Is the Intoxilyzer 9000 Accurate?

The Short Answer

Not always.

The Intoxilyzer 9000 is Michigan’s newest evidentiary breath-testing instrument. It is often assumed to be reliable because it is digital, modern, and approved for use in DUI cases throughout the State. In practice, however, newer technology does not eliminate error. In some cases, it conceals it.

At the Barone Defense Firm, we do not rely solely on the breath-test ticket printed by the police. We conduct a deeper evaluation by reviewing the Intoxilyzer 9000’s internal diagnostic records—known as COBRA (Computer Online Breath Archives) data—which frequently reveal instability and malfunction that never appear on a standard police report.

Why the Intoxilyzer 9000 Is Often Treated as “Bulletproof”

Many Michigan DUI cases proceed on the assumption that the Intoxilyzer 9000 is inherently accurate because it is computerized and newer than prior machines.

That assumption overlooks a basic reality: breath-testing instruments do not measure blood alcohol directly. They estimate alcohol concentration from breath samples collected under variable physiological and environmental conditions.

When a machine struggles to obtain a “valid” sample, that struggle is not incidental. It is scientific evidence.

Beyond the Breath-Test Ticket: What the Machine’s Own Data Shows

A breath-test ticket shows only the final number. It does not show how many attempts were rejected, how often the machine flagged instability, or what occurred before a result was finally accepted.

In a two-year audit conducted by Barone Defense Firm of one Michigan Intoxilyzer 9000 instrument (Serial No. 90-004872), we identified a recurring pattern of instability:

• Exception codes occurred in 40.6% of all test attempts across 219 subjects
• 17.4% of attempts were labeled “Unacceptable” samples, reflecting unstable breath profiles
• Nearly 40% of subjects were required to provide three or more breath attempts, with some individuals required to blow ten or more times before a result was accepted

These findings do not mean every Intoxilyzer 9000 performs the same way. Performance varies by instrument, maintenance history, environment, and operator.

However, when a machine rejects a substantial percentage of breath attempts before producing a reportable number, it raises a legitimate question:
Is the result reflecting stable analytical measurement—or the point at which the instrument finally accepted a breath variation that satisfied its programmed criteria?

Our experience reviewing instruments across the state shows that exception patterns differ by location. Some instruments perform more consistently than others. The only way to know is to review the underlying diagnostic data, not just the printed ticket.

The exception and instability patterns identified through these audits are not theoretical observations. We have relied on this underlying diagnostic data in:

• motions to suppress chemical test results,
• evidentiary hearings challenging breath-test reliability,
• negotiations with prosecutors where instrument instability materially affected case resolution,
• and jury trials where the reliability of the machine was directly at issue.

Courts and prosecutors do not evaluate a breath test based solely on the printed ticket. When underlying diagnostic data reveals repeated instability, exception codes, or forced breath attempts, it becomes part of the evidentiary analysis.

Each case turns on its own facts. But breath-test reliability is not presumed; it must be supported by stable instrument performance.

What Is COBRA Data—and Why It Matters

The Data Police Reports Do Not Show

COBRA data records the Intoxilyzer 9000’s internal diagnostics in real time. While the breath-test ticket shows only the final number, COBRA data shows how the machine was performing when that number was generated.

This data is critical because it reveals whether the machine was struggling, unstable, or detecting interference during the test.

Common Problems Revealed in Intoxilyzer 9000 COBRA Data

Radio Frequency Interference (RFI)

The Intoxilyzer 9000 continuously monitors for electromagnetic interference from radios, squad-car electronics, and nearby devices. If RFI warnings appear in the COBRA data, the analytical environment was not properly controlled.

Purge Failures and Baseline Instability

Before analyzing a breath sample, the machine must clear residual alcohol from its system.
Warnings such as “Purge Failure” or “Marginal Purge” indicate that the baseline may have been contaminated—artificially elevating the reported result.

Chemical Interferents

The Intoxilyzer 9000 uses infrared spectroscopy to detect ethanol, but it can misidentify compounds such as acetone as alcohol. Acetone may be present in individuals with diabetes, those following ketogenic diets, or during fasting.

Interferent alerts and other diagnostic flags are sometimes noted on the Michigan Forensic Breath Alcohol Analytic Report. However, a full assessment of instrument performance typically requires examination of the underlying COBRA data, which may contain additional diagnostic information and context beyond what appears on the printed ticket.

These systemic limitations in breath testing have also been examined in published defense scholarship. A discussion of how breath-testing instruments can generate reportable results despite internal instability appears here: Introducing the Intoxilyzer 9000 – Michigan’s New Breath Test Instrument.

Why the Machine Prints a Number Even When It Is Failing

Breathing Technique Can Change the Result

One of the most important—and least understood—facts about breath testing is that there is no single “true” breath alcohol concentration.

Peer-reviewed research demonstrates that breathing patterns alone can materially affect reported breath alcohol concentrations. Dr. A. W. Jones, a widely recognized forensic alcohol researcher, has published extensively on how breath-to-blood partition ratios and exhalation variables influence measured results. Similarly, Dr. Dennis Simpson, a Michigan-based university researcher, documented how longer exhalation duration can significantly increase reported breath alcohol values in the same individual:

• Longer or forced exhalations can significantly increase reported results
• Shorter, natural breaths often produce lower readings
• Differences of 0.02 to 0.04 g/210L can occur in the same person based solely on how the breath is delivered

When an Intoxilyzer 9000 repeatedly rejects samples, officers often instruct subjects to “keep blowing,” unintentionally encouraging breathing patterns known to inflate results.

The machine prints a number not because it has discovered a biological truth, but because it has finally accepted a breath pattern that satisfies its internal criteria.

These principles are discussed in more technical detail in a peer-reviewed Michigan Bar Journal article on breath and blood test uncertainty authored by Patrick Barone: Breath and Blood Tests in Intoxicated Driving Cases Why They Currently Fail to Meet Basic Scientific and Legal Safeguards for Admissibility

Decoding the Histogram: The Breath Test’s Digital Fingerprint

What the Histogram Can Reveal

The Intoxilyzer 9000 generates a histogram—a graph showing breath flow over time.

That graph can reveal problems such as:

Mouth Alcohol Patterns
A sharp spike followed by a sudden drop may indicate residual mouth alcohol from acid reflux or GERD that the machine failed to exclude.

Forced-Air Patterns
Extended or flattened curves can show that a subject was forced to exhale beyond a natural breath, capturing deep-lung air where alcohol concentration can be significantly higher.

These patterns are invisible unless the data is preserved and analyzed.

Operator Certification Still Matters

Michigan Administrative Rule R 325.2658

Under Michigan Administrative Rule R 325.2658(1), Intoxilyzer operators must maintain current certification and re-certify every two years.

Because the Intoxilyzer 9000 relies on manual entry and operator oversight, certification errors matter. If an officer’s certification lapsed—even briefly—the breath test may be legally inadmissible.

Frequently Asked Questions

Because Intoxilyzer 9000 diagnostic data can be overwritten or purged, early action matters.

Next Steps

If the Intoxilyzer 9000 was used in your Michigan DUI case, the reported number is only part of the story.

A review of the machine’s diagnostic and COBRA data can reveal whether instability or exception patterns affected the result. To speak with an attorney at Barone Defense Firm, call (248) 306-9158 or 1-877-ALL-MICH (877-255-6424).