Police Fail to Correct a Linearity Problem on In-House Annual Inspection

Updated: May 1

Let's suppose that you obtain disclosure or Freedom of Information documents respecting In-house Periodic Inspections by your local police service of their approved instruments. What can you do with that information?


Purpose of this example cross-examination:


To obtain an admission from the CFS scientist that the disclosed annual maintenance records indicate that the police failed to take any steps to correct control tests on inspection that indicated the instrument's response had shifted well beyond not only a 3 mg% tolerance, but well beyond a 10mg% tolerance.


To suggest to the CFS scientist that the calibration curve has shifted.


To suggest to the CFS scientist that the instrument's response is no longer as accurate and precise as it should be per manufacturer's specifications.


To suggest that police are not following their very generous 10 mg% tolerance SOPs.


To obtain an admission that measurement results on the instrument are no longer reliable.

To obtain an admission that the police did not stop use of the instrument and report the non-conformance to a superior.


[Make the disclosed or FOI documents an exhibit first.]

Q. All right. So, at this point in time we still have linearity for this particular instrument. Let’s take a look at June the 20th, 2012, page 28. Now, with respect – we’re looking at page number... A. Twenty-eight?

Q. ...28. So, page 28, it appears that there was testing that was done at three different values of 50, 100 and 300. A. Yes. Q. And with respect to those – that testing at 50, 100 and 300 again, the tolerance that the police are looking for is plus or minus 10 percent. A. Correct. Q. And the values, the two values that are obtained at 50, are 43 and 43. A. Yes.

Q. And those results are outside of 10 percent. A. They are, yes.

Q. There’s no note on that periodic inspection of the author of the periodic inspection stopping use of the instrument. He documented it, but there’s no indication of the author of that document stopping use of the instrument and


reporting it to a superior. A. My belief is that [the breath testing supervisor] was the program coordinator for [local] Regional Police for the breath program. Q. Yes.

[ If "you would have to ask him what happened here" should there not be an O'Connor order for contemporaneous documentation by the breath testing supervisor?]

A. During that time. You would have to ask him what happened here, but yes, an inspection here, it appears that the low standard of 50 milligrams of alcohol in 100 millilitres of blood is reading low. Q. Right. But there’s no notation in the documents to indicate that [the breath testing supervisor] has noticed or noted in any way, that he is getting a low calibration check result at 50 milligrams per 100 mils that’s outside of the – both outside of the 3 percent, in the manufacturer’s specifications, and outside of the 10 percent of the – of the standards that seem to be applied by [the local] Regional Police. A. That's correct. Q. And those numbers of course, are also outside of the plus or minus 5 – is it plus or minus 5 percent that the Alcohol Test Committee uses for purposes of an evaluation of a new type of approved instrument? A. I would have to look at the recommendation.


[see Canadian Society of Forensic Science Alcohol Test Committee

Equipment Standards and Evaluation Procedures] Q. All right, we’ll have – perhaps we can check that later. So – but what the interesting thing is, that the numbers of 93 and 92 at 100 milligrams per 100 mils seem a bit low, but they’re within the 10 percent. A. Correct. Q. Are they also within the 3 percent? A. No, they’re not. Q. And there’s no notation about that, obviously, on the document. A. On this document, no.


Q. And with respect to the results at 300 the results obtained were 303 and 300. A. Those are acceptable. Q. Yes. So, I just want to put to you the hypothesis that the calibration curve on this instrument, or at least if not the curve, then the way that the instrument interacts with the real world has shifted. We have an indication here of a change in the calibration curve that doesn’t seem to be lining up as a parallel curve up or down, the way that you postulated earlier in your evidence. A. Okay. So, the instrument has a reduced response... Q. Yes. A. ...to the alcohol standard at that concentration. I can’t explain why the result is what it is. There are number of possible reasons why the calibration checks could be low, for both the 50 and the 90 even though the 90 is – the 93 and the 92 are within the plus or minus 10 percent, you’d have to ask [the breath testing supervisor] why he chose to proceed with testing and then indicate below that the instrument is in proper working order. Q. All right, well, in layperson’s terms. A. Yes. Q. The results at 50 are a lot lower than they should be. The results at 100 are a bit lower than they should be. And the results at 300 are where they should be or slightly above. A. That's correct. So, the instrument has a reduced response to a concentration of alcohol at 100 and would actually report the result as somewhere around 92 or 93 and even further down in the concentration range, would be reporting a 50 as a 43.


Q. So, one inference that we could draw from this document is that the calibration curve has shifted. A. No. The – so the calibration curve is fixed. Q. Yes.

[He's right but - Do some serious research into the quadratic equation used by the manufacturer to set up the calibration curve. You will find that although the x's in the equation may be part of every instrument's firmware, there are parameters set for that instrument during each individual instrument's factory calibration. Those parameters then stay the same until the next factory or authorized service centre's calibration (re-calibration). However, the response of the instrument to a breath sample will change over time for reasons discussed elsewhere in this blog. Ultimately that means that the instrument's response over time won't match its calibration curve (the parameters set during factory calibration). That's why IR instruments need to be re-calibrated at regular calibration intervals. See also "Calibration Interval" NRC 6.8, NCSLI RP-1, Establishment & adjustment of Cal Intervals 2007, Simplified Calibration Interval Analysis. The failure of the Canadian criminal justice system to require "metrological supervision" of police and compulsory "Calibration intervals" is a huge systemic failure in our justice system. Quaere: Section 8 breach? ]

A. That doesn’t change from the time that it’s actually determined and put into the instrument using those standards. Q. Yes. A. It’s the response to the instrument that shifts Q. Yes. A. So, a calibration curve, from calibration curve to calibration – so if you calibrated the instrument one year, it’s possible that the next year you could recalibrate the instrument and it’s possible that the curve could shift. All right? As it can with any instrument, including ours. But for that particular time, it is fixed. It doesn’t shift. It doesn’t move. All right? And it’s the response of the instrument to the alcohol that has shifted and therefore the instrument is reading low or in this case, maybe the proper procedure wasn’t followed with respect to how the standard was used. Q. But we have an indication that at the very least, the calibration, if I can put it, of the instrument, has changed.


[In asking this question I thought I was asking whether calibration had changed in the sense that the "measurement" results of the instrument no longer matched reality (the SI unit traceability required by the Weights and Measures Act - see OIML definition of "measurement" as a comparison). He answered the question referring to whether or not the software parameters in the individual instrument set at the factory had changed.] A. No. There’s nothing anywhere in the published literature or in presentations in the international community or otherwise, that shows that the calibration shifts. It’s actually the response of the instrument to a concentration of alcohol has shifted. Q. The linearity had shifted?


A. No, ‘cause linearity is the range from 0 to 600. Q. Yes. A. So, it’s just simply the response has shifted. Q. I see, but.... A. Not the calibration curve ‘cause it’s fixed and the calibration curve is fixed in the instrument and is monitored by the software from the time that it is established. Q. But the – and so what’s happened is the response of the instrument has changed and yet the calibration curve in the electronics, saved in the electronics of the instrument is still the same. A. Correct. Q. And so, in terms of results, the measurement results, that we are likely to receive from that instrument are less reliable. A. They’re going to be lower, yes.


#linearity #nonconforming #calibration #annualmaintenance #inspection

#tolerance #linearity #nonconforming #calibration #annualmaintenance #inspection

8 views0 comments