Calibration and Noise

Precise calibration measurements were performed on the APV25S1, which was mounted on a single-chip hybrid (fig. , p. ). Small capacitors are integrated in the PCB (printed circuit board) hybrid to allow pulsing by an applied voltage step. The placement of such capacitors with an associated voltage divider is difficult in terms of precision and excess noise introduced to the corresponding channels.

Pulse scans were performed for both internal and external calibration. The results with a nominal charge injection of $1\,\rm MIP$ ($22500\,\rm e$) are shown in fig.  for both peak and deconvolution modes. The waveform with external pulsing is not perfectly straight due to limited precision of the external delay used for this scan. While the shapes for internal and external calibration are congruent, their amplitudes differ considerably. The precision of the internal calibration circuit is limited for reasons discussed in section , p. . For this particular sample, the internal calibration charge is about $40\%$ higher than the externally applied charge with an estimated accuracy of $\pm 20\%$.

Figure: Comparison of the APV25S1 response to internal and external calibration pulsing for both peak and deconvolution modes.

With a known calibration amplitude, the APV noise can be obtained in absolute numbers. We will use the external calibration for this evaluation, since it is the worse case in terms of absolute noise. Thus, the resulting numbers rather give an upper limit for the noise figure.

After subtracting the noise of the readout system, the resulting equivalent noise charge (ENC) for channels with no capacitive load is shown in tab. . These numbers perfectly agree with the design values and with manufacturer's tests [64].

Table: Comparison of the APV25S1 equivalent noise charge design values to measurements with external calibration. The noise of the readout system has been subtracted.

Mode	Design	Measurement
	$\rm ENC\,[e]$	$\rm ENC\,[e]$
Peak	250	267
Deconvolution	400	425