Pulse Height Spectroscopical Measurements
3.2.3 Fine Gain
The fine gain is to fine adjust the gain and therefore the energy to channel ratio. This gain
is just a mathematical factor used for calculation of the filter. The highest dynamic range
(ratio of low energy cutoff to maximum energy in spectrum) and best resolution can be
achieved when using a rather high coarse gain and a low (e. g. 0.5) fine gain. With higher
count rates (>10kcps) it is however recommended to stay with fine gain above 1 and to
reduce coarse gain.
The adjustment range is 0.5 to 6.5, default is 1. This adjustment range was chosen to be
compatible to MCA166 which has a range of 0.5 to 1.5, and to extend it also to larger
values which is useful for real high count rates.
3.2.4 Trigger Filter
The trigger filter is applied continuously to the incoming signal to check for events (voltage
steps). Also here are selections possible which represent trade off between time
resolution and sensitivity. The simplest trigger filter is the -1, 1 filter which just checks the
difference between subsequent incoming values and therefore performs a single
differentiation. Even more sensitive is the -1, 0, 1 filter which checks the difference
between incoming signals of 200ns distance.
The best pulse pair resolution (ability to distinguish between close following events) is
however achieved with a double differentiating filter as the 1, -2, 1. But double
differentiating increases the noise level and therefore this filter is not as sensitive. A
compromise is the 1, 0, -2, 0, 1 trigger filter, which has still a good time resolution but a
better sensitivity then the 1, -2, 1.
Table 11: Properties of different trigger filters. Dynamic range tested with GL2010R,
amplification 5*0.5, 4096 channels, shaping time 2µs flat 0.8µs
Trigger Filter
Pulse Pair
Resolution
Best Dynamic Range
Possible
Lower Energy Cutoff
in % of Full Scale
-1, 0, 1
860ns
1:625
0.16%
-1, 1
760ns
1:353
0.28%
1, 0, -2, 0, 1
330ns
1:325
0.30%
1, -2, 1
240ns
1:169
0.60%
Default is the 1, 0, -2, 0, 1 filter as a good compromise for both sensitivity and time
resolution. In case of a low count rate (<5kcps) and a required high dynamic range (low
energy part of the spectrum also important) the filter -1, 0, 1 is recommended. The 1, -2, 1
is best at high count rates and where the lower end of the spectrum is not really important.
The -1, 1 is more a bad compromise and just here because it is the shortest and needs
the least processor power.
29
