From rutherfo@physics.arizona.edu Tue Jun 1 14:37:48 2004 Date: Thu, 27 May 2004 14:37:30 +0200 (CEST) From: John P. RUTHERFOORD To: atlas-larg-h6-comb@cern.ch Subject: CBT-EC2 "Update" 27 May 2004 CBT-EC2 update J.Rutherfoord The ATLAS EMEC/HEC/FCal combined test beam run (CBT-EC2), sometimes called "crack studies", is in progress in the H6 beam line in the North Area. Many people are now at CERN but we are not in a smooth data-taking mode yet. This report is an update on current progress and an attempt to identify problems. Because it takes some time to collect information, some of this may already be out-of-date as it is being written. Also, since I am not an expert on all the items covered here, it is possible that I have not understood what I was told. So please be aware that this is an attempt to summarize a lot of information and that it is subject to errors, hopefully not too egregious. The H1 cryostat was filled with liquid argon on Tuesday (25 May) after cooling over the weekend. Attempts to determine the liquid argon purity have not yet been successful. The same can be said of the EMB cryostat in the H8 line. For the H1 cryostat it appears that a HV line is severed within the cryostat. If this is true, then we won't have any purity information for this run. We may be able to infer something from the H8 line if that purity monitor can be made to work ok. Right now the noise levels are much higher than anticipated. HV has been applied to the modules. There have been some hardware and software problems with the ISEG HV system so it is hard to be certain of the status of the modules. But it appears there are a few shorts (nothing alarming) and a few questionable channels. The HV system software experts from Milano are looking at the deficiencies and the malfunctioning modules are being replaced. One problem is that we occasionally find the HV is turned off, perhaps because something tripped. There are both hardware and software trips for various reasons. The software is supposed to report a trip and its cause. But, so far, these "trips", if that is what is happening, are not understood. The FCal Mod 0 calibration pulser board has been modified by the CERN electronics experts (Claude-Pierre Marin and Claude Millerin). They soldered a 150 pF capacitor across the 50 Ohm output resistor on each of the 128 pulser channels. This will filter out the high frequency components coming from the rising edge of the pulse, putting the resulting pulse within the frequency response of the preamp. This should solve the non-linearity problem seen in the FCal Combined TestBeam Run of 2003. Beam returned to the North Area late Tuesday afternoon. A.Savin was able to time most of the trigger right away and has now completed the trigger and, in addition, has timed in the ADCs, TDCs, latches, and beam chambers. This includes the Warm Tail Catcher (WTC). The experts will have to check that the beam chambers are working properly. And the HV will have to be adjusted on the WTC phototubes, and the timing checked again. The ITEP BPCs are working fine and can be used to tune the beam. The MWPCs (JINR responsibility) look good too but the experts will have to verify this. The DAQ is reading out all the beam-line devices, including everything read out via CAMAC. This includes the "trigger" latches, WTC, and beam chambers. At this point the data goes as far as pcfcal02 where Petr Gorbounov writes out files that mimic the byte-stream format of ATLAS. This allows the offline software developers to exercise their monitoring routines. It remains to transfer this data to the ROS computer and merge it with the calorimeter data. But the calorimeter data is not being read out yet. So far five time samples from four of the Front End Boards (FEBs) is being read into the Read Out Driver (ROD). Then the ROD data is being read to the Filar (PCI) interface in the ROS computer. This data can be read by a diagnostic program but not yet by the TDAQ software. J.Huber hopes to have this working by Friday (28 May). It is not yet clear when the ROD data and the Beam data will be merged in the ROS. The ROD firmware is not yet doing what we require for the test beam run. Jacques Colas has promised that the ROD firmware experts will be available to us starting on Tuesday 1 June and that updated firmware will be available for use to test. At present it is not clear how much trouble there will be in getting working 1) the merging of the data in the ROS computer, 2) the firmware in the ROD to handle more time samples and multiple gains (there's some confusion on my part whether multiple gains work with the old firmware), and 3) the additional PU which has not yet been installed in the ROD. We need this additional PU to read out the other four FEBs. Leonid could get one whenever we are ready for it. There are probably other issues which 1) I don't understand and 2) which none of us have yet foreseen. It will be crucial that we have all the experts here at CERN while we try to get the many facets of the DAQ running properly. There are many detailed features which must be implemented and tested before we can be confident that the DAQ is delivering data that is rock-solid and unambiguous.