56
3 SAMPA Chip Implementation
3.2.3.1 First Baseline Correction
Time projection chambers, due to their inherent construction, suffer from a few
signal effects that can be corrected through application of some filtering. A lowfrequency spurious signal can be present that is in the order of a kilohertz or less.
This signal perturbs the detector signal by shifting its baseline by an amount that is
close to constant, i.e. a change of less than one ADC count during an event. This
type of signal perturbation could, for instance, be caused by environmental changes
e.g. drift in temperature and voltage, environmental variations due to differences in
placement in the detector or by manufacturing or process variations.
Another type of signal perturbations is caused by systematic effects, like those
related to triggering of the detector, which affect the signal in terms of superimposed
noise patterns.
To cope with the first effect, a constant value can be used, in case the variation is
very small, or an IIR filter can be used if the change is larger. The IIR filter is further
referred to as the Variable Pedestal (VPD) filter.
The analogue front-end is constructed to add a small positive offset to the baseline
of the input signal of about 100 ADC counts. Since process variations, temperature
and supply voltage can impact, by shifting the baseline wither positively or negatively,
the added offset provides enough margin so that input baseline of the signal should
not appear below 0, which would render the data from channel unusable. Since this
shift varies from channel to channel and chip to chip, there is an option to set a
constant value per channel. The value to be used can be determined by temporarily
enabling one of the other baseline correction filters, which can provide the calculated
baseline through a registry entry. If the detector environment is stable and does not
suffer from other signal perturbation effects, the constant value subtraction might be
enough, which would enable the detector to save some power by disabling the other
filters.
The VPD is best used if the baseline is close to constant within one event, but
changes slightly in-between events, like what is shown in Fig. 3.10a. Since the VPD
filter can be configured so that it only updates its calculated baseline outside an event,
where there should be no detector signals, it will be effective for these cases. It will
acq #1
acq #2
acq #3
acq #4
acq #5
long-term
perturbation
(a) Example of long-term perturbations
useful signal
perturbation
(b) Example of gating perturbations
Fig. 3.10 Slow perturbation examples
3 SAMPA Chip Implementation
3.2.3.1 First Baseline Correction
Time projection chambers, due to their inherent construction, suffer from a few
signal effects that can be corrected through application of some filtering. A lowfrequency spurious signal can be present that is in the order of a kilohertz or less.
This signal perturbs the detector signal by shifting its baseline by an amount that is
close to constant, i.e. a change of less than one ADC count during an event. This
type of signal perturbation could, for instance, be caused by environmental changes
e.g. drift in temperature and voltage, environmental variations due to differences in
placement in the detector or by manufacturing or process variations.
Another type of signal perturbations is caused by systematic effects, like those
related to triggering of the detector, which affect the signal in terms of superimposed
noise patterns.
To cope with the first effect, a constant value can be used, in case the variation is
very small, or an IIR filter can be used if the change is larger. The IIR filter is further
referred to as the Variable Pedestal (VPD) filter.
The analogue front-end is constructed to add a small positive offset to the baseline
of the input signal of about 100 ADC counts. Since process variations, temperature
and supply voltage can impact, by shifting the baseline wither positively or negatively,
the added offset provides enough margin so that input baseline of the signal should
not appear below 0, which would render the data from channel unusable. Since this
shift varies from channel to channel and chip to chip, there is an option to set a
constant value per channel. The value to be used can be determined by temporarily
enabling one of the other baseline correction filters, which can provide the calculated
baseline through a registry entry. If the detector environment is stable and does not
suffer from other signal perturbation effects, the constant value subtraction might be
enough, which would enable the detector to save some power by disabling the other
filters.
The VPD is best used if the baseline is close to constant within one event, but
changes slightly in-between events, like what is shown in Fig. 3.10a. Since the VPD
filter can be configured so that it only updates its calculated baseline outside an event,
where there should be no detector signals, it will be effective for these cases. It will
acq #1
acq #2
acq #3
acq #4
acq #5
long-term
perturbation
(a) Example of long-term perturbations
useful signal
perturbation
(b) Example of gating perturbations
Fig. 3.10 Slow perturbation examples