62
3 SAMPA Chip Implementation
Fig. 3.13 Example of BC2 stuck baseline issue [9]
effective way of recovery, and as it requires that data is lost before it can recover,
an auto-reset functionality is implemented in the SAMPA. If the baseline is found
to be outside the thresholds for more than a given number of samples
6 the filter
either reset the baseline to a pre-programmed value configured through the registry,
or to the same baseline that the BC3 filter has calculated, as the BC3 filter does not
get stuck. The GdSP has a separate reset for the BC2 baseline controlled from the
configuration register, but this method has the same drawbacks as soft-resetting the
device.
As an extra protection, an absolute threshold for the baseline, can be configured.
3.2.3.5 Baseline Correction 3
Conventional linear filters like the BC1 and BC2 have been used in the past in particle
detector experiments and are useful for the task, but in some exceptional cases, they
can end up in a failure mode wherein the baseline of the signal shifts to outside the
thresholds of the filter and the filter stops to operate. In general, the filter has a max
slope that can be handled before it needs to be reset or the thresholds need to be
changed.
A new non-linear filter called BC3 [10] has been included, which does not have
any thresholds. This makes the baseline tracking free of dead areas and the need for
fine-tuning of configurations. The main idea behind this filter is to always follow the
baseline, but with a limited slope configurable per channel. The slope parameter can
be set individually for the up and down-slope.
The output signal y[n] is expressed as a function of the input u[n] as follow:
6 Programmable from 0 to 1023 with a 4 count granularity.
3 SAMPA Chip Implementation
Fig. 3.13 Example of BC2 stuck baseline issue [9]
effective way of recovery, and as it requires that data is lost before it can recover,
an auto-reset functionality is implemented in the SAMPA. If the baseline is found
to be outside the thresholds for more than a given number of samples
6 the filter
either reset the baseline to a pre-programmed value configured through the registry,
or to the same baseline that the BC3 filter has calculated, as the BC3 filter does not
get stuck. The GdSP has a separate reset for the BC2 baseline controlled from the
configuration register, but this method has the same drawbacks as soft-resetting the
device.
As an extra protection, an absolute threshold for the baseline, can be configured.
3.2.3.5 Baseline Correction 3
Conventional linear filters like the BC1 and BC2 have been used in the past in particle
detector experiments and are useful for the task, but in some exceptional cases, they
can end up in a failure mode wherein the baseline of the signal shifts to outside the
thresholds of the filter and the filter stops to operate. In general, the filter has a max
slope that can be handled before it needs to be reset or the thresholds need to be
changed.
A new non-linear filter called BC3 [10] has been included, which does not have
any thresholds. This makes the baseline tracking free of dead areas and the need for
fine-tuning of configurations. The main idea behind this filter is to always follow the
baseline, but with a limited slope configurable per channel. The slope parameter can
be set individually for the up and down-slope.
The output signal y[n] is expressed as a function of the input u[n] as follow:
6 Programmable from 0 to 1023 with a 4 count granularity.