BNC PB-5 - Manual (Page 8)

7
used for low-amplitude tail pulses, the flat top portion will become longer as amplitude
decreases. Since some amount of flat top is desirable this may not be a problem
especially when using long tail pulses. However, low-level pulses below one or two
volts can have up to 700 ns flat top when attenuation is not used.
1.5.3 Linearity Measurements
In order to achieve the best statistical distribution and low drift, it is best to use the
fastest ramp time (30 seconds) and the maximum number of cycles (999). This
combination is convenient for long runs (about 8.3 hours). Depending upon the
frequency of the PB-5 and the number of channels under test, the ramp can be restarted
the following day and repeated for the number of days necessary to achieve the required
statistical accuracy (see Section 1.7).
1.5.4 CLAMP Mode
The amplitude of long exponential tail pulses may decrease with increasing rep-rates.
This happens when the duty cycle exceeds the time requirements for full baseline
recovery prior to the next pulse. To maintain the pulse amplitude, activate the CLAMP
Mode. To optimize performance in this mode the DELAY must be set greater than
3.0 µs.
1.6 TESTING A PREAMPLIFIER
Select PULSE SETTINGS from the main menu and select PULSE TOP TAIL from the
sub menu. The rise time will automatically be 50 ns and the tail time should be set long
compared to the decay time of the preamplifier (tail time is typically set to 500 µs or
more). Sometimes an experimenter may be concerned about using this pulse for the test
signal of a charge-sensitive preamplifier especially when solid-state detectors have
shorter rise times. However, the 50 ns pulse rise time is perfectly satisfactory for use in
testing linearity, stability, and resolution of the preamplifier. The following comments
are provided to clarify this matter.
It can be shown that the amount of injected charge from the pulser into a preamplifier is
given by Q = CV, where C is the coupling capacitance and V is the pulse amplitude.
The only restraint is that C be much smaller than the input capacitance of the
preamplifier. As long as the rise time of the injected pulse is much shorter than the
decay time-constant of the preamplifier, essentially all the charge will be collected. A
corollary of this is that the test pulser rise time need not be as short as the detector pulse
to simulate the same charge. These conditions are fulfilled by the Model PB-5 which
provides a rise time of 50 ns compared with the usual preamplifier decay time-constant
of 50 µs or more.
1.7 MEASURING DIFFERENTIAL LINEARITY IN A MCA
The model PB-5 has a built in ramp generator for measuring the differential linearity of
a multichannel analyzer (MCA). Differential nonlinearity (DNL) in a MCA describes
the change in relative width of one or more channels with respect to the average width of
all the channels. DNL can be determined by manually setting a pulse amplitude to both