Measurement System Analysis: An MSA guide with a focus on UV measurement

The proof of capability of measuring systems is essential to ensure the reliability and accuracy of measuring devices. In UV measurement technology, especially when using UV measuring devices, UV sensors and calibration lamps, the validation of measurement capabilities is of central importance. In general, the measurement system analysis (MSA) confirms the suitability of a measuring device for a specific measurement task.

The proof of capability of measurement systems was originally developed by the automotive industry. However, there is no specific standard that defines these procedures. However, the following values and forms are typically used.

Various statistical methods and quality criteria are used for the proof of capability. These include the evaluation of the repeatability, linearity and stability of the measuring devices.

Note that the terms "ability" and "capable" are synonymous with "aptitude" and "suitable". Both terms are therefore to be regarded as equivalent.

The MSA measuring equipment capability analysis proposes three procedures for this purpose.

• Method 1 examines the accuracy and repeatability of a measuring system. A known standard is required to determine the accuracy or deviation from the standard.
• Method 2 examines the repeatability of a measuring system, including the influence of the operator.
• Method 3 is generally used for automated measuring systems.

Measuring equipment capability analysis MSA and UV?

In practice, regular calibration of UV measuring devices is necessary to ensure consistent measurement results and to take possible ageing and contamination into account.

Standards and calibration lamps are available in our calibration laboratories, which is why the proof of capability can be carried out as an additional test according to procedure 1.

Here, the accuracy is known from the calibration and the repeatability is checked using, for example, 25 measurements.

Procedure 2 can only be carried out by the user himself, as the influence of the positioning of the UV measuring device and the operator on the repeatability cannot be simulated in the laboratory.

So what is done in MSA procedure 1?

It is determined whether a measuring device is suitable for the intended use under operating conditions. In practice, method 1 is used with or without a standard. The application with standard is advantageous as it is known and stable in terms of time and temperature.

Two characteristic values are evaluated, which are referred to as C_gk and C_g and compare the repeatability with the tolerance. The formulas are:

This means:

T = Tolerance

x_g = Average value of the measured values of the UV measuring device

x_m = nominal value of the standard

s_g = Standard deviation of the measured values of the UV measuring device

If no standard is available for measurement reasons, the calculation is omitted

of C_gk. In this case, with the help of a suitable measurement object, only the

repeatability C_g can be determined.

If a normal exists, the deviation of the mean value can be recorded as a statistical component and the so-called capability parameter c_gk is determined. If there is no standard, the capability parameter C_g is determined purely from the random component. In this case, the scattering range of the measuring system is increased to 4 times the standard deviation s_g and the weighting factor for the tolerance is adjusted.

C_gk values and C_g values greater than 1.33 indicate that the measuring device is sufficiently accurate in relation to the tolerance range and has sufficiently low scatter. The formula for C_gk and C_g can be rearranged and it can be seen that the value of 1.33 corresponds to 40 times the tolerance in relation to the repeatability. This means that the UV measuring device will scatter less than 1/40 of the tolerance.

Other factors for the quality of a measurement include high repeatability, traceability to national and international calibration standards and low measurement uncertainties. Calibration allows the deviation of a UV measuring device from the standard to be corrected. So that this is minimized.

Example: Target value 100 W/m² with tolerance +/- 10 W/m²

The tolerance is therefore 20, namely -10 to +10 W/m².

In accordance with the MSA specification, the measuring device must have a resolution that does not exceed 5% of the specified tolerance (+/-). So if the tolerance is +/-10 W/m², the resolution must be at least 1 W/m².

Approx. 25 measurements are now used to check whether the repeatability is sufficient. For this purpose, the measured values should lie within the tolerance approximately at the target value. The mean value and the standard deviation (e.g. in EXCEL =STABW.N(..)) are calculated from the 25 measurements.

Example (without normal):

25 measurements are taken, which have a standard deviation of 0.3 W/m².

The capability parameter c_g in the example is: c_g = 0.2 * 20 / (4 * 0.3) = 3.33

The value c_g is greater than 1.33 and therefore the proof of capability of the measuring system is deemed to have been passed.

In simple terms, the required tolerance is set in relation to the 4-fold standard deviation and evaluated.

Measuring range and linearity for the MSA

Other elements of the proof of capability are linearity and stability.

For the linearity test, the UV sensor must respond linearly over the desired measuring range. 

In a capability/linearity test, a distinction is made between the following situations:

• Proof of linearity on several standards (e.g. length standards)
• Without normal: linearity is verified separately

In addition to calibrations, we therefore offer linearity testing of UV sensors in the laboratory. Comparison receivers or source addition methods are used for this.

The background to this is that calibration is usually carried out at two points: 0 and at the target irradiance. This is done according to international standards. From a mathematical point of view, this results in an equation with an unknown constant that can be solved by two points.

In the UV measuring device, the device sensitivity is usually set via a resistance potentiometer or a calibration factor. This corresponds to an "unknown" constant.

The linearity check therefore checks this assumption in the measuring range. Errors due to offset and saturation are also detected.

For technical reasons, however, we cannot meet all test points exactly, as most UV lamps cannot be dimmed without changing their spectrum. We would be happy to discuss the range with you.

Stability and long-term stability are important for reliable UV measurements. These can be ensured and documented by regular comparative measurements and calibrations.

In summary, the guideline on the "Proof of capability of measuring systems" describes important steps for the validation and quality assurance of measuring devices. Consistent application of these steps can ensure the reliability and accuracy of UV measuring systems.

Once the UV measuring device has been characterized using a measuring system analysis according to method 1, other influences can be easily determined on site during use according to method 2.