7.3.5 AMPLITUDE LINEARITY
The amplitude calibration measurements were performed relative to an LeCroy
1882F 12-bit charge ADC residing in FASTBUS. Prior to all calibration
measurements a pedestal run (for the ADC and for the DSCs) and a sampling-speed
calibration run were done. In order to test the dynamic behavior of the DSC,
the calibration were performed with signals of very short rise and fall times
and a FWHM of 40 ns. The signal has been split and recorded simultaneously
by the FB ADC and DSCs. The recorded waveform is integrated over 100 ns
(DSC integral value), the same gate length used for the FB ADC's.
An initial calibration measurement was done by recording a signal with an
amplitude of 100 mV (FWHM of about 40 ns) with both, the DSC and the
FB ADC. The DSC integral value for each event is found by integrating the
signal over a range of 100 ns. The mean and rms-value of the DSC integral
ad and the FB ADC aref for 500 events
are:
DSC integral: ad =1903±17.3 ADC counts·ns (0.9%) (
7.8)
FB ADC: aref =523±4.8 ADC counts (0.9%) (7.9)
The value for the noise of the 100 mV signal (0.9%=0.9 mV (rms))
found by the DSC is in agreement with the FB ADC information. The noise is
coming from pick-up on the signal lines in the counting house. It exceeds the
typical noise of the FB ADC of 0.8 ADC counts (rms).
The amplitude calibration was performed with an "analog" NIM-signal attenuated
from -300 mV to -25 mV, which corresponds to the dynamic range of
signals from the PM's of the CsI-crystals. For each voltage 500 events have
been recorded and the mean- and rms-values are computed. The DSC integral
ad value is plotted as a function of the FB ADC value
aref and fitted with a straight line (Fig 7.12).
Figure
7.12:
The DSC integral value ad displayed as a function of the FB ADC
aref fitted with a straight line (left). Deviations from the linear
fit for the 300 mV input range (right).
The linearity la relative to the FB ADC is again defined as
maximum deviation from the linear fit. It is 25ADC counts·ns, which
corresponds according to Eq. (7.8) to 1.3 mV. This value should be considered
as an upper limit on the linearity, since the FB ADC values are very sensitive
to baseline shifts due to the attenuation of the NIM-pulse. In addition it
includes the non-linearity of the FB ADC[17]
±%
. As value for the linearity of the DSC one obtains
la<±1.3 mV (±0.43% of 300 mV). (
7.10)
The linearity la relative to the FB ADC was also measured by
applying DC voltages between 10 and 300 mV. The data are analyzed as explained
above and the resulting linearity agrees with the value obtained from the
"dynamic" calibration in eq. 7.10.
[17] The integral non-linearity of the 1882F ADC
is specified as <(0.25 of reading + 2 cnts). At 1500 ADC counts (range of
the linearity measurement) this corresponds to 5.75 ADC counts. Together with a
sensitivity of 50fC/ADC count, a input signal width of 40ns and the 50 W
resistance this leads to an equivalent input non-linearity of 0.4mV