PSD
EM User Manual
Version
1.1 (5/4/2000)
Author
:
J. Knödlseder
1.
SCOPE O
F
THE DOCUMENT
The
purpose of this document is to
provide
a user manual for the PSD EM subassembly. This manual should introduce the user
in the general functionality of the system, and should provide a reference
document that allows for a proper configuration and usage of the system.
2.
REFERENCE
DOCUMENTS
RD1 EM
software version 228 / V1.07
RD2 PSD
Software description (TBD)
RD3 SPI
INTERFACES SPECIFICATION, SPI-SI-0-1324-CNES, TBD
3.
GENERAL
FUNCTIONALITY
The
scope of this paragraph is the presentation of the general functionality of the
PSD subassembly in order to clarify the meaning of the PSD parameters.
3.1.
Multiplexing
The
PSD subassembly is equipped with 19 input channels (also called
mux
channels
),
each connected to one of the 19 fast detector output channels. Each of the 19
channels can be enabled or disabled individually in order to activate or to
suppress pulse-shape discrimination for individual detectors. On the EM, only
channels 0-6 (7 channels) are implemented. These channels correspond to
connectors J14 - J20.
Pulse
Shape Discrimination for background reduction is only meaningful for events
that are contained within one germanium detector. Therefore it has been decided
to have only one PSD system that treats all 19 detectors with an input
multiplexing. If signals occur on several PSD inputs within TBD, the event is
considered as a multi-detector event and is not treated by the PSD subassembly.
3.2.
Triggering
After
multiplexing, there is a hardware triggering system that is common to all
detector channels. The hardware trigger (or front-end trigger) is characterised
by three quantities: the
FRONT
END TRIGGER
LEVEL
(FET)
,
the
LOWER
LEVEL DISCRIMINATOR (LLD)
,
and the
TIME
WINDOW (TW)
.
All three quantities are configurable by telecommand. The major aim of the
hardware threshold is to trigger the PSD system on detector pulses and to
separate real event triggers from noise triggers. Consequently, the optimum
trigger parameters will depend on the noise in the system. The 3 front-end
trigger parameters are the most critical parameters of the PSD system since
noise triggers can easily exceed the expected event rates, leading directly to
an exceed in time-tag rates sent to DFEE.
The
following logic has been implemented for the 3 front-end trigger parameters:
- The
detector current must first rise above the FET level which aims in
discriminating real pulses against noise. The FET threshold should usually be
set just above the noise level.
- After
a FET, the detector current must rise above the LLD within a time interval
specified by TW.
A
forth quantity, the
UPPER
LEVEL DISCRIMINATOR (ULD)
is fixed and discriminates events that exceed a given threshold. In the EM,
this level roughly corresponds to TBD.
3.3.
Digitisation
After
triggering, the detector current is digitised by 4 interleaved ADCs at a
frequency of 25 MHz, leading to a time resolution of 10 ns. The ADCs have a
resolution of 10 Bits, but only the 9 most significant Bits are used for the
analysis. For the EM, the conversion function is roughly 7 digits / mV. With a
typical baseline of 45 digits, this corresponds to a maximum voltage of 67 mV
for a maximum ADC value of 511.
3.4.
Veto
The
PSD system may receive a veto signal from the DFEE in order to suppress event
processing. In the case that no DFEE cable is connected to the PSD, no veto is
active, i.e. all events are processed by the PSD.
The
Veto signal is hardwired in the PSD and cannot be influenced by configuration
commands.
3.5.
Gain
range
There
are two gain ranges available in the PSD. For the EM, these gain ranges
correspond to TBD. The gain range is common to all detectors and may be
selected by configuration command. The front-end trigger characteristics are
not influenced by the gain range selection.
3.6.
Event
identifier emission
The
PSD system communicates with the DFEE subassembly by two signals:
- A
time-tag signal that is emitted if a trigger occurred in the PSD system
- A
data bus for identification of valid events (DFEE ID)
If
a front-end trigger occurs in the PSD system without a Veto signal active, a
time-tag is emitted to the DFEE system. The PSD system is then integrating over
the measured pulse shape using a 9 Bit hardware integrator in order to estimate
the energy of the event. This energy estimate is then compared to a validity
range, specified for each detector individually by the
LOWER
ENERGY THRESHOLD (LET)
and the
UPPER
ENERGY THRESHOLD (UET)
.
The LET and UET are configurable for each detector by configuration commands.
If the pulse integral falls within the validity interval, an event identifier
is emitted to the DFEE.
The
relation between the hardware integrator 9 Bit value and the energy is rather
complex since the integral is composed of both the pulse area and the baseline
area. The baseline may vary as function of time and event frequency.
3.7.
HSL
Scientific
data is transmitted on the HSL from the PSD subassembly to the DPE subassembly.
The scientific data consists of:
- event
data
- curve
data
The
number of curves that are transmitted in a HSL data frame, and the frequency of
curve data transmission can be configured using configuration commands. The
transmission rates may differ between the operational mode and the calibration
or diagnostic mode.
Detector
pulses that were accumulated during cycle n of the 8 Hz HSL clock are sent to
the DPE during cycle n+1. A new cycle begins with the falling edge of the 8 Hz
clock. At this moment the PSD subassembly prepares the data frame in a FIFO
buffer. It may happen that at the moment of the 8 Hz falling edge not all
events of cycle n are yet processed, hence they would be lost since they are
not allowed to be sent in cycle n+2. For this reason, a post processing of 10
events at maximum is allowed after the falling edge of the 8 Hz clock occurs.
The FIFO buffer is only prepared after this postprocessing. The maximum number
of events to be postprocessed can be set by configuration commands.
3.8.
Pulse
shape discrimination
Pulse
shape discrimination is performed by comparing measured detector pulse shapes
to a library of pulse templates. As result of the comparison, an event may be
flagged as single site event (most significant Bit of the PSD word set to 0) or
as multiple site event (most significant Bit of the PSD word set to 1). Event
selection statistics are maintained in the housekeeping telemetry.
Library
templates can be uploaded using telecommands for each individual detector. The
libraries are stored in EEPROM. Two library sets may be uploaded for each
detector. The selection of the library set is done via configuration command.
Further, the number of time-steps used for the pulse shape analysis and the
number of library templates are also configurable.
3.9.
DSP32 software
The
PSD subassembly houses a DSP32C digital signal processor. This processor
handles all communications with the environment (LSL, HSL, DFEE) and performs
the scientific analysis of the digitised pulses shapes. The software in the PSD
is composed of two parts:
- The
functional software (eng.s) that handles all interfaces and the data
accumulation and preparation
- The
scientific software (science.s) that analyses the pulse shapes and handles the
library management
In
the EM, the functional software has version number 228 and the scientific
software has version number V1.07.
4.
COMMANDING
THE PSD
The
scope of this paragraph is to provide command descriptions which allow the
modification of configurable PSD parameters. The command structure is defined
in RD3.
4.1.
Enabling
/ disabling detector channels
Aim
:
Enable
or disable individual detector channels.
Command
sequence :
|
Configuration
command
|
Action
|
Affected
Bytes
in
command
|
|
01
Hex
|
Enables
/ Disables individual detectors
|
3-5
& 7-9
|
Verification
sequence :
|
Housekeeping
command
|
Action
|
Affected
Bytes
in
command
|
|
01
Hex
|
Read
detectors enable / disable status
|
3-5
& 7-9
|
Range
:
Flags
: 0 or 1
Default
:
All
detectors are enabled in all scientific modes (all Bits are set to 0).
Procedure
:
no
details
Notes
:
- Different
set of detector channels can be enabled or disabled for either operational /
calibration mode or for diagnostics mode. Operational / calibration mode
channel selections are determined by Byte 3-5 of the configuration command.
Diagnostics mode channel selections are determined by Byte 7-9 of the
configuration command (see RD3 for the detector - Bit correspondence).
- A
Bit set to
0
means
enable,
a Bit set to
1
means
disable
the detector channel.
4.2.
Specify
pulse trigger characteristics
Aim
:
Define
the pulse trigger characteristics to adapt to the actual noise level and
detector gain.
Command
sequence :
|
Configuration
command
|
Action
|
Affected
Bytes
in
command
|
|
01
Hex
|
Sets
FET, LLD, and TW
|
5-6
& 9-10
|
Verification
sequence :
|
Housekeeping
command
|
Action
|
Affected
Bytes
in
command
|
|
01
Hex
|
Reads
FET, LLD, and TW
|
5-6
& 9-10
|
Range
:
|
Parameter
|
Minimum
|
Maximum
|
Unit
|
Comment
|
|
FET
|
0
|
7
|
Dec
|
higher
FET value means lower FET threshold
|
|
LLD
|
0
|
7
|
Dec
|
lower
LLD value means lower LLD threshold
|
|
TW
|
0
|
7
|
Dec
|
time
window(ns) = 300 + TW * 160
|
Default
:
|
Parameter
|
Value
|
Unit
|
Comment
|
|
FET
|
2
|
Dec
|
|
|
LLD
|
4
|
Dec
|
corresponds
roughly to 630 keV
|
|
TW
|
1
|
Dec
|
corresponds
to 460 ns
|
Conversion
function / table :
FET:
The following conversion function has been established between lower pulse area
threshold (pulse area at half maximum) and FET (the pulse area is the net
integral over the digitised pulse, baseline subtracted; this is the most
precise measure of the event energy that is avaliable to the PSD subassembly):
|
channel
|
conversion
function
|
|
0
|
PA
= 900 - 658.6 * FET
|
|
1
|
TBD
|
|
2
|
TBD
|
|
3
|
TBD
|
|
4
|
TBD
|
|
5
|
TBD
|
|
6
|
TBD
|
Using
the relation between PA and energy as given in section 4.4, the following
conversion function result between lower energy threshold (in keV) and FET:
|
channel
|
conversion
function
|
|
0
|
Energy
(keV) = 102.37 - 84.82 * FET
|
|
1
|
TBD
|
|
2
|
TBD
|
|
3
|
TBD
|
|
4
|
TBD
|
|
5
|
TBD
|
|
6
|
TBD
|
LLD:
The following conversion function has been established between lower pulse area
threshold (pulse area at half maximum) and LLD:
|
channel
|
conversion
function
|
|
0
|
PA
= -500 + 500 * LLD
|
|
1
|
TBD
|
|
2
|
TBD
|
|
3
|
TBD
|
|
4
|
TBD
|
|
5
|
TBD
|
|
6
|
TBD
|
Using
the relation between PA and energy as given in section 4.4, the following
conversion function result between lower energy threshold (in keV) and LLD:
|
channel
|
conversion
function
|
Minimum
|
Maximum
|
|
0
|
Energy
(keV) = -491 + 212 * LLD
|
0
|
993
keV
|
|
1
|
TBD
|
TBD
|
TBD
|
|
2
|
TBD
|
TBD
|
TBD
|
|
3
|
TBD
|
TBD
|
TBD
|
|
4
|
TBD
|
TBD
|
TBD
|
|
5
|
TBD
|
TBD
|
TBD
|
|
6
|
TBD
|
TBD
|
TBD
|
Note
that there is no strict LLD / FET - energy relation since for a given energy
the peak height varies with the shape of the pulse; the above relation should
be considered as an average relation.
Procedure
:
In
the following a procedure is described which allows to set the front-end
characteristics for various noise conditions:
1.
Initialise:
Set FET=0, LLD=7, TW=1
- Determine
FET value:
The FET serves to discriminate real pulse triggers against noise triggers. For
this reason the FET threshold should lie above the noise level. To determine
the critical threshold, set the FET to its minimum (0) which corresponds to the
highest FET threshold. Monitor the registered pulse shapes in either
configuration or diagnostics mode. Rise the FET (from 0 - 7) until noise
triggers occur (visible either by a noisy curve display or an increase in the
jitter of the curve start time). Fix the FET to the highest level (= lowest
value) where no noise triggers occur (for security one might lower the FET 1-2
digits below this value).
- Determine
LLD value:
This value mainly sets the lower energy threshold (see above). Lower LLD (from
7 - 0) until the desired energy level was reached. If noise triggers occur,
lower FET again until they disappear.
- Set
TW value:
This value should be set between 0-1. For TW=0 less noise triggers should occur
than for TW=1, but scientific data could be biased for TW=0. Thus to reduce
noise triggers one could switch from TW=1 to TW=0.
Notes
:
FET
:
Increasing
the FET value
reduces
the FET threshold level (inverse relation)! Or in other words, increasing the
FET value makes the PSD more sensitive. If the FET threshold level is too low
(i.e. the FET value too high), noise triggers will appear.
LLD
:
The
relation of LLD value to detector current is a linear function with a possible
offset.
TW
:
The
minimum time window (TTW=0) corresponds to 300 ns, the maximum (TTW=7)
corresponds to 1560 ns. For TTW=7 the time window is ineffective since it
exceeds the pulse accumulation time.
4.3.
Specify
gain range
Aim
:
Switch
between normal and extended PSD gain range.
Command
sequence :
|
Configuration
command
|
Action
|
Affected
Bytes
in
command
|
|
01
Hex
|
Set
gain flag
|
6
& 10
|
Verification
sequence :
|
Housekeeping
command
|
Action
|
Affected
Bytes
in
command
|
|
01
Hex
|
Read
gain flag
|
6
& 10
|
Range
:
|
Flag
|
Gain
range
|
|
0
|
0
- 2 MeV (TBC)
|
|
1
|
0
- 6 MeV (TBC)
|
Default
:
Flag
= 0 (0-2 MeV gain range) (TBC)
Procedure
:
no
details
Notes
:
none
4.4.
Specify
energy thresholds
Aim
:
Specify
the energy range for each detector channel for which a event ID will be emitted
to the DFEE sub-system and which will be analysed by PSD subassembly.
Command
sequence :
|
Configuration
command
|
Action
|
Affected
Bytes
in
command
|
|
02
Hex
|
Set
lower energy thresholds (LET) for detectors 0-8
|
5-22
|
|
03
Hex
|
Set
lower energy thresholds (LET) for detectors 9-18
|
3-22
|
|
04
Hex
|
Set
upper energy thresholds (UET) for detectors 0-8
|
5-22
|
|
05
Hex
|
Set
upper energy thresholds (UET) for detectors 9-18
|
3-22
|
Verification
sequence :
|
Housekeeping
command
|
Action
|
Affected
Bytes
in
command
|
|
02
Hex
|
Read
lower energy thresholds (LET) for detectors 0-8
|
5-22
|
|
03
Hex
|
Read
lower energy thresholds (LET) for detectors 9-18
|
3-22
|
|
04
Hex
|
Read
upper energy thresholds (UET) for detectors 0-8
|
5-22
|
|
05
Hex
|
Read
upper energy thresholds (UET) for detectors 9-18
|
3-22
|
Range
:
|
Parameter
|
Minimum
|
Maximum
|
Unit
|
Comment
|
|
LET
|
0
|
511
|
Dec
|
values
> 511 will be interpreted as 511
|
|
UET
|
0
|
511
|
Dec
|
values
> 511 will be interpreted as 511
|
Default
:
|
Parameter
|
Value
|
Unit
|
Comment
|
|
LET
|
10
|
Dec
|
0
keV for EM
|
|
UET
|
511
|
Dec
|
~
8 MeV for EM
|
Conversion
function / table :
The
following conversion functions have been determined between the pulse area (PA)
and the LET and UET parameters for the different EM channels:
|
channel
|
conversion
function
|
|
0
|
PA
= -4969.5 + 49.231 * LET (UET)
|
|
1
|
TBD
|
|
2
|
TBD
|
|
3
|
TBD
|
|
4
|
TBD
|
|
5
|
TBD
|
|
6
|
TBD
|
The
following conversion functions have been determined between the pulse area (PA)
and the photon energy in keV for the different EM channels:
|
channel
|
conversion
function
|
|
0
|
PA
= 658.6 + 2.358 * Energy (keV)
|
|
1
|
PA
= 601.6 + 2.362 * Energy (keV)
|
|
2
|
PA
= 597.6 + 2.304 * Energy (keV)
|
|
3
|
TBD
|
|
4
|
TBD
|
|
5
|
TBD
|
|
6
|
TBD
|
The
typical energy resolution for the pulse area (PA) is between 2-3 % (1 sigma).
The conversion function is based on the analysis of calibration data taken with
a germanium detector and
60Co
and
228Th
sources. For the calibration only multiple site events were used since they
provide a better signal to noise ratio for gamma-ray lines. Note that for a
given energy, the pulse area slightly increases with increasing TTP while the
scatter of the pulse area decreases (see Fig. 1). Note that for a narrow TTP
range the pulse area energy resolution is between 1-2 % (1 sigma).
Fig.
1: Mean (left) and standard deviation (right) of the pulse area for 2615 keV
photons. The pulse area increases with increasing TTP while the scatter of the
pulse area diminishes with increasing TTP.
Based
on the above relations the following conversion functions have been determined
between the LET and UET parameters and the photon energy in keV for the
different EM channels:
|
channel
|
conversion
function
|
LET
= 10
|
UET
= 511
|
|
0
|
LET
(UET) = 114.32+ 0.0479 * Energy (keV)
|
0
keV
|
8281
keV
|
|
1
|
TBD
|
|
|
|
2
|
TBD
|
|
|
|
3
|
TBD
|
|
|
|
4
|
TBD
|
|
|
|
5
|
TBD
|
|
|
|
6
|
TBD
|
|
|
The
actual conversion functions on the SPI EM may deviate from this due to gain and
pulse baseline differences. In the ideal case, the conversion function should
be redetermined using the following procedure.
Procedure
:
In
order to determine the threshold conversion function, the following procedure
has been applied at CESR:
- Connect
the germanium detector fast output to one of the activated PSD channels
- Collect
~2000 HSL frames (~5 minutes) of
228Th
source data with the LET/UET set to narrow intervals (110-112, 120-122,
140-142, 180-182, 220-222, 240-242, 260-262, 280-282).
- Determine
the mean pulse area for all files and derive a (linear) relation between pulse
area and LET/UET value.
- Collect
~10000 HSL frames (~20 minutes) of
228Th
/
60Co
source data (
228Th
source very close to detector to dominate over room background,
60Co
at 1.5 m). The
228Th
source has 2 well separated strong gamma-ray lines at 583 keV and 2615 keV. The
583 keV is blended with the 511 keV line and is of limited use. The
60Co
source has to nearby gamma-ray lines ray 1173 keV and 1333 keV.
- Determine
the pulse area spectrum for single and multiple events. Gamma-ray lines should
be dominant in the multiple event spectrum.
- Fit
the 1173 keV, the 1333 keV and the 2615 keV lines in the multiple spectra and
determine the linear calibration relation between pulse area and gamma-ray
photon energy.
Eventually,
the 2.6 MeV
228Th
line may fall out of the PSD energy range for different gain adjustments. If
this happens than another source like
60Co
may be used in conjunction with
228Th.
Notes
:
none
4.5.
Specify
8 Hz post-processing
Aim
:
Specify
the number of events that are postprocessed after 8 Hz falling edge and before
HSL FIFO buffer preparation.
Command
sequence :
|
Configuration
command
|
Action
|
Affected
Bytes
in
command
|
|
02
Hex
|
Set
number of events for 8 Hz post processing
|
3
|
Verification
sequence :
|
Housekeeping
command
|
Action
|
Affected
Bytes
in
command
|
|
02
Hex
|
Read
number of events for 8 Hz post processing
|
3
|
Range
:
|
Parameter
|
Minimum
|
Maximum
|
Unit
|
Comment
|
|
postprocessed
events
|
0
|
10
|
Dec
|
0
means no post processing
|
Default
:
|
Parameter
|
Value
|
Unit
|
Comment
|
|
postprocessed
events
|
1
|
Dec
|
1
event is postprocessed
|
Procedure
:
no
details
Notes
:
Note
that this parameter has an impact on the HSL timing. Each post processed event
may add up to 1.25 ms processing time.
4.6.
Define
curve transmission rates
Aim
:
Specify
the curve transmission rates in the different scientific modes.
Command
sequence :
|
Configuration
command
|
Action
|
Affected
Bytes
in
command
|
|
0A
Hex
|
Set
curve transmission rates
|
3-6
|
Verification
sequence :
|
Housekeeping
command
|
Action
|
Affected
Bytes
in
command
|
|
0A
Hex
|
Read
curve transmission rates
|
3-6
|
Range
:
|
Parameter
|
Minimum
|
Maximum
|
Unit
|
Comment
|
|
8
Hz rate
|
0
|
255
|
Dec
|
period
in units of 125 ms
|
|
subrates
|
0
|
5
|
Dec
|
number
of shapes
|
Default
:
|
Parameter
|
Value
|
Unit
|
Comment
|
|
OP
mode 8 Hz rate
|
32
|
Dec
|
1
shape every 4 seconds (1 per 32 8 Hz clck)
|
|
OP
mode subrate
|
0
|
Dec
|
not
relevant
|
|
CALIB
& DIAG mode 8 Hz rate
|
0
|
Dec
|
use
subrate information
|
|
CALIB
& DIAG mode subrate
|
5
|
Dec
|
5
shapes per 8 Hz clock
|
Procedure
:
no
details
Notes
:
8
Hz rate has priority with respect to subrate information. The 8 Hz rate
specifies the rate at which 1 pulse shape is sent in the HSL data (i.e. 32
means send 1 pulse shape every 32 cycles = 4 seconds). Only if the 8 Hz rate is
set to 0, subrate information is used. The subrate specifies how many shapes
are sent in the HSL data (i.e. 5 means send 5 pulse shapes per HSL frame).
5.
DISCRIMINATION
CONTROL
5.1.
A/D
offset control settings
Aim
:
Adjust
the software gain and offset correction for the 4 interleaved ADCs. This
adjustment is needed to optimise the scientific analysis of the pulse shapes.
Command
sequence :
|
Configuration
command
|
Action
|
Affected
Bytes
in
command
|
|
06
Hex
|
Set
gain and offset adjustments
|
3-10
|
Verification
sequence :
|
Housekeeping
command
|
Action
|
Affected
Bytes
in
command
|
|
06
Hex
|
Read
gain and offset adjustments
|
3-10
|
Range
:
|
Parameter
|
Minimum
|
Maximum
|
Unit
|
Comment
|
|
gain
adjustment
|
-128
|
127
|
Dec
|
signed
8 Bit integer
|
|
offset
adjustment
|
-128
|
127
|
Dec
|
signed
8 Bit integer
|
Default
:
|
Parameter
|
Value
|
Unit
|
Comment
|
|
gain
adjustment
|
0
|
Dec
|
|
|
offset
adjustment
|
0
|
Dec
|
|
Conversion
function / table :
The
pulse shape is corrected in software using the formula
pcorrecti
= pulse * gain + offset
where
gain
= 1.0 + (gain adjustment value) * 0.0005
and
offset
= (offset adjustment value) * 0.05
Procedure
:
no
details
Notes
:
none
5.2.
Library
selection and control
Aim
:
Specify
the library set, the number of time, and the number of library templates used
for pulse shape discrimination.
Command
sequence :
|
Configuration
command
|
Action
|
Affected
Bytes
in
command
|
|
07
Hex
|
Set
library selection and control for detectors 0-6
|
3-28
|
|
08
Hex
|
Set
library selection and control for detectors 6-12
|
3-28
|
|
09
Hex
|
Set
library selection and control for detectors 13-18
|
3-26
|
Verification
sequence :
|
Configuration
command
|
Action
|
Affected
Bytes
in
command
|
|
07
Hex
|
Read
library selection and control for detectors 0-6
|
3-28
|
|
08
Hex
|
Read
library selection and control for detectors 6-12
|
3-28
|
|
09
Hex
|
Read
library selection and control for detectors 13-18
|
3-26
|
Range
:
|
Parameter
|
Minimum
|
Maximum
|
Unit
|
Comment
|
|
Library
set
|
0
|
1
|
Dec
|
only
two sets (0,1) fit into memory
|
|
Number
of time steps
|
6
|
64
|
Dec
|
|
|
Number
of templates
|
0
|
38
|
Dec
|
|
Default
:
|
Parameter
|
Value
|
Unit
|
Comment
|
|
Library
set
|
0
|
Dec
|
only
one library set is loaded in EEPROMs
|
|
Number
of time steps
|
64
|
Dec
|
|
|
Number
of templates
|
26
|
Dec
|
|
Procedure
:
no
details
Notes
:
The
number of time steps is required to reconstruct the information that is
compressed into the 16 Bits of the PSD information word.
5.3.
Library
upload
Aim
:
Library
templates may be uploaded into the PSD system and stored in EEPROM after
calibration of the system. Each detector has at maximum 2 attributed sets of
library templates.
Command
sequence :
|
Upload
command
|
Action
|
Affected
Bytes
in
command
|
|
0B
Hex
|
Send
library template definition and data items 0-3
|
3-20
|
|
0C
Hex
|
Send
data items 4-13
|
3-32
|
|
0D
Hex
|
Send
data items 14-23
|
3-32
|
|
0E
Hex
|
Send
data items 24-33
|
3-32
|
|
0F
Hex
|
Send
data items 34-43
|
3-32
|
|
10
Hex
|
Send
data items 44-53
|
3-32
|
|
11
Hex
|
Send
data items 54-63
|
3-32
|
Verification
sequence :
Read
HK0 (status) and make sure that no CRC error occurred.
Range
:
Bytes
3-6 specify the reference of the library template.
|
Parameter
|
Minimum
|
Maximum
|
Unit
|
Comment
|
|
Detector
selection
|
0
|
18
|
Dec
|
|
|
Curve
selection
|
0
|
37
|
Dec
|
special
value 255 loads library parameter block
|
|
Set
number
|
0
|
1
|
Dec
|
|
|
Data
items
|
1
|
64
|
Dec
|
usually
64
|
Default
:
not
applicable
Procedure
:
no
details
Notes
:
Normally,
the curve selection parameter is comprised between 0 and 37. However, if a
curve selection parameter of 255 is specified, the data items are interpreted
as library parameter information and not as a library template. Library
parameter information is needed for fine tuning of the scientific PSD software
and for setting the discrimination characteristics.
Library
template building using the EM as standalone system :
The
following recipes may be applied to build single site interaction pulse shape
template libraries using the EM without having access to the energy
information. For the calibration a 228Th source is used which leads to a
double-escape gamma-ray line at 1560 keV.
- Set
LLD=7 in order to suppress the low energy gamma-ray tail in the spectrum (see
calibration relations of section 4.2).
- Set
LET-UET around the 1560 keV energy range. For example: for EM channel 0 a
window around 190 may be used, i.e. LET=185 and UET=195 (see calibration
relations of section 4.4).
- Accumulate
data placing a
228Th
source as close as possible to the germanium detector.
- During
data analysis, select only those event with pulse areas close to 1560 keV. For
example: for EM channel 0 select only pulses with PA= [4300,4380] (see section
4.4). One may further select only those pulses that were flagged by the PSD as
single site interaction, which, however, introduces a bias with respect to the
actual active library. To improve the signal to noise ratio one could establish
a more refined relation between pulse area and energy as function of TTP value,
an select only those events which are around the double escape line.
6.
SOFTWARE
MAINTENANCE
The
PSD system is equipped with a DSP32C digital signal processing unit. In the EM,
the DSP32C software resides in EEPROM which can be reprogrammed via an IDE
interface (the EM PSD box has to be opened for this reprogramming). On
start-up, the EEPROM code is copied into RAM and executed there.
The
software is designed as modules which are connected via jump or call vectors.
Software can be maintained either by replacing code directly in RAM or by
adding software modules and redirection of the jump or call vectors. For this
purpose a specific software maintenance mode is foreseen which can be accessed
when the system is set to configuration mode. The DSP32C code for software
maintenance resides in EEPROM and consists of a reduced command interpreter.
When in software maintenance mode, the only commands allowed are memory upload,
memory dump, and status HK0 (TBC).
7.
Monitoring
the PSD
7.1.
Technical
housekeeping data
7.1.1.
Voltage
control
Aim
:
Collect
information about the PSD subassembly voltages.
Command
sequence :
|
Housekeeping
command
|
Action
|
Affected
Bytes
in
command
|
|
12
Hex
|
Read
voltages
|
3-10
|
Conversion
function / table :
TBW
Notes
:
none
7.1.2.
Temperature
control
Aim
:
Collect
information about the temperatures in the PSD subassembly.
Command
sequence :
|
Housekeeping
command
|
Action
|
Affected
Bytes
in
command
|
|
12
Hex
|
Read
temperatures
|
11-18
|
Conversion
function / table :
TBW
Notes
:
none
7.1.3.
Software control
Aim
:
Collect
information about actual software performance and status. Mainly used for
debugging. Also used for the reconstruction of the rate time information.
Command
sequence :
|
Housekeeping
command
|
Action
|
Affected
Bytes
in
command
|
|
12
Hex
|
Command
count, last received command code and identifier, last DFEE ID sent, 8 Hz counter
|
19-26
|
|
13
Hex
|
Events
and curves in buffers, error count since last power-on, last error type
|
3-8
|
|
1C
Hex
|
RAM
checking, analogue control
|
17-22
|
Conversion
function / table :
The
command count
is a roll-over counter of the number of commands received by the PSD EM,
including the housekeeping commands.
The
last
received command code
Byte contains the last command code that was received by the PSD EM,
excluding
housekeeping commands. The following table summarises possible return values:
|
Value
|
Meaning
|
|
43
Hex (C)
|
Configuration
command
|
|
44
Hex (D)
|
Memory
dump command
|
|
49
Hex (I)
|
Library
upload command
|
|
4C
Hex (L)
|
Memory
upload command
|
|
4D
Hex (M)
|
PSD
mode command
|
|
50
Hex (P)
|
Parameter
command
|
|
52
Hex (R)
|
Software
maintenance command
|
The
last
received command identifier
Byte
contains the last command identifier that was received by the PSD EM,
excluding
housekeeping commands. The possible return values depend on the command code :
|
Command
code
|
Possible
command identifiers
|
|
43
Hex (C)
|
01
- 0A Hex (command identifier)
|
|
44
Hex (D)
|
Most
significant Byte of start address
|
|
49
Hex (I)
|
0B
- 11 Hex (library upload identifier)
|
|
4C
Hex (L)
|
Most
significant Byte of start address
|
|
4D
Hex (M)
|
53
(S), 58 (X), 43 (C), 59 (Y), 44 (D) (all hex)
|
|
50
Hex (P)
|
parameter
identifier
|
|
52
Hex (R)
|
4D
(M), 52 (R)
|
The
last
HSL identifier sent to DFEE
contains the last 16 Bit identifier that was sent to the DFEE.
The
8Hz
counter
contains the actual value of the PSD internal 8 Hz roll-over counter. The least
3 significant Bits (Bits 5,6,7) of this counter are always 0 on the EM.
Notes
:
RAM
checking is not implemented on the EM. The corresponding Bytes should always be
0.
7.2.
Scientific
housekeeping data
7.2.1.
Channel rates
Aim
:
Collect
information about the detector channel rates, i.e. the event rates that are
passed to the analysis software (TBC).
Command
sequence :
|
Housekeeping
command
|
Action
|
Affected
Bytes
in
command
|
|
13
Hex
|
Compressed
channel rates for detectors 0-17
|
9-26
|
|
14
Hex
|
Compressed
channel rate for detector 18
|
3
|
Conversion
function / table :
The
channel rates for each detector are compressed into 8 Bits using the following
scheme:
|
7
|
6
|
5
|
4
|
3
|
2
|
1
|
0
|
|
exponent
|
mantisse
|
Decompression
leads to a 16 Bit value given by
value
= mantisse * 2 ^ (exponent + 4)
Units
for the decompressed value are events per 64 seconds.
Notes
:
none
7.2.2.
Selection statistics
Aim
:
Collect
information about the scientific selection statistics, i.e. the number of
events that have been identified as single interaction events or multiple
interaction events per detector.
Command
sequence :
|
Housekeeping
command
|
Action
|
Affected
Bytes
in
command
|
|
14
Hex
|
Compressed
selection statistics for detectors 0-10
|
5-26
|
|
15
Hex
|
Compressed
selection statistics for detectors 11-18
|
3-18
|
Conversion
function / table :
The
selection statistics for each detector are compressed into 8 Bits using the
scheme described in section 7.2.1. Decompression leads to a 16 Bit value. Units
for the decompressed value are events per 64 seconds.
Notes
:
none
7.2.3.
Rate history
Aim
:
Track
the history of global front-end LLD and ULD rates (no breakdown per detector).
The LLD rates are the number of pulses within 2 seconds that led to a front-end
trigger (i.e. pulses that exceeded the LLD within TW after a FET trigger). The
ULD rates are the number of pulses within two seconds that exceeded the ULD.
These parameters may be useful for deadtime estimation.
Command
sequence :
|
Housekeeping
command
|
Action
|
Affected
Bytes
in
command
|
|
15
Hex
|
Rate
history for 2 second intervals 1-2
|
19-26
|
|
16
Hex
|
Rate
history for 2 second intervals 3-8
|
3-26
|
|
17
Hex
|
Rate
history for 2 second intervals 9-14
|
3-26
|
|
18
Hex
|
Rate
history for 2 second intervals 15-20
|
3-26
|
|
19
Hex
|
Rate
history for 2 second intervals 21-26
|
3-26
|
|
1A
Hex
|
Rate
history for 2 second intervals 27-32
|
3-26
|
Conversion
function / table :
Uncompressed
16 Bit values. Units are triggers per 2 seconds.
Notes
:
none
7.2.4.
Library status
Aim
:
Collect
information about pulse baseline and noise running averages.
Command
sequence :
|
Housekeeping
command
|
Action
|
Affected
Bytes
in
command
|
|
1B
Hex
|
Read
running averages
|
3-26
|
|
1C
Hex
|
Read
running averages
|
3-16
|
Conversion
function / table :
The
running average is given by
ravg
= 0.25 * (running average value)
where
'running average value' is the unsigned 8 Bit integer value as read from
housekeeping.
Notes
:
Noise
running averages are not implemented on the EM. The content of all Bytes should
be 0.
