X-Ray Output QC: Reproducibility & Linearity

X-ray output reproducibility and linearity are the two QC tests that confirm a radiographic system delivers a consistent, predictable air kerma — the foundation on which every technique chart, automatic exposure control curve, and dose estimate depends. Reproducibility verifies that repeated exposures at a fixed technique produce nearly identical output (coefficient of variation ≤ 0.05), and linearity verifies that air kerma per milliampere-second stays constant as the mAs station changes (adjacent stations differing by no more than 0.10 of their sum). Both criteria are set by the FDA performance standard in 21 CFR 1020.31 and are verified by a qualified medical physicist. ¹

Introduction

Radiographic imaging assumes a stable relationship between the numbers a technologist selects on the console and the radiation actually delivered to the patient and detector. Select 200 mA and 0.1 s, and the system is expected to produce the same air kerma every time, and roughly twice that air kerma at 200 mAs versus 100 mAs. When that assumption breaks down, everything built on top of it — the technique chart, the automatic exposure control (AEC) calibration, the exposure index target, and the dose estimate — becomes unreliable. ¹²

Two closely related quality-control tests protect that assumption. Reproducibility answers the question: if I make the same exposure repeatedly, how consistent is the output? Linearity answers a different question: as I move across tube-current and time stations, does the output track the selected mAs in a predictable, proportional way? A system can pass one and fail the other, which is why both are evaluated at acceptance and during the annual physics survey. ²³

This article explains what these tests measure, the exact tolerances in the federal performance standard, the worked math a physicist uses to grade them, how they connect to image quality and patient dose, and the practical pitfalls that cause borderline units to drift out of compliance. DRPS performs this testing as part of its diagnostic radiography physics services across Florida, Maryland, Virginia, Washington DC, California, Nevada, Pennsylvania, New York, New Jersey, and Delaware.

Topic Explanation

What do reproducibility and linearity actually measure?

Reproducibility (also called constancy of output) is the shot-to-shot consistency of radiation output when the technique factors are held fixed. A medical physicist places a calibrated dosimeter in the beam, selects a single technique (for example, 80 kVp, 200 mA, 100 ms), and makes a series of exposures without changing any control. The spread in the measured air-kerma values, expressed as a coefficient of variation, is the reproducibility metric. ¹

Linearity is the proportionality between selected mAs and delivered air kerma. Here the physicist varies the mAs — either by changing mA at fixed time, changing time at fixed mA, or using the console's mAs selector — and measures how the air kerma per mAs changes from one station to the next. In an ideal system, air kerma per mAs is a constant: doubling mAs doubles the air kerma. In a real system, small deviations are allowed, but they must stay within a defined band. ¹

The distinction matters because the two failure modes have different causes. Poor reproducibility usually points to an unstable generator, timer, or tube (arcing, contactor bounce, inadequate power supply, or line-voltage fluctuation). Poor linearity usually points to a calibration problem in the mA or time stations, space-charge effects at low mA, or timer inaccuracy at short exposure times. For related generator and beam-quality checks, see our guide to half-value layer and kVp radiography QC.

Where the tolerances come from

The controlling U.S. requirement is the FDA electronic-product performance standard for radiographic equipment, 21 CFR 1020.31. It is one of the few places where numeric radiographic-output tolerances are written directly into federal regulation rather than into a professional guidance document. Manufacturers must certify to it, and medical physicists routinely apply the same numbers as field acceptance criteria. ¹

Professional-society documents build on that regulatory floor. The ACR–AAPM Technical Standard for Diagnostic Medical Physics Performance Monitoring of Radiographic Equipment (revised 2021) defines who is qualified to test, what parameters are evaluated, and the expected cadence, while AAPM Report No. 74 (Quality Control in Diagnostic Radiology) and AAPM Report No. 150 / Task Group 150 (digital radiographic systems) describe the test methods and typical acceptance limits used in practice. ²³⁴

Key Technical Principles

Reproducibility: the coefficient of variation

Reproducibility is graded with the coefficient of variation (CV), the ratio of the sample standard deviation to the mean of the measured air-kerma values:

CV = \frac{s}{X ˉ} = \frac{1}{X ˉ} \frac{1}{n - 1} i = 1 \sum n (X_{i} - \overset{ˉ}{X})^{2}

where $X_{i}$ are the individual air-kerma readings, $\overset{ˉ}{X}$ is their mean, $s$ is the sample standard deviation, and $n$ is the number of exposures.

The regulation is explicit about both the limit and the method. Per 21 CFR 1020.31, the coefficient of variation of the air kerma shall be no greater than 0.05 for any specific combination of technique factors, and compliance is determined from 10 consecutive measurements taken within a time period of one hour. The standard further requires that all technique-factor controls be reset between measurements and that the percent line-voltage regulation for each measurement fall within ±1 of the mean value, so the test reflects the equipment rather than an unstable power supply. ¹

Worked example. Suppose a physicist records the following 10 air-kerma readings (in µGy) at a fixed 80 kVp, 10 mAs technique:

2.01, 2.04, 1.98, 2.00, 2.03, 1.97, 2.02, 1.99, 2.05, 1.96.

The mean is $\overset{ˉ}{X} = 2.005 μ Gy$ . The sample standard deviation is approximately $s = 0.030 μ Gy$ . Therefore:

CV = \frac{0.030}{2.005} \approx 0.015

A CV of about 0.015 (1.5%) is comfortably within the 0.05 limit, indicating good reproducibility. A unit creeping toward 0.04–0.05 is still compliant but is often an early warning of generator, timer, or tube-contactor degradation that deserves attention before it fails. ¹²

Linearity: the coefficient of linearity

Linearity is evaluated from air kerma per mAs at each station. Define $X_{k} = K_{k} / (mAs)_{k}$ , where $K_{k}$ is the measured air kerma at station $k$ . The regulatory criterion in 21 CFR 1020.31 for any two consecutive mAs settings (or two settings differing by no more than a factor of 2 when selection is continuous) is: ¹

∣ X_{1} - X_{2} ∣ \leq 0.10 (X_{1} + X_{2})

An equivalent and widely reported form is the coefficient of linearity (CL), computed for adjacent stations:

CL = \frac{X _{m a x} - X _{m i n}}{X _{m a x} + X _{m i n}} \leq 0.10

Worked example. A physicist measures air kerma at a fixed 80 kVp while stepping through mAs stations, holding time constant and changing mA:

mAs station	Measured air kerma $K$ (µGy)	Air kerma per mAs, $X = K / mAs$ (µGy/mAs)
5 mAs	4.85	0.970
10 mAs	10.10	1.010
20 mAs	20.60	1.030
40 mAs	40.80	1.020

Check the most divergent adjacent pair, 5 mAs and 10 mAs, with $X_{1} = 0.970$ and $X_{2} = 1.010$ :

∣ 0.970 - 1.010 ∣ = 0.040 versus 0.10 (0.970 + 1.010) = 0.198

Because $0.040 \leq 0.198$ , the pair passes. Expressed as a coefficient of linearity across these two stations:

CL = \frac{1.010 - 0.970}{1.010 + 0.970} \approx 0.020

A CL near 0.02 (2%) indicates excellent linearity. Deviations tend to be largest at the extremes of the mAs range — very low mA (space-charge limited) and very short exposure times (timer resolution) — so physicists deliberately include those stations in the test rather than sampling only the mid-range. ¹²

Reproducibility and linearity QC at a glance

The table below summarizes the two core output tests alongside the closely related generator checks a physicist typically evaluates in the same session. The reproducibility and linearity limits are the federal 21 CFR 1020.31 values; the kVp and timer figures are commonly applied acceptance tolerances rather than fixed federal numbers, so confirm the exact value against the current ACR–AAPM technical standard and your facility QC program.

QC test	What it verifies	Common quantitative criterion	Primary basis
Exposure reproducibility	Shot-to-shot output consistency at fixed technique	CV of air kerma ≤ 0.05 over 10 exposures in 1 hour	21 CFR 1020.31 ¹
Exposure linearity	Air kerma/mAs constant across mAs stations	\|X₁ − X₂\| ≤ 0.10 (X₁ + X₂) for adjacent stations	21 CFR 1020.31 ¹
kVp accuracy	Measured vs. selected tube potential	Commonly within about ±5% or ±5 kV (verify)	AAPM 74 / ACR–AAPM ²³
Timer / exposure-time accuracy	Delivered vs. selected exposure time	Commonly within roughly ±5–10% at clinical times (verify)	AAPM 74 ²
Half-value layer	Beam quality / minimum filtration	Meets 21 CFR 1020.30 minimum HVL for the kVp	21 CFR 1020.30 ⁵

Clinical Impact

Reproducibility and linearity are not abstract compliance metrics — they are the physics that makes technique charts and AEC behave predictably. Three clinical consequences follow directly.

First, repeat rates and unnecessary dose. If output is not reproducible, an identical technique can under-expose one image and over-expose the next. Under-exposed images that get repeated add dose and workflow burden; over-exposed images add dose silently because modern digital detectors mask over-exposure with post-processing. Output instability is therefore a hidden contributor to the repeat-and-reject burden analyzed in our repeat/reject analysis guide.

Second, AEC and exposure-index reliability. Automatic exposure control assumes the tube delivers a predictable air kerma so the AEC can terminate the exposure at the right detector signal. Nonlinear output distorts that relationship across stations, and the effect propagates into the exposure index used to monitor technique — the metric discussed in our digital radiography exposure index article and in the automatic exposure control overview.

Third, dose estimation and comparison. Diagnostic reference levels, protocol optimization, and any dose audit assume that "X mAs at Y kVp" corresponds to a known air kerma. When linearity drifts, those estimates lose their footing, and comparisons across rooms or over time become unreliable. Consistent, linear output is what allows a facility to trend and optimize dose with confidence.

Practical Optimization Tips

A defensible output-QC session follows a repeatable workflow. The details below reflect common practice; the physicist adapts them to the equipment and the current technical standard.

1. Set up the dosimeter correctly

Use a calibrated diagnostic dosimeter with a current calibration traceable to a national standard, and select a chamber or solid-state detector appropriate for the kVp and dose-rate range.
Position the detector on the central ray at a fixed, documented source-to-detector distance, free of backscatter material behind it where possible, and collimate to cover but not overfill the sensor.
Record temperature and pressure if an ion chamber requiring correction is used.

2. Run reproducibility with the regulatory method

Choose a clinically representative technique (a mid-range kVp/mAs commonly used on that unit).
Make 10 consecutive exposures within one hour, resetting the technique controls between exposures as the standard specifies, and watch line-voltage stability.
Compute the mean, standard deviation, and CV; compare against the 0.05 limit and against the unit's historical baseline. ¹

3. Run linearity across the real operating range

Step through mAs by changing mA at fixed time, then repeat by changing time at fixed mA, because the two probe different subsystems.
Deliberately include the lowest mA and shortest time stations, where linearity most often fails.
Compute air kerma per mAs at each station and evaluate every adjacent pair against the 0.10 criterion. ¹

4. Interpret, document, and trend

Do not stop at pass/fail. A unit at CV = 0.045 or CL = 0.09 is compliant but drifting; note it and shorten the recheck interval.
Keep the raw readings, not just the summary, so year-over-year trends are visible.
Tie findings back to service history — a linearity failure right after a tube or generator replacement points to a calibration step that was missed.

Common pitfalls to avoid

Sampling only mid-range stations. Linearity failures hide at the extremes; a mid-range-only test can miss them.
Ignoring line-voltage regulation. A noisy supply can masquerade as poor reproducibility; the ±1% line-voltage check isolates the equipment. ¹
Using an uncalibrated or energy-mismatched detector. The tolerances are only meaningful against a properly calibrated dosimeter.
Treating a "pass" as the end of the analysis. Trending borderline values prevents unplanned downtime and failed accreditation surveys.
Confusing reproducibility with linearity. They are separate tests with separate causes; report and act on them independently.

Regulatory Considerations

Radiographic output QC sits at the intersection of the FDA performance standard for the equipment and state radiation-control rules for the facility. The two work together.

21 CFR 1020.31 (radiographic equipment) establishes the reproducibility (CV ≤ 0.05) and linearity (0.10) requirements as a federal performance standard that manufacturers certify to and that physicists apply in the field. ¹
21 CFR 1020.30 (diagnostic x-ray systems and their major components) sets related requirements, including minimum half-value-layer (beam-quality) requirements by kVp and indication of technique factors. ⁵
State radiation-control programs regulate the use of the x-ray machine — registration, the requirement for periodic medical physics surveys, and record retention. X-ray-producing machines are regulated by the FDA (for the equipment) together with state or Agreement-State radiation-control authorities (for facility use), a different framework from the NRC's jurisdiction over radioactive material. Among the jurisdictions DRPS serves, Florida, Maryland, Virginia, California, Nevada, Pennsylvania, New York, and New Jersey administer their own radiation-control programs, while Washington DC and Delaware sit under direct federal oversight for byproduct material; confirm the specific survey and record requirements with the authority having jurisdiction.

The professional standard of practice — the ACR–AAPM technical standard and AAPM QC reports — defines the qualified medical physicist's role in acceptance testing and the annual performance evaluation, and accrediting bodies expect that a physicist has verified output performance. For the accreditation context, see our overview of ACR accreditation physics requirements. ²³

Frequently Asked Questions (FAQs)

What is x-ray output reproducibility?

Reproducibility is the consistency of radiation output when the same technique factors are selected repeatedly. Under 21 CFR 1020.31, the coefficient of variation of the air kerma for any specific combination of technique factors must be no greater than 0.05, based on 10 consecutive measurements taken within one hour.

What is x-ray output linearity?

Linearity describes how air kerma per unit tube-current–time (air kerma/mAs) behaves as the mAs station changes. Under 21 CFR 1020.31, the average air-kerma-to-mAs ratios at any two consecutive settings must not differ by more than 0.10 times their sum, so that doubling the mAs approximately doubles the output.

What is the tolerance for reproducibility of a radiographic unit?

The federal performance standard requires the coefficient of variation of air kerma to be no greater than 0.05 (5%) for a fixed technique, measured over 10 consecutive exposures within one hour. Many physicists apply this same 0.05 limit as their acceptance criterion during annual surveys.

How is x-ray output linearity calculated?

For two adjacent mAs stations with average air-kerma-per-mAs values X1 and X2, the requirement is |X1 − X2| ≤ 0.10 (X1 + X2). Equivalently, the coefficient of linearity (Xmax − Xmin)/(Xmax + Xmin) across adjacent stations should be no greater than 0.10.

Why do reproducibility and linearity matter for patient dose?

If output is not reproducible, the same technique can under- or over-expose from shot to shot, driving repeats or unnecessary dose. If output is not linear, automatic exposure control and technique charts cannot predict dose reliably across stations, which undermines both image quality and dose optimization.

Who performs x-ray output QC testing?

Acceptance testing and the annual performance evaluation are performed or supervised by a qualified or board-certified medical physicist, using a calibrated dosimeter. Technologists may perform some routine constancy checks, but the physicist establishes baselines and confirms compliance with regulatory tolerances.

How often should output reproducibility and linearity be tested?

Both are checked at acceptance and after major service on the tube or generator, and are typically re-evaluated during the annual medical physics survey. Facilities may also add periodic constancy checks between annual surveys as part of the radiation safety and QC program.

Key Takeaways

Reproducibility and linearity are the foundation of predictable output. Technique charts, AEC, exposure index, and dose estimates all assume the tube delivers a consistent, proportional air kerma.
The tolerances are federal. 21 CFR 1020.31 sets CV ≤ 0.05 for reproducibility (10 exposures in one hour) and |X₁ − X₂| ≤ 0.10 (X₁ + X₂) for linearity across adjacent stations. ¹
They fail for different reasons. Reproducibility failures point to generator, timer, tube, or power-supply instability; linearity failures point to mA/time calibration, space-charge, or timer-resolution effects.
Test the extremes. Linearity problems concentrate at the lowest mA and shortest exposure times, so those stations must be included.
Trend, don't just grade. A compliant-but-drifting unit (CV ≈ 0.045, CL ≈ 0.09) is an early warning worth acting on before it fails.
A qualified physicist owns the test. Acceptance and annual evaluation with a calibrated dosimeter are part of the professional standard of practice and accreditation expectations. ²³

Conclusion

Output reproducibility and linearity are among the oldest tests in diagnostic radiography QC, and they remain central precisely because so much depends on them. A radiographic system that cannot reliably reproduce its output, or whose output does not scale predictably with mAs, undermines every downstream tool the department uses to control image quality and dose. The federal performance standard gives clear, testable numbers — a coefficient of variation of 0.05 and a linearity band of 0.10 — and a qualified medical physicist turns those numbers into a defensible acceptance and annual-survey program. Facilities that trend borderline results rather than treating each survey as a one-time pass/fail catch drift early, reduce repeats, and keep their dose optimization on solid footing. ¹²

How DRPS Can Help

Diagnostic Radiation Physics Services performs acceptance testing and annual performance evaluations of radiographic equipment, including output reproducibility, linearity, kVp and timer accuracy, half-value layer, and AEC performance, using calibrated dosimetry and documented against the applicable federal and state requirements. Our board-certified medical physicists provide diagnostic radiography physics testing, accreditation support, and broader medical physicist consulting across our service locations, including Florida, Maryland, Virginia, Washington DC, California, Nevada, Pennsylvania, New York, New Jersey, and Delaware.

A strong output-QC program is not just about passing a survey. It is about giving the clinical team a system they can trust to deliver the dose they intend, every time.

Related Resources

References

U.S. Food and Drug Administration. 21 CFR 1020.31: Radiographic equipment. Code of Federal Regulations. ecfr.gov
Gray JE, et al. American Association of Physicists in Medicine. Quality Control in Diagnostic Radiology. AAPM Report No. 74. Madison, WI: Medical Physics Publishing; 2002. aapm.org
American College of Radiology, American Association of Physicists in Medicine. ACR–AAPM Technical Standard for Diagnostic Medical Physics Performance Monitoring of Radiographic Equipment. Revised 2021. aapm.org
American Association of Physicists in Medicine Task Group 150. Acceptance Testing and Quality Control of Digital Radiographic Imaging Systems. AAPM Report No. 150. aapm.org
U.S. Food and Drug Administration. 21 CFR 1020.30: Diagnostic x-ray systems and their major components. Code of Federal Regulations. ecfr.gov
U.S. Food and Drug Administration. Nationwide Evaluation of X-Ray Trends (NEXT): national surveys of radiographic technique and patient dose. fda.gov
International Electrotechnical Commission. IEC 61223-3-1: Evaluation and routine testing in medical imaging departments — Acceptance and constancy tests — Imaging performance of radiographic X-ray equipment. Geneva: IEC. iec.ch