60
3 SAMPA Chip Implementation
+
-L1
Data in 13
13
DS enable
DS enable
13
13
Data out
+
+
+
+
+
+
+
Z -1
Z -1
Z -1
Z -1
0
1
K1
-L2
-L3
-L4
K2
K3
K4
L30
+
-L1
Data in 13
13
DS enable
DS enable
13
13
Data out
+
+
+
+
+
+
Z -1
Z -1
Z -1
Z -1
0
1
K1
-L2
L3
K4
K2
-K3
-L4
L30
Fig. 3.11 Schematic overview of the IIR filter for the digital shaper
accumulated sum. The filter uses acceptance thresholds to exclude pulses from the
average calculation, as inclusion of pulses would introduce undershoot in the output
signal of the filter after a pulse, effectively introducing a dead time in the signal.
When the signal is outside an acceptance threshold, either the higher threshold or
the lower, the calculation of the average is frozen until the input signal again returns
inside the thresholds (see Fig. 3.12a).
Inside the acceptance window, the correction is done according to the following
equation:
x[n] = x[n − 1] + u[n] − u[n − M]
M = 2, 4 or 8
(3.4a)
y[n] = u[n] −
x[n]
M
M = 2, 4 or 8
(3.4b)
Where u[n] is the input sample, x[n] is the sum of the M last samples and y[n]
is the corrected sample value. When the current sample is inside the acceptance
window, the average that is subtracted from a given sample is the average of this
sample and the previous 1, 3 or 7 (depending on the configuration) samples that were
inside the thresholds.
When a fast variation in the input signal is present, e.g. a pulse, the input signal will
pass out of the acceptance window set by the thresholds and subsequently be excluded
from the calculation of the baseline. When outside the acceptance thresholds, the
value of the sample is corrected with the last updated baseline value from when the
previous last sample was inside the acceptance window.
A setting to additionally exclude a programmable number of samples before and
after any pulse that passes out of the exclusion thresholds is available. This might be
useful if a slight undershoot is always expected to follow a pulse, or if the thresholds
have been set rather high so there actually is a few samples that is part of the pulse
before and after it passes the thresholds.
3 SAMPA Chip Implementation
+
-L1
Data in 13
13
DS enable
DS enable
13
13
Data out
+
+
+
+
+
+
+
Z -1
Z -1
Z -1
Z -1
0
1
K1
-L2
-L3
-L4
K2
K3
K4
L30
+
-L1
Data in 13
13
DS enable
DS enable
13
13
Data out
+
+
+
+
+
+
Z -1
Z -1
Z -1
Z -1
0
1
K1
-L2
L3
K4
K2
-K3
-L4
L30
Fig. 3.11 Schematic overview of the IIR filter for the digital shaper
accumulated sum. The filter uses acceptance thresholds to exclude pulses from the
average calculation, as inclusion of pulses would introduce undershoot in the output
signal of the filter after a pulse, effectively introducing a dead time in the signal.
When the signal is outside an acceptance threshold, either the higher threshold or
the lower, the calculation of the average is frozen until the input signal again returns
inside the thresholds (see Fig. 3.12a).
Inside the acceptance window, the correction is done according to the following
equation:
x[n] = x[n − 1] + u[n] − u[n − M]
M = 2, 4 or 8
(3.4a)
y[n] = u[n] −
x[n]
M
M = 2, 4 or 8
(3.4b)
Where u[n] is the input sample, x[n] is the sum of the M last samples and y[n]
is the corrected sample value. When the current sample is inside the acceptance
window, the average that is subtracted from a given sample is the average of this
sample and the previous 1, 3 or 7 (depending on the configuration) samples that were
inside the thresholds.
When a fast variation in the input signal is present, e.g. a pulse, the input signal will
pass out of the acceptance window set by the thresholds and subsequently be excluded
from the calculation of the baseline. When outside the acceptance thresholds, the
value of the sample is corrected with the last updated baseline value from when the
previous last sample was inside the acceptance window.
A setting to additionally exclude a programmable number of samples before and
after any pulse that passes out of the exclusion thresholds is available. This might be
useful if a slight undershoot is always expected to follow a pulse, or if the thresholds
have been set rather high so there actually is a few samples that is part of the pulse
before and after it passes the thresholds.