_ TRIGGER


Previous:Simple Data Taking	Table of Contents	Next:Scalers

Chap 4: Trigger

4.1 First Level Trigger

The BLAST trigger consists primarily of programmable CAMAC modules from LeCroy. There are two CAMAC crates, one per sector. Communication with the CAMAC crates is through a Creative Electronic Systems CBD8201 branch driver. This is located in a VME crate in Rack 3, which also contains the scalers. The program ~/pro2003/trigger/trig loads settings to the various modules from a user-specified file. trig also provides a graphical user interface to examine, modify and save settings.
Detector signals are brought on RG58 cable to patch panels at the upstream end of each detector subframe, and from there to the D Tunnel. The high voltage cables also run from the D Tunnel through the patch panel to the detectors. The neutron detectors are an exception; the cables run directly between the detectors and the D Tunnel. The phototube signal cables are a total of 190 feet in length, whether running direct or through the patch panel.
The phototube signals are sent through splitter panels, where a passive circuit splits the signal in half; the prompt portion goes to the discriminators, while the other half goes through a passive delay line chip (Data Delay Devices 2214-500A) for 500ns of delay to the ADCs. The phototube signals for each Cerenkov box are added in a CAEN N407 analog adder, before going to the splitter panel.
The programmable trigger consists of discriminators, delays, logic modules to AND the ends of scintillator bars, ORs to fan together groups of scintillators, and MLUs (memory lookup units) to form triggers from user-defined patterns of input bits. Discriminator outputs are also sent to passive delays whose outputs go to scalers and to TDC stops. The discriminated signals from the end of each TOF (time of flight scintillator) are also sent to CAEN C561 mean timers. The mean time of each TOF hit is ORed together to provide a Common Strobe signal for the TS; this gives a stable start time for the event, independent of out-of-plane angle.
There is an MLU for each sector, with 16 input bits each; 6 output bits from each sector MLU go to the cross MLU (XMLU). A logic signal from the laser flasher goes to another XMLU input bit. Each MLU has a panel (above the CAMAC crate) to patch input signals to input bits, rather than making octopus cables. Programming the MLU hit patterns requires some care. A "1" for an input bit means there must be a signal there; a "0" means there must not be a signal there. If you don't care whether that particular detector (or group of detectors) fires, you must set an "X" for that input bit. Be sure you don't program multiple output patterns for the same input pattern; this will cause problems in the MLU. It's very easy to do when using X's, so check the input patterns carefully.
There are a number of standard trigger setting files in /home/daq/blast/pro2003/trigger suitable for loading with trig. See the README in that directory for an explanation.

Medium Large

4.2 Second Level Trigger

We use a second-level trigger to reject events with particles that fired the TOFs but left no information in the wire chambers. The wire chamber hit wire information is taken from the back of the 1877 TDCs using special boards. The boards put out logic signals in groups of 16 wires; these signals are cabled to a logic board in the back of each FASTBUS crate. The cabling and logic are such that a "Sector Wire Chamber Hit" signal is produced only if there is a hit in each of the inner, middle and outer chambers of the sector. The boards for each sector are cabled together to produce the OR of the two sectors, "WC Good". The AND of the two sectors is also available, although this is not appropriate for most of our triggers (e.g. an electron-neutron coincidence).

Individual groups of 16 wire inputs can be jumpered TRUE to take them out of the trigger, for example if the voltage to certain wires has to be turned off. Then the trigger for that sector effectively becomes a hit wire in each of two chambers; the third chamber is always true. A hot wire will produce the same effect.

The "WC Good" TTL signal is converted to NIM, and then ANDed in NIM logic with the Level 2 Start from the TS. The result is sent to the Level 2 Pass input to the TS; the NOT of the result is ANDed with the Wire Chamber TDC stop signal (it comes after any real wire chamber information) and sent to the Level 2 Fail input to the TS.

This second level trigger has cut our event rate by approximately a factor of 10, decreasing our dead time. Further improvements can be made by requiring the AND of the wire chambers in both sectors for events where a TOF fired in each sector (the so-called "e,e'p" events). This will be tested in the 2004 Spring running.