5.7 Multipoint Precision Check
A multipoint precision check is used to determine the linear response of the instrument across its
operating range. The instrument is supplied with span gas at multiple known concentrations, typically
zero and at least four up-scale points, spread across the operating range of the instrument. The
observed concentrations are compared to expected values and the linearity of the instrument assessed
against local applicable standards.
Note: The instrument is inherently linear and the gain should not be adjusted at each individual
point. Non-linearity indicates a problem with the instrument (refer to Section 7 Non Linear
Response). Ensure that the instrument has been running for at least one hour before any
calibration is performed to ensure sufficient stability.
1. Ensure a suitable span source is connected to the instrument from a gas calibrator (Ecotech
recommends the Serinus Cal 3000) through the Calibration Port (refer to Section 5.5.1).
2. Generate and record the displayed span concentration for (at least) five stepped points (of known
concentrations) evenly spaced across the instruments measurement range (refer to the example
below).
3. Then using a program such as MS Excel, create an X Y scatter plot of expected concentration versus
the recorded instrument response and use linear regression to calculate the line of best fit and the
correlation factor (R
2
) – refer to the users local applicable standards.
Example for an instruments measurement range of 500 ppb:
a. For the 1st concentration, set the gas dilution calibrator to supply 400 ppb Ozone to the
instrument.
b. Allow the instrument to sample the calibration gas until a prolonged stable response is
achieved (the amount of time this takes is impacted by the calibration setup) and record the
instrument response.
c. Repeat the above steps using concentrations of 300 ppb, 200 ppb, 100 ppb and a Zero point.
Note: To highlight hysteresis errors, it is advisable to run the multipoint check in both descending
and ascending order - refer to applicable local standards.
d. Graph the results and use linear regression to determine a pass or fail as per applicable local
standards.
To Next Page
Loading...
Need help?
General
