Calibration
17
6 Calibration
What Do We Calibrate?
This section focuses on the procedure for offset and gain error calibration
of the input channels. The offset is a constant DC offset, and the gain
error is an error of the amplification factor in the input and output paths.
Both offset and gain errors are gain-dependent. The gain settings will
influence the offset estimate and the gain error itself. Therefore, it is
necessary to calibrate for all the voltage ranges.
The purpose of calibration is to ensure accuracy by estimating the
offset value and the gain error in all possible ranges (gain-settings).
Once this calibration data is stored, the input measurement and the
signal source can be adjusted back to the most accurate value.
The
PHOTON
+
uses sigma-delta A/D and D/A converters and an anti-
aliasing filter with a very high cutoff frequency. As a result, the
amplitude characteristics of the input and output are not frequency
dependent. Satisfactory flatness is guaranteed over the whole useful
frequency range. Therefore, we do not recommend calibration of the
amplitude vs. frequency characteristics.
We use the signal source (from the OUTPUT channel) of the
PHOTON
+
to calibrate the input channels. The key is to have a good signal
reference measurement, provided by a high-precision voltmeter. If the
voltmeter meets accuracy specifications, it does not matter where the
signal is generated.
Signal Source Used for Calibration
A single sine tone generated by the OUTPUT channel is used for both
input and output calibration. Multiple frames are captured, windowed,
and averaged to reduce the variance of the sine tone amplitude
estimation.
The amplitude flatness (ripple) within the analysis frequency range is
about 0.01dB. This very low number is due to digital implementation
of the anti-aliasing filter (for the inputs) and the reconstruction filter (for
the output). The gain characteristics are frequency-independent within
the analysis frequency range.
