn’t linear then you have to use a mathematical equation to get the curve – to produce the curve from the results that are obtained, because you’re not having a one to one relationship. So, if you have a concentration of one, you expect a response of one, and in a linear relationship, if you have a concentration of two, you would expect the concentration to be two. But in a quadratic system, what happens is, is that concentration of two will be equal to much higher value. So, it’s not a one to one relationship.

[See also the blog entry "Instrumental Deviations from Beer-Lambert Law" with excerpts from the university textbook on creation of the calibration curve to "fit"]

Q. So C.M.I. in this document says the relationship is non-linear. A. Correct. Q. And secondly, they say that it is a quadratic polynomial curve fit. A. Yes.

Excerpt from Exhibit. Note that the manufacturer acknowledges that the relationship is non-linear. The relationship is unique to each instrument. The relationship is unique to each re-calibration. The variables in the formula are adjusted automatically DURING RE-CALIBRATION, not subject time of use.

Q. And the format of that polynomial curve fit is – and they have an equation of milligrams per litre times 10,000 equals A times percentage absorbance squared, plus B times percentage absorbance, plus C. That that’s the – the curve that they are building in the analytical system of the Intoxilyzer 8000 or 8000C. A. Correct. Q. Right? A. Yes. Q. The bottom line is, it creates a curve. A. Yes. Q. And that.... A. Which is not a straight line. Q. Right. And that it’s – yes. A. It’s kind of like a – I’m trying to think

how you would describe it. It’s almost like a “J” that continues out. Q. I suppose if we.... A. I’m trying to think of another way to describe how.... Q. Well, can I suggest a parabola or a portion of a parabola? A. Yes. Q. I think we all know what a parabola is. A. Yes. Half of one. Q. Right. A. Yeah. Q. So, this curve is created and this curve is unique to every individual Intoxilyzer 8000C instrument and every calibration or recalibration of that instrument. A. Correct. Q. It’s not as if there is the same calibration curve across all Intoxilyzer 8000s or all Intoxilyzer 8000Cs. A. The process will be the same it’s just the response of that particular instrument will be for that particular instrument.

Q. Yes. And so, the instrument as part of the analytical theory, creates something called a second order coefficient, a first order coefficient, and a zero order coefficient. A. Yes, which are the A, B and C.

Excerpt from Exhibit. The variables within the instrument's firmware for A, B, and C are set during the factory or service-centre calibration or re-calibration.

Q. Yes. And those numbers as a result of that autocalibration sequence, the number comes up with an actual number for the zero order coefficient – I’m on the page where it has minus 2-9-9 point 8-7, first order coefficient, 13-70 point 3-0 and second order coefficient, 9 point 5-7.

A. Correct. Those are mathematical values that are derived from that. Q. Right. And the instrument is doing that calculation as well, calculating a percentage absorbance value at each of 0 milligrams per 100 mils, this is at the top....

MR. BISS: And I’m sorry, I – the page Your Honour that has – go down through analytical theory, and I’m sorry, these aren’t – I didn’t put page numbers on them. It’s the one that has analytical theory milligrams per litre times 10,000. Do you see that page? THE COURT: Yes. MR. BISS: All right. Q. So, the instrument during the autocalibration sequence comes up with these zero order, first order and second order coefficients, but it also does a calculation, a percentage absorbance, and I’m – that’s why I’m looking at the top of the page now, and the instrument has two filters; one at about 3.4 microns and one at about 9.4 microns. A. Yes.

Excerpt from Exhibit. Note the calculation of values methodology.

Q. And 9.4 micron filter, that wavelength that is used for the quantification of alcohol, of ethyl alcohol in the sample chamber. A. Well, they’re both used for the quantification of alcohol. Only the 9.4 micrometre wavelength is the one that’s actually reported on the display and on the test record card. The 3.4 micron or micrometre wavelength is to check for the presence of interferences on the person’s breath. Q. Right, and by an interferant, meaning something like acetone?

A. Correct. Q. Right? Which is isopropanol. A. Yes. Q. All right. So, let’s look at the right hand side of the upper part of the page. The instrument has calculated a percentage absorbance for 0 milligrams per – milligrams per litre, I think that is. Um, yes, which translated as 0 milligrams per 100 mils of 0.204. A. Correct. That’s what it says. Q. All right and it does some standard deviation in um, is that relative standard deviation or relational standard deviation? A. Relative standard deviation. So the variability associated with it. Q. And then comes up with a – all right, then – and then does the same thing for the other solution at 40... A. Each of the.... Q. ...milligrams per 100 mils, 99 milligrams per 100 mils, 300 milligrams per 100 mils. A. Yes, it does it... Q. Right? A. ...for each of the solutions.

Q. And then it applies – the next pages have to do with applying those numbers to, as an example, to the creation of the curve, of the unique curve for that Intoxilyzer 8000 or 8000C and that particular calibration. A. Yes. Q. Right?

Excerpt from Exhibit. Calculation of A, B, and C using 40 mg/100mls alcohol standard.

MR. BISS: So, Your Honour can we mark this document as a – as an exhibit as well, please?

THE COURT: Yes, I think that’s Exhibit 26. And 26 is a further excerpt from the C.M.I. manufacturer’s manual for the Intoxilyzer 8000. EXHIBIT NUMBER 26: Further excerpt from the C.M.I manufacturer’s manual for the Intoxilyzer 8000 – produced and marked.

MR. BISS: Q. Now, again, without that kind of calibration being run at the C.M.I. lab, and in the absence of any other calibration of the instrument, without that calibration, the instrument would not be reliable across the range of 0 to 600 milligrams per 100 mils. A. Well, if you didn’t calibrate the instrument it wouldn’t give you – it would give you nonsense. It wouldn’t give you anything. Q. But it wouldn’t be reliable. A. Correct.

The following is slide show giving an example of the steps required to conduct an auto-calibration on an 8000C:

#crossex #calibration

#autocalibration #calibrationcurve #reliability