Repository navigation
Expand file tree
/
Copy pathdaq.tex
More file actions
72 lines (66 loc) · 4.37 KB
/
Copy pathdaq.tex
File metadata and controls
72 lines (66 loc) · 4.37 KB
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
The DAQ will merge data to form events from the LArTPC,
photon detector and beam detector readouts using the
artDAQ data acquisition toolkit using a farm of commercial
computers connected with an Ethernet switch. ArtDAQ is
in use on several experiments at Fermilab. We are using it
on the 35t prototype, so we will have considerable
experience by the time of the CERN test.
The data collection for the CERN test will operate in a mode
similar to that forseen for the underground detectors. In order
to collect data from non-beam interactions such as proton decay
candidates or atmospheric neutrinos, data will be continuously
read in to the artDAQ data receiver nodes and processed through
the artDAQ system in quanta corresponding to time intervals fixed
from the time of the beginning of the run. These are then
transferred through the switch to a set of event building nodes
which work in parallel, each node receiving all the data from all
the detectors for the time intervals it is responsible for processing.
There will be 32 parallel incoming data streams from the LArTPCs
and 16 streams from the photon detectors.
There will be an additional stream from the trigger board (the same
board as built by Penn for the 35t test will be used) which will receive input
of the spill gate, warning of extraction, and pattern-unit bits from trigger
counters and other beamline instrumentation such as Cerenkov counters [Which
section are these described in?, should we refer to them from here?]. The
existing trigger board takes as input logical signals from the various sources,
and provides both a hardware trigger signal as well as a trigger word
specifying which triggers are active in any given event (beam spill, cosmic
veto, calibration trigger, random trigger, etc.). The trigger board is built
around a MicroZED development board carrying a Xilinx Zynq 7020 and is
therefore highly configurable, both through firmware (``process logic'') side
and software (``process system'') side. With small hardware modifications, the
board could take analog inputs if, for example, the signals from beam
instrumentation or veto counters are provided in a raw form from their
front-end.
Synchronisation across all the input sources is essential in order
that artDAQ can bring together the data from the input streams correctly for
processing by the event building nodes. The data receiver nodes will provide
overlap by repeating the data at the boundaries of the time intervals so
that a particle whose data spans two time intervals can be collected.
The time synchronisation is provided to the RTM back-module on the LArTPC
readout crates, to the SSP photon detector readout and to the trigger board from
a GPS based time synchrononisation distribution system originally designed
for the NOvA experiment. This system includes functionality to calibrate and
correct for the cable delays, and to send synchronisation control signals to
the readout at predetermined times.
The event building nodes will select time regions of interest within the time
intervals they are processing and form these into events to be written to
disk. The algorithms to select the events may be as simple as looking for
a trigger bit in the trigger board data stream, or may involve looking
for self-triggered events in the LArTPC data. An aggregation task, which
is part of artDAQ will handle the parallelized event building processes by
merging the output events into a single stream and writing them to disk.
To avoid oversized output data files, when a predetrmined file size is reached,
the aggregator will switch to writing to a new file. The collaboraion
requests to CERN, data links of sufficient bandwidth to transfer these files
from the CENF to the CERN data center, and from there to locations
worldwide for analysis.
Improved versions of the software systems which are being prototyped at the
35t test will available for the CERN test including (a) Run control which
controls and monitors the DAQ processes and allows run starts and stops to
be performed by the operator (b) online monitoring (c) slow control of
voltages and temperatures being used by the electronics (this may not be
comprehensive by the time of the CERN protoype, but we plan on prototyping
the readout of some of the quantities). The trigger board includes facilities
for generating calibration pulses and for identifying the event times of
the calibration events.