= Spirometer calibration verification
\b[Spirometer calibration verification] checks whether a <spirometer> measures a known input within specified limits. It differs from <measurement calibration>, which establishes the relation between input and indication, and from an adjustment that changes that relation. A satisfactory verification applies to the tested equipment, assembly and conditions; it does not validate a person's expiration maneuver.
The primary 2019 technical statement specifies daily verification with a 3-L syringe cycled over multiple flows. Its stated measured-volume accuracy requirement is ±3%, and a filter used for testing is also used during verification.<Standardization of Spirometry 2019 Update by Brian L. Graham and others>[Calibration and Quality Control, 3-L syringe and multiple-flow verification]{ref} The known volume is useful precisely because it can be delivered at different speeds.
An original gain-and-offset model explains this design. If measured flow is $Q_m=gQ+b$, integrating one transfer of known volume $V_0$ over duration $T$ gives
$$
V_m=gV_0+bT.
$$
For the same $V_0$ at two durations,
$$
b=\frac{V_{m,2}-V_{m,1}}{T_2-T_1},\qquad
g=\frac{V_{m,1}-bT_1}{V_0}.
$$
A single transfer cannot distinguish gain from zero offset in this model; two different durations can.
Take stipulated 3.000 L transfers lasting 1 s and 6 s, reported as 3.020 and 3.120 L. The inferred offset is $0.100/5=0.020\,\mathrm{L\,s^{-1}}$, while $g=1.000$. The slow transfer is 4% high, exceeding the cited 3% limit. If both transfers instead read 3.060 L, this model gives $g=1.020$ and $b=0$: a 2% gain error. These are constructed instrument responses, not a calibration record for an actual device.
<Identifying flow-sensor gain and offset with timed transfers> relies on the model assumptions. Nonlinear response, leaks or an inaccurately known syringe volume can produce other patterns. Reference-volume uncertainty belongs in the uncertainty budget, and repeated transfers share it rather than independently re-estimating it. A test at several flows is therefore more informative than repeating one convenient delivery speed.
Back to article page
