¶ 1. Defendant appeals the civil suspension of his driver’s license and the admission of the breath-test results in his crimin.1 prosecution for driving under the in.1uence of intoxicating liquor (DUI). Defendant contends that because the test results *280were obtained after the testing machine registered a “fatal error,” the breath-test an.1ysis did not meet the requisite performance standards, and thus the necessary foundation for admissibility was not laid. We affirm the court’s decision denying suppression in the crimin.1 case, and reverse and remand the civil suspension.
¶ 2. On December 4, 2011, at around 3 a.m., an officer of the Burlington Police Department stopped defendant after observing him speeding and driving erratically. Based on the officer’s observations, he commenced a DUI investigation. After the officer had defendant perform standard field-sobriety tests, he arrested defendant and transported him to the police station for DUI processing. At the stationhouse, defendant agreed to provide an evidentiary breath sample. The officer used a DataMaster infrared breath-testing machine to conduct the breath an.1ysis. During the first attempt, the machine produced an error message of “standard out of range.” The officer restarted the machine and attempted again to obtain a test result. This time, the machine produced a result of .229 without an error message.
¶ 3. Defendant requested a second test result. The officer attempted another test but received another “standard out of range” error message. Again, the officer restarted the machine and obtained a result of .260 without indication of error.
¶ 4. Defendant was charged with DUI subject to crimin.1 proceedings pursuant to 23 V.S.A. § 1201(a)(2) and a civil license suspension under 23 V.S.A. § 1205. Defendant filed a motion to suppress and dismiss in both cases, claiming that after receiving a standard-out-of-range error, the officer’s training instructed him to use a different machine. Because the officer failed to follow the correct procedure, defendant argued that the result was not reliable, citing 23 V.S.A. § 1205(h). Defendant filed a supplemental motion to suppress and dismiss arguing that the discrepancy between the two tests negated their reliability and made them inadmissible. In support, defendant submitted a letter and affidavit from an expert.
¶ 5. On April 11, 2012, the court held a hearing on defendant’s motions. The day before, the State had filed a motion to allow its chemist to testify by telephone on the basis that the witness would be inconvenienced by the travel. Defendant’s attorney objected. The trial court denied the State’s motion because the State provided defendant insufficient notice of its request, and this prevented defendant from properly preparing for or conducting an *281effective cross-examination in both the civil and crimin.1 cases. The court then indicated its intent to dismiss the crimin.1 case based on the State’s lack of an expert. Pursuant to the State’s request, the court agreed to delay entering dismissal for ten days.
¶ 6. The court proceeded with the suppression hearing in the civil suspension proceeding. Without objection, the court admitted the chemist’s affidavit, the DUI processing form, the officer’s affidavit and the DataMaster tickets.
¶ 7, The Burlington police officer testified for the State. He described administering the breath-alcohol test to defendant. He explained that after he received the first error message, he understood from his training that he was supposed to transport defendant to a different location and use a different machine, but he decided not to proceed in this fashion. The officer testified that based on his experience, the DataMaster machine sometimes produces an error message when there is alcohol emitting from the defendant’s person in an en.1osed area. He therefore went ahead and restarted the machine and conducted a test. When describing his attempt to obtain a second test result, he explained that he received a second standard-out-of-range message before restarting the machine and obtaining a usable result. On cross-examination, he agreed that the standard-out-of-range message was a “fatal error.” He also agreed that, according to his training, he was supposed to proceed to a different machine after encountering such an error. Defendant introduced the section of the manual produced by the Vermont Crimin.1 Justice Training Council that catalogued this message as a fatal error.
¶ 8. Defendant presented expert testimony from a forensic consultant, who formerly worked as a state employee. The witness explained that she had reviewed the breath-test memory reports and status record summaries for the instrument used on defendant, as well as maintenance records from the Department of Health. She opined that it was not a “good choice” for the officer to proceed -with using the machine after it produced the standard-out-of-range message because the message was an indication that the simulator vapor was not reading within the required five-percent accuracy. She further testified that the particular machine’s reliability was suspect because the breath test summaries showed a history of errors, and because apparen.1y the machine failed its accuracy testing shortly after it was used on defendant and was sent back to the manufacturer. The witness also ques*282tioned the reliability of defendant’s tests because of the discrepancy between the results. She opined that the disparity between the tests was greater than is typically seen. On cross-examination, she conceded that a standard-out-of-range message does not always indicate that a test result is unreliable. She maintained, however, that given the discrepancy between the results, and the machine’s history of problems, she believed the tests were not reliable.
¶ 9. The court took the civil suspension matter under advisement. The court also informed the parties that it would allow the State to file a supplemental motion, and would wait ten days before dismissing the crimin.1 proceeding.
¶ 10. The State submitted a post-hearing motion to reconsider the dismissal of the criminal case. The State argued that under State v. Rolfe, 166 Vt. 1, 686 A.2d 949 (1996), and 23 V.S.A. §§ 1203(d) and 1205(h), the results of an infrared breath test are admissible if the State shows that the an.1ysis was performed by an instrument that meets the performance standards set by the Department of Health,1 and the instrument met those standards at the time of the test. The State alleged that defendant could contest the foundation facts, but could not otherwise challenge the admissibility of the test. According to the State, defendant did not contest the facts relevant to admission, in.1uding that: the DataMaster met the relevant standards, operated according to those standards during the tests, and was operated by a person certified to administer the test. Because these foundation facts were unchallenged, the State claimed that suppression was not the proper remedy. In the alternative, the State asked to proceed in the crimin.1 case without the result of the test, given the other evidence of intoxication.
¶ 11. In response, defendant claimed that he was indeed attacking the admissibility of the test in the crimin.1 proceeding. He alleged that because the officer did not follow the proper procedure following the error message, the DataMaster failed to meet the applicable performance standards, and the results were' inadmissible.
¶ 12. The court issued a written order on July 2, 2012. As to the civil suspension, the court found that the State’s expert’s affidavit *283provided a sufficient basis to establish that the test in this case was performed by an instrument that meets the standards of the Department of Health rules and that the instrument met those standards while an.1yzing the sample in this case. The court found that although the machine returned a standard-out-of-range message, it subsequen.1y returned a breath-alcohol reading, which indicated that the machine was working at the time of the test. The court con.1uded that the results were not invalid simply because the officer did not follow directions from the training manual. Rather, he elected to restart the machine based on other training and experience. The court explained that the officer’s failure to follow the instructions in the training manual did not warrant exclusion of the test result because deficiencies in the procedure were different from evidence refuting the reliability of the instrument. The court emphasized that even defendant’s expert conceded that the error message did not necessarily indicate that the results taken before or after the message were invalid. Therefore, the court con.1uded that defendant had failed to rebut the presumption in § 1205(h), and entered judgment for the State.
¶ 13. As to the crimin.1 case, the court denied the motion to suppress, con.1uding that defendant’s challenge went to the weight of the evidence, not admissibility. The court also con.1uded that there were no grounds to dismiss the crimin.1 charge because the State had sufficient evidence of impairment other than the test to demonstrate a prima facie case. The court denied defendant’s motion to reconsider its decision. Defendant then entered a condition.1 guilty plea in the crimin.1 case and appealed both judgments.
¶ 14. We begin with the crimin.1 proceeding. Defendant was charged with operating a vehicle under the in.1uence of intoxicating liquor. 23 V.S.A. § 1201(a)(2). Defendant filed a motion to suppress the results of his evidentiary breath test, which the court ultimately denied. On appeal from denial of a motion to suppress, we give deference to the court’s factual findings and review the legal questions de novo. State v. Fletcher, 2010 VT 27, ¶ 8, 187 Vt. 632, 996 A.2d 213 (mem.).
¶ 15. Defendant’s motion to suppress was based on a challenge to the admissibility of the breath-test results. The relevant statute sets out the foundation required to admit a breath *284test: “The an.1ysis performed by the state shall "be considered valid when performed according to a method or methods selected by the department of health.” 23 V.S.A. § 1203(d). We have held that the sentence should be read with “the term ‘admissible’ rather than ‘valid.’ ” Rolfe, 166 Vt. at 12, 686 A.2d at 957. Therefore, a breath-test result is “admissible if the State shows that the an.1ysis was performed by an instrument that meets the performance standards contained in the rules of the Department of Health, and the instrument met those performance- standards while employed to an.1yze the sample.” Id. at 13, 686 A.2d at 957. Defendants can contest the foundation facts, but “may not otherwise challenge the admissibility of a test result although they may, of course, urge the jury to give it no weight.” Id.
¶ 16. Defendant argues that the State offered insufficient evidence to demonstrate the necessary foundation facts. Initially, the court agreed that without expert testimony, the State lacked a sufficient basis to demonstrate the foundation facts. The court reconsidered its decision, however, and held that because admissibility is a determination for the court not subject to the rules of evidence, V.R.E. 104(a), the court could consider the state chemist’s affidavit in assessing admissibility. On appeal, defendant does not challenge the court’s ability to consider the affidavit in assessing admissibility, but instead argues that the affidavit alone is insufficient because defendant rebutted it with testimony from his expert.
¶ 17. The threshold for admissibility is set by § 1203(d). State v. McQuillan, 2003 VT 25, ¶ 7, 175 Vt. 173, 825 A.2d 804 (statute sets “the required foundation for admissibility of breath tests”).
As long as the State demonstrates that the an.1ysis of the challenged sample was performed by an instrument that meets the [Health] Department’s performance standards, the defendant may not otherwise challenge the admissibility of the test result; rather, the defendant can only contest the foundation facts or urge the factfinder to give little or no weight to the test.
Id. Here, the State met those requirements. The state chemist averred that the DataMaster machine used on defendant met the performance standards established by the Department of Health. The affidavit further stated: “The reporting of an alcohol concen*285tration of a person’s breath by the DataMaster is evidence that the instrument had successfully met all intern.1 and extern.1 quality control reviews and had been operating properly at the time the breath sample was an.1yzed.” In other words, the fact that the officer was able to obtain a test result indicates that the machine was operating properly at the time of the test.2 Therefore, the chemist’s affidavit was sufficient to meet the standard for admissibility of the breath-test result.
¶ 18. Defendant raises two issues he claims negate the admissibility of the tests. We consider each in turn. First, defendant argues that the officer’s failure to follow the procedures set forth in the training manual after receiving the standard-out-of-range message renders the test result inadmissible. In support of this argument, defendant’s expert testified that the officer did not administer the tests in accordance with the officer’s training. The officer agreed that according to his training, when he received the error message, he was supposed to turn off the DataMaster machine and use a different one. This testimony is supported by the admitted portion of the manual, which instructs that an officer “should consider [certain messages, in.1uding the ‘standard out of range’] ‘fatal errors’ and proceed to a different DataMaster.” Although we agree that the officer did not follow his training and the procedures set forth in the manual, we con.1ude that this failure implicates the reliability of the tests and not their admissibility.
¶ 19. As explained above, the admissibility standard is met when the test is performed according to standards set forth by the Department of Health. None of those standards dictate how the officer is to employ the DataMaster. The statute’s directive to the Department of Health is to set performance standards, not the manner of collection. McQuillan, 2003 VT 25, ¶ 11. Further, although the training manual specifies that a machine should not be used after a standard-out-of-range message has been received, *286the Department’s rules do not incorporate the manual.3 Thus, this Court has held that an officer’s failure to follow procedures outlined in the training manual does not affect whether the State has satisfied the foundation requirements for admissibility.4 State n Massey, 169 Vt. 180, 187, 730 A.2d 623, 628 (1999). Thus, if a defendant claims that the proper procedures were not followed in administering the test, the argument goes to weight, not admissibility. McQuillan, 2003 VT 25, ¶ 12. Although defendant styles his objection as one implicating whether the machine was working properly at the time of the tests, his evidence did not show that a result obtained after an error message meant the machine was operating outside the performance standards set by the Department of Health, only that the result might be unreliable.
¶ 20. This distinction is confirmed by the testimony of defendant’s expert. As to the significance of following the correct procedure, defendant’s expert offered that continuing to use the DataMaster after receiving the error implicated the reliability of the results. She did not opine that the DataMaster was incapable of meeting performance standards after returning an error message, or that the error message indicated some inherent nonfunctioning of the particular machine. The expert also did not rebut the state chemist’s statement that a machine’s reporting of a breath-alcohol concentration indicates that all required intern.1 and extern.1 quality controls have been met. See Vezina, 2004 VT 62, ¶ 5 (“[T]he fact that the DataMaster was able to detect the problem during the second sequence does not give reason to *287surmise that the instrument malfunctioned during the first one; rather, it is evidence that, had the problem been present during the first test, the instrument would have not issued a report”). Thus, while defendant’s expert’s testimony was relevant to show that the test result was unreliable in this case, it did not undermine the admissibility of those results.
¶ 21. Defendant’s second attempt to undercut the foundation facts for admissibility is based on the discrepancy between the two results. Defendant’s expert testified that there was a “large difference” between the two results and this raised a concern. Again, we con.1ude that the discrepancy between the two test results is insufficient to undermine the foundation facts necessary for admissibility. There is no requirement in the Department of Health rules that the two tests be within a certain percentage of one another. Further, the expert did not define how much of a discrepancy is too much or opine that the discrepancy indicated the machine did not meet Department of Health performance standards. The crux of her testimony was that she had concerns about whether the instrument was working properly. While this evidence is relevant to the reliability of the machine and whether the test results should be given weight, it does not undercut the fact that the machine met the performance standards necessary for admissibility of test results.5 Therefore, we con.1ude that the test results were admissible and affirm the court’s denial of defendant’s motion to suppress in the crimin.1 proceeding.
¶ 22. Next, we turn to the civil suspension adjudication. Civil suspension is a summary proceeding “designed to serve the remedial purpose of protecting public safety by quickly removing potentially dangerous drivers from the roads through purely administrative means.” State v. Anderson, 2005 VT 80, ¶ 2, 179 Vt. 43, 890 A.2d 68. At the fin.1 hearing in a civil suspension proceeding, the issues are limited and in.1ude “whether the test *288results were accurate and accurately evaluated.” 28 V.S.A. § 1205(h)(1)(D).
¶ 23. To achieve the goal of having “a speedy and summary procedure to get drunk drivers off the roads” the statute employs the use of presumptions. Anderson, 2005 VT 80, ¶¶ 3, 8 (discussing rebuttable presumption that test result over 0.08 taken within two hours of operation indicates person’s alcohol concentration was over 0.08 at time of operation); see also State v. Pluta, 157 Vt. 451, 453-54, 600 A.2d 291, 292-93 (1991) (describing how rebuttable presumption in a civil suspension proceeding aids in summary resolution). The standard for admissibility of a test result remains 23 V.S.A. § 1203(d). See Rolfe, 166 Vt. at 14, 686 A.2d at 958. Once the test is admitted, a presumption regarding reliability arises: “Evidence that the test was taken and evaluated in compliance with rules adopted by the department of health shall be prima facie evidence that the testing methods used were valid and reliable and that the test results are accurate and were accurately evaluated.” 23 V.S.A. § 1205(h)(1)(D). This presumption is not about admissibility, but concerns the burden of production. It allows the State to prove summarily that the testing methods were valid and reliable and that the results were accurate. Rolfe, 166 Vt. at 13, 686 A.2d at 958. It “shifts the burden of going forward with evidence challenging the test result on the defendant.” Id. at 14, 686 A.2d at 958. To rebut the presumption, defendant must present evidence to show that “the reliability and validity of the testing methods and the accuracy of the test results” is not true in defendant’s particular case. Id.
¶ 24. As set forth above, the State satisfied the § 1203(d) test regarding the admissibility of the evidence based on the affidavit from the state chemist. Therefore, the State had the benefit of the § 1205(h)(1)(D) presumption regarding the validity and reliability of the test results.
¶ 25. The question is therefore whether defendant’s evidence was sufficient to rebut the statutory presumption. See V.R.E. 301 (explaining effect of statutory presumptions). The trial court con.1uded that defendant’s evidence was insufficient because it was theoretical and not specific to defendant.
¶ 26. To rebut the presumption, defendant was required to “produce evidence fairly and reasonably tending to show that the real fact is not as presumed.” State v. Giard, 2005 VT 43, ¶ 9, *289178 Vt. 544, 871 A.2d 976 (mem.) (quotation omitted). Defendant’s expert testified that she had concerns about the reliability of the samples obtained in defendant’s case based on the officer’s failure to follow procedure, the history of the machine’s errors, and the discrepancy between the results. She explained that the standard-out-of-range error indicates that the “simulator vapor ... is not reading within the five percent accuracy required.” Therefore, her opinion was that tests taken after receiving such an error message could be compromised. She further opined that there was some question as to this particular machine’s accuracy based on the machine’s history of error messages and the difference between the two results. Given that defendant’s expert provided opinions specific to the instrument used to obtain defendant’s test results and about the particular circumstances surrounding collection of defendant’s tests, we con.1ude that there was sufficient evidence to rebut the presumption. Cf. State v. Wells, 172 Vt. 603, 606, 779 A.2d 680, 684 (2001) (mem.) (holding defendant failed to rebut statutory presumption where evidence was not particular to the facts and circumstances of the case). We cannot uphold the trial court’s con.1usion to the contrary.
¶27. Our conclusion is consistent with our recent decision in State v. Spooner, 2012 VT 90, 192 Vt. 465, 60 A.3d 640. In that case, a trooper performed a series of breath tests to measure the defendant’s blood-alcohol content. He obtained a reading of .101 on the first test; his second attempt returned a “standard out of range” error; the third returned a report of “invalid sample”; and on his fourth try he obtained a reading of .109. The State sought civil suspension of defendant’s operating license, and defendant argued that he was denied a second test. The defendant claimed that the second test was not reliable because the trooper failed to comply with his training, which instructed him to proceed to a different machine after receiving the error message. The trial court con.1uded that the trooper’s failure to comply with operating protocol undermined the reliability of the second breath-test result and thus deprived defendant of the benefit of a second test. The State appealed. On appeal, this Court emphasized that the question was not the admissibility of the breath-test evidence, but the reliability of the test result, and this was a question of fact reviewable for clear error. Id. ¶¶ 10-11. We acknowledged the statutory presumption of reliability, and explained that where an officer does not comply with correct operating procedures, “the *290fact finder is free to find the converse of the presumption.” Id. ¶ 13. Because evidence indicated the trooper did not comply with the proper procedures, we con.1uded that the court’s finding that the second test was unreliable was not clearly erroneous. Id. ¶ 18.
¶28. Thus, Spooner confirmed that an operator’s failure to abandon a machine that returned a standard-out-of-range message may undercut the reliability of any test result obtained thereafter, but does not affect its admissibility. It further confirmed that if a defendant produces evidence that a sample was collected in a manner inconsistent with an officer’s training, this is sufficient to rebut the presumption of reliability in a civil suspension proceeding. Id. ¶ 13. Because defendant introduced such evidence in this case, as in Spooner, he rebutted the statutory presumption in § 1205.
¶ 29. Once defendant rebutted the presumption, the State retained the burden of persuasion to demonstrate that the tests were indeed reliable. Giard, 2005 VT 43, ¶ 12. To support its case, the State relied on the expert’s affidavit and the arresting officer’s testimony that, based on his experience, the error message at times results from alcohol being emitted from defendant’s person in an en.1osed space.
¶ 30. “Whether a test is reliable or accurate is a factual finding.” Spooner, 2012 VT 90, ¶ 11. This Court reviews the trial court’s factual findings for clear error. Id. Here, the trial court did not weigh defendant’s evidence against the State’s or make a decision regarding reliability because the court determined that defendant had failed to rebut the statutory presumption. On the evidence before us, the trial court can determine that the test results are either reliable or unreliable. This is primarily a choice between the expert evidence presented through the State’s affidavit and the evidence of defendant’s expert witness. In essence, the State’s expert says that the machine’s reporting of an alcohol concentration in a person’s breath means that the machine is working properly and reliably; defendant’s expert witness disagrees on the facts of this case. It is up to the factfinder to resolve the con.1ict. We reverse and remand the judgment in the civil suspension proceeding for the trial court to make this assessment in the first instance.
¶ 31. Defendant’s fin.1 argument with respect to the civil suspension proceeding relates to the reliability of his second test. *291Under the statute, a defendant is entitled to request a second test. 23 V.S.A. § 1202(d)(5). Assuming that the first test is deemed reliable, defendant argues that he was denied a reliable second test. This argument relates solely to the civil suspension proceeding because a valid second test is not an element of a criminal DUI charge. See Spooner, 2012 VT 90, ¶ 17. In making this argument, defendant primarily relies on Spooner, wherein we affirmed the trial court’s finding that the second test was not reliable and therefore that the State failed to carry its burden of demonstrating that it complied with the requirements of § 1202. See 23 V.S.A. § 1205(h)(1)(E) (requiring State to demonstrate in a civil proceeding that it complied with § 1202). The question of the reliability of the second test is a factual one for the trial court to assess in the first instance. In light of our remand to the trial court to determine the reliability of both tests, this issue should be addressed by that court in the first instance if necessary.
The civil suspension is reversed and remanded. The order denying the motion to suppress in the crimin.1 proceeding is affirmed.
At the time this case was filed, § 1205 delegated rule-making authority to the Department of Health. Due to a legislative change, the Department of Public Safety is now responsible for adopting the relevant rules. See 2011, No. 56, § 16 (eff. March 1, 2012).
This point is the center of the disagreement between this opinion and the dissent. The dissent says summarily “The machine was not working.” Post, ¶ 32. To the contrary, the State’s expert affidavit states that the machine was working properly if it reports an alcohol concentration, which it twice did here. This is the conclusion of State v. Vezina, 2004 VT 62, ¶ 5, 177 Vt. 488, 857 A.2d 313 (mem.), a case cited by both this opinion and the dissent.
Following a legislative change, the Department of Public Safety is now delegated with the rule-making authority over breath-testing devices. We note that the Department of Public Safety’s rules do incorporate the manual, instructing: “The operator will follow procedures incorporated in the Vermont Crimin.1 Justice Training Council Student Manuals in effect at the time of testing and approved by the Commissioner of Public Safety.” See Vt. Dep’t of Public Safety, Breath and Blood Alcohol Analysis Rule (C)(II), Code of Vt. Rules 28 060 002 (2013). We make no judgment on the significance of this regulatory change.
Defendant cites cases from Missouri to support his contention that an officer’s failure to turn off a breath-testing machine after an error message makes the result inadmissible. See, e.g., Kennedy v. Dir. of Rev., 73 S.W.3d 85, 87 (Mo. Ct. App. 2002) (con.1uding that there were sufficient facts to support court’s finding that there was an insufficient foundation for breath test results after officer used machine that had returned an error message), overruled on other grounds by Verdoorn v. Dir of Rev., 119 S.W.3d 543 (Mo. 2003). The Missouri statute differs from the Vermont statute, however, in that it in.1udes a provision requiring an operator to cease using a machine that is not functioning properly. See id.
We have held that a disparity in two test results does not make the results inadmissible in at least two nonprecedential three-justice memorandum decisions. See State v. Howe, Nos. 2006-429, 2006-432, 2007 WL 5313288, at *3 (Vt. Mar. 28, 2007) (unpub. mem.) (holding that Department of Health rules do not require agreement between two successive results); State v. Springer, Nos. 2006-433, 2006-434, 2007 WL 5315017, at *4 (Vt. Feb. 28, 2007) (unpub. mem.) (explaining that tests need not be compared to be admissible, arguments go to weight of evidence). By our decision today, we make this a precedential holding.