The following excerpt about breath testing is taken from a portion of Chapter 5 from The Georgia DUI Defense Trial Practice Manual, © 1996 William C. Head. The reader should understand that the material provided is in no way comprehensive, but outlines certain legal attacks on breath testing. Historically, many breath tests (and some refusals) have been excluded prior to trial on other grounds other than the ones mentioned in the excerpt.

5.1.2.a. Is the Intoxilyzer 5000 better, and why?

During most of 1994, the State of Georgia tested two machines in field testing before choosing the Intoxilyzer 5000 as the "sole" machine to be used for breath testing. The other machine was the Datamaster II, manufactured by National Patent in Mansfield, Ohio. The Datamaster is used by South Carolina, Washington, Arkansas and several other states.

In some ways, the design of the 5000 is undoubtedly better. In others, it is not as "good" as other devices. Many Georgia law enforcement officers preferred the Intoximeter 3000 because almost no test subject was unable to provide a sufficient "blow" on the machine. The contrary has proven to be true on the Intoxilyzer 5000.

The 5000 (series 768) is supposedly superior in that it has five wavelength "filters" to the 3000's one wavelength "filter." The machine has an internal reference standard centered at 3.80 microns, plus two detectors centered at 3.40 and 3.47 microns. These filters help distinguish between alcohol and acetone. The machine is also fitted with a filter to detect acetaldehyde (a common by-product found in human breath when alcohol is metabolized by the liver). It also has a filter which is designed to detect toluene, a very common chemical in paint products. In addition, the machine has a filter which will detect water vapor, as do all machines of the 5000 series. (564, 566, 568 and 768 series.)

The filters for the 5000 are designed to detect and SUBTRACT out from the "reading" any interfering substances if properly calibrated. However, the GBI has required CMI to set up a different protocol for the Georgia version of the 5000. The GBI has trained all officers to discontinue any breath test which produces a readout indicating an interfering substance, and to take the subject to a hospital for blood testing. The machine (as configured in Georgia) is programmed to merely report the interfering substance, not try to deduct it and render an adjusted reading. In other words, the procedure is to not chance that the machine can accurately subtract out these substance readings, but get an accurate test of the person's blood whenever a subject's test indicates the presence of any interfering chemicals. This "safeguard" is deigned to protect the accuracy and reliability of a subject's test result.

If the machine is not operating properly to detect and report an interferant, the value obtained can be artificially high and the test operator would be unaware of that fact.

The blood test result will be much more likely to give an accurate analysis of all chemicals in the person's system, including alcohol. Until the July, 1995 inspections on Georgia's Intoxilyzer 5000 machines, no systematic testing for interfering substances was attempted by the area supervisors, because the GBI did not supply them with any chemicals with which to test for interfering substances. Even then (and at the time of publication of this book) only acetone is mixed with the alcohol solution to test for this single interfering substance.

The "rules" for this inspection procedure are not part of the GBI Rules for implied consent. They are only written down in a memo from the GBI to the area supervisors. In fact, until early 1995, there was not even a training manual for the Intoxilyzer 5000 given to students who had taken the course update (the so-called "refresher" course). When Georgia first purchased the Intoxilyzer 5000 in 1994, there was no manual on the 768GA. However, one has been published since. Prior to that, only a dozen pages of material from the manufacturer's service manual as documentary support during their four hour training course.

The "refresher" course given in Georgia was only 4 hours in length, compared to 40 hours required in Florida's program for their 5000 operators. Hence, the "refresher course" operators will not only SAY they know nothing about the machine and how it works, they DO KNOW NOTHING! Hence, cross-examination of the officer on the length of his/her training, plus the issuance of subpoena duces tecum for the written course materials utilized in his/her training course may be highly informative.

The "regular" training course for the Intoxilyzer 5000, which was taught to officers who were new to infrared testing in Georgia, is basically the same two-day (16 hour) course that has been taught on the Intoximeter 3000 for over 12 years. A manual was created around February 1, 1995, and is substantially the same text as the old Intoximeter 3000 training manual (including some incorrect and outdated information).

The 5000 has a heated collection tube (which is supposed to prevent condensation of warm, alcohol-laden breath) to provide the subject's breath intake. This is an improvement over the 3000, which had an unheated collection tube.

The 5000 also has a "slope" detector which is supposed to detect "mouth alcohol," such as when someone has burped, or has alcohol in their mouth. However, experiments have proven that the slope detector of the 5000 does not always work. One of the flaws of the machine is its susceptibility to variations in test results based upon the manner of "blowing" into the tube. This so-called "sampling error" can be caused by hyperventilating before testing, or by "overblowing" (packing the sample chamber with a forced "blow" such that extra molecules of ethanol are inside the chamber). Moreover, the slope detector can be "fooled" by having the subject perform a slow, steady blow into the machine, rather than the hard, forced blow suggested by the officer.

Beginning January 1, 1995, the state has mandated by statute that all subjects be given two tests on the breath machine [O.C.G.A. § 40-6-392(a)(1)(B)]. The computer program in Georgia's series 768GA machines was custom-designed to Georgia's specifications. Yet, the first computer chip software program utilized in the 5000 machines was flawed and the average test subject was generally incapable of providing a second "blow" into the device in a manner that would yield a numerical result. Some machines were defective when shipped into the state and would not readily accept ANY blow from a subject. For example, at one point, the City of Suwanee reported 8 of 10 subjects "refusing" the test, despite the fact that most subjects blew until they were dizzy.

Reports were rampant across the state about similar problems with the machines. In early shipments, as many as one-third of all machines shipped to Georgia were immediately returned to the manufacturer for corrective work. Despite the Intoxilyzer 5000's full database capabilities, sketchy hand-written records of these problems are being maintained by the state. This is due to the fact that Georgia's machines are not downloaded---EVER. These computer-generated records, if kept by the state, would provide a history of the machine's unreliability. This omission of use of the computer database by the state seems to be "by design."

In support of this contention, consider that these machines are 100% computer-friendly, and have a Z80 microprocessor chip and a serial port which can be downloaded by modem or by on-site computer. This way, a complete, accurate history of defects and aborted tests could be kept, yet the GBI has taken no steps to keep this data, and apparently has no plans to do so in the future. This is a continuation of the "ostrich" approach that has been the trademark of the GBI Division of Forensic Sciences for over 10 years. If this evidence was available to the judge and jury, proof of failure and test errors would undoubtedly be a compelling reason for a jury to not believe machine test results.

The series 768 machine has approximately seven (7) different configurations. Although the series of machine used in Georgia (the 786) is the only breath testing device in use in America which is capable of providing a "flow rate" of the subject's volume and "flow" of breath, Georgia opted not to purchase this technology. Such machines (when configured with this more advanced technology) do not have the balky "pressure switch" at all. They are equipped with a thermistor. A printout of this "flow rate" data would instantly tell the operator that the machine had malfunctioned, and that the subject was feigning compliance with the officer's request to "blow."

Even worse, even though these machines are fully capable of being fitted with an adapter which permits the capture and retention of the subject's actual breath "sample," the GBI opted not to have this important part installed on the machines being commissioned for Georgia. The samples (if collected) have a 98% accuracy rating for up to two years.

Although the 5000's capability to accept a "normal" blow and trigger the machine's pressure switch is critical to a valid test being obtained, no periodic maintenance procedures exist whereby the functionality of the pressure switch is checked. To check this (since Georgia purchased the 5000 with a pressure switch, and not a model utilizing a thermistor), a spirometer or other gauge is needed. Georgia's machine was purchased without the available self-contained "spirometer" technology.

Hence, Georgia selected and bought the Cadillac of breath machines, but left off the steering wheel (the thermistor model with the breath sample collection system), the brakes (the computer database) and speedometer (the volume/flow indicator). These omitted attachments or options help the state obtain convictions, but circumvent fair play for citizens who are asked to submit to tests on these devices.