Criminal defense attorney Jeff Sheridan grew concerned about the reliability of the state’s method for testing drunken drivers when clients kept telling him they tested for alcohol concentrations out of the ballpark for the number of drinks they consumed.
Sheridan watched the state adjust its Intoxilyzer machine in the past due to glitches. So it didn’t surprise him to think the current machine might have problems.
In 2006, he asked a judge for the “source code,” the computer code that controls the Intoxilyzer 5000EN by analyzing a person’s blood-alcohol content and converting the data into a numerical reading. The machine is used to test most of the 30,000-plus drivers arrested each year in Minnesota for drunken driving.
He struck it lucky in a Dakota County courtroom, breaking new ground in the DWI (driving while impaired) law when First Judicial District Court Judge Richard Spicer ordered the state to provide the source code to a drunken-driving defendant, who was challenging the revocation of his driver’s license.
“It was the first time anybody had been asked (for the code),” said Spicer, the first judge in Minnesota to order the source code to be produced. “I think they waited for just the right case.”
What might have been an open-and-shut matter, however, has been prolonged because the state didn’t have the source code and the manufacturer refused to disclose it, citing trade-secret rights.
Jeff Sheridan: Eagan attorney Jeff
Sheridan believes DWI defendants
should be allowed to analyze the
computer code that controls the state’s
Intoxilyzer 5000EN. He says the state
never completed a full validation study
when it replaced the machine’s
software.
As more and more judges followed Spicer’s lead, prosecutors began advising police to switch to blood or urine draws, instead of risking that a judge would throw out breath results.
While everyone waits for the state Supreme Court to decide whether defendants should be provided the source code — and if so, under what conditions — prosecutors and defense attorneys continue to disagree over whether the source code is even relevant.
While law enforcement agencies in Scott County have resumed breath tests, Dakota and Carver counties continue using the more time-consuming blood and urine tests until the Supreme Court issues a decision.
Not surprisingly, prosecutors argue the Intoxilyzer machine has been proven to be effective and defense attorneys are just on a wild goose chase, trying any tactic to get their clients off the hook.
Dakota County Attorney Jim Backstrom, president of the Minnesota County Attorneys Association, calls it an “undocumented fishing expedition aimed at delaying or avoiding justice.”
Criminal attorneys allege there are plenty of reasons to doubt the reliability of the machine, and the computer behind it is key to knowing whether the Intoxilyzer really is failure-proof.
Sheridan sees the matter as one of justice.
“If we’re going to use the machine and only the machine to put people in prison, why can’t we know how it’s going to work? Is there some national security at stake?” Sheridan said. “It seems so ridiculous to even have to ask for this. It’s so clearly discoverable.”
Does it work?
The Intoxilyzer 5000EN relies on infrared technology to determine the alcohol concentration of drivers. Versions of the instrument have been in use in Minnesota since 1984.
When a drunken-driving suspect blows into the machine, an infrared light causes alcohol molecules to vibrate or absorb light at a particular frequency. The difference in light emitted and received is computed to determine the percentage of blood-alcohol content.
Defense attorneys across the country have been requesting Intoxilyzer source codes to see if there are any errors that could be affecting the machines’ reliability.
In New Jersey, a study of the Alcotest 71000, a machine similar to the Intoxilyzer, found more than 19,400 potential errors in the machine’s source code. After a three-year legal battle, however, the state’s Supreme Court upheld use of the machine.
Backstrom said the source code itself is not going to tell anything about the reliability of the machine. “The source code basically converts the English language into machine language so the machine can do its work and come up with documentation for the results,” he said.
Criminal attorneys argue that the source code is important because the computer is the brains behind the Intoxilyzer 5000EN — and those brains are old, relying on an antiquated Z-80 microprocesser.
“The Intoxilyzer is nothing more than a glorified computer,” said defense attorney Sam McCloud of Shakopee. “It doesn’t even use the most up-to-date computer technology — it uses computer technology from the ’80s.”Prosecutors and the state of Minnesota maintain the machine has been tested rigorously by the state. “There are a lot of procedures and safeguards put in place to ensure these tests are accurate,” said Backstrom. “Intoxilyzer machines have been in place and utilized across America for many years. They are extremely reliable machines.”According to a fact sheet put out by the Minnesota Department of Public Safety, the Intoxilyzer was subjected to validation testing before being put to use, including tests that used simulator solutions with a known alcohol concentration, as well as testing on live subjects.“Based on the results of this validation testing, the (Minnesota Bureau of Criminal Apprehension) concluded that the instrument performed properly, yielding accurate and reliable breath-alcohol measurements,” the department said. “Each time the instrument is used, the results are measured against a simulator, which would indicate a potential problem with the instrument, at which time it would be examined and taken out of service, if necessary.”
Each time the machine is used, the suspect provides at least two breath samples and each sample is tested twice. The numbers must be within 0.02 percent of each other and the lowest result is used, if they vary. Simultaneously, the machine tests a solution with a known alcohol concentration.
Enough study?
The Intoxilyzer 5000EN has had its “brains” (two computer silicon chips) replaced three times, said Sheridan, but the state only performed a control study the first time.
That means the computer that currently controls the machine never went through a validation study to compare breath measurements against blood or urine samples taken from the same people.
One of the times the computer was changed was after the state Legislature decided samples taken from defendants had to measure within 0.02 percent of each other in order to be counted, Sheridan said. Before, there was no minimum range.
The machine’s computer was replaced again, Sheridan said, when it was discovered the machine wasn’t kicking out samples outside the 0.02-percent range like it was supposed to.
