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.c. Weaknesses of the Intoxilyzer 5000

(1) The slope detector of this machine can be fooled be utilizing a slow, steady blow into the collection tube. In a demonstration conducted in Nevada on February 3, 1995, at the National DWI seminar of the Minnesota Society for Criminal Justice (MSCJ), in Nevada, Dr. Richard E. Jensen obtained a readout of 0.472 for a subject who had swished Scotch whiskey in his mouth. The slope detector DID NOT WORK.

(2) The air blank portion of the machine always reads 0.000, regardless of whether the ambient air sample has alcohol or other volatile chemicals (e.g., from paint fumes or solvents) in it. At the MSCJ National DWI seminar, this was proven by having an attorney (who had been drinking) blow into the collection tube for the 5000. The reading was approximately 0.060. Then, the machine's port for air blank intake was exposed to an open glass of Scotch whisky during the air blank sample. During the air blank mode (mode A), it read 0.000. Then, the same attorney (3 minutes later) blew into the machine again, rendering a 0.000 result!

The machine was malfunctioning, but gave a false test result without alerting the operator to the problem. Because the ambient air detector caused the baseline to go to "0.000," the 0.06 reading was reported as a false negative and therefore gave a ZERO reading.

(3) Despite the 5-filter series being used, the machine is still non-specific for ethanol. With five filters, it is (theoretically) the best machine available, but all infrared machines are prone to the same shortcoming. When combined with the state's refusal to install computer modem lines or to require periodic downloading of data to assure that all tests are being accounted for, a test subject would never know for certain whether or not an errant "readout" had been obtained. Because the machine is designed to allow the collection and retention of each breath sample (at a cost of ± $1.00), the state could provide a means of verifying the accuracy of each test via gas chromatography analysis. Plus, the state has no meaningful protocol or rules in place whereby the machine MUST be checked against these other chemicals to assure that all common interfering chemicals CAN be detected.

Moreover, since most machines are being run continuously, with no intervening diagnostics checks, machine error can only be surmised by the defense attorney, and argued to a jury under the aegis of "probable cause." This is a poor substitute for evidence.

(4) The state's failure to purchase a version of the 5000 which can check and report the volume/rate of breath sample of the subject creates doubt about any alleged refusal on the machine. When a person makes an effort to blow (as evidenced by puffed cheeks and getting red in the face) a strong argument to exclude evidence of a refusal can be made. See Department of Public Safety v. Orr, 122 Ga. App. 439, 177 S.E.2d 164 (1970) and Burson v. Collier, 226 Ga. 427, 175 S.E.2d 660 (1970).

(5) CMI has made many changes in its software and machine configuration in recent years. Often, these changes are hastily sent to users without adequate pre-testing to assure that it works in all applications. This was highlighted in Georgia when the original software would not accept a second "blow" on most attempts by subjects, and CMI changed the software. No new regulations or hearings were held on the propriety of this significant change in the "brain" of the 5000, pursuant to the Administrative Procedure Act; yet, the "guts" of the machine have been changed.

With decent regulations and rules, this lack of meaningful testing protocol would operate to void or place in doubt all breath tests on the 5000. However, Georgia's appellate courts have had a dismal history when it comes to reviewing due process claims against the state's breath machines. Lattarulo v. State, 261 Ga. 124, 401 S.E.2d 516 (1991); State v. Richardson, 186 Ga. App. 888, 368 S.E.2d 825 (1988); Magher v. State, 199 Ga. App. 508, 405 S.E.2d 327 (1991); Blanos v. State, 192 Ga. App. 835, 386 S.E.2d 714 (1989); Kuptz v. State, 179 Ga. App. 150, 345 S.E.2d 670 (1986); Oxley v. State, 210 Ga. App. 296, 435 S.E.2d 705 (1993).

Other states have shown a more enlightened approach to this subject and have suppressed all breath machine tests where proper procedure was not followed. State v. Flood, 523 So.2d 1180 (Fla. App. 1988); State v. Wilson, 116 Idaho 771, 780 P.2d 93 (1989); Commonwealth v. McGinnis, 511 Pa. 520, 515 A.2d 847 (1986).

The denial of pre-trial access to a breath machine does not constitute a violation of the Sixth Amendment right of confrontation. Confrontation in a criminal trial really means the right to ask questions and secure answers from the witness confronted. Kuptz v. State, 179 Ga. App. 150, 345 S.E.2d 670 (1986).

Denial of a pre-trial motion for access to a breath testing device is harmless and was no basis for reversal under O.C.G.A. § 40-6-391(a)(1). Gilbert v. State, 262 Ga. 840, 426 S.E.2d 155 (1993).

(6) The machines are too sophisticated to be repaired locally, and this results in maintenance being neglected or having needed maintenance delayed. The old Intoximeter 3000 machines could be completely disassembled by the area supervisors and virtually any repair could be made in the field. By contrast, South Carolina's Datamaster machines are all linked by modem to S.L.E.D. (State Law Enforcement Division) in Columbia, and can be diagnosed by phone, and often "repaired" by clearing and resetting a computer component over the phone line, without on-site repair. Likewise, the factory can make other repairs from Ohio via phone line.

At the end of 1995, only one Georgian, TFC Soffie Thigpen (an area supervisor) could repair certain internal parts of an Intoxilyzer 5000. All other repairs are made at the factory in Owensboro, KY.