1. Overview of EMU Software Rick Wilkinson

2. Slice Test DAQ We succeeded in using Slice Test DAQ code to take test beam data, combining chamber and trigger trackfinder data through an XDAQ Event Builder. We handled some big problems in 2003! – C++ VME interface and initialization – Chamber + trackfinder event building – XML configuration – Data unpacking and ROOT analysis For 2004, we need maintenance and extensions to all of the above, plus: – Run Control and User Interface – Optimize performance – Conditions – Monitoring – Integration with HCAL?

3. Raw Data and ORCA Right now, ORCA does not make anything that looks like raw data. Small summaries “digis” of each nonzero strip and wire channel are stored separately. Not grouped into SLINK readouts, so we can’t use ORCA to simulate rates easily. We want to run a level 3 ORCA filter unit inside our event builder, so that’s a good reason to do it right, and soon (Kreuzer, Wilkinson) Right now, we can read our raw data into ORCA –Our raw data format committed into ORCA CVS now. –Rewritten from test beam package to use bit fields, for easier interpretation, and maybe speed. Working on: Writing the correct raw format from ORCA Changing readout scheme, and associating readouts with the correct chamber Will require some ORCA changes, too. ORCA just isn’t written to store “blob” data!

4. Fitting Clusters We were forced to switch to a new, common minimizer, newly written by LCG. We use it to fit clusters of three strips to a Gatti function to find the hit position. So Tim (UCD) became a beta tester of the minimizer, and helped the LCG group track down problems. Is there a good analytical way to get the hit position?

5. Segment Building The way ORCA Muon trackfitting works, it first finds patterns using segments, then adds every point on the segment one by one into a Kalman filter fit. Not really robust. One bad point, or a point with a bad error, can mess up the whole track. Is it possible to do better? Doing it robustly might require building lots of track candidates, which means lots of CPU-expensive propagations through iron. Vladimir Palichik has made improvements to the segment building algorithm, to save TeV muons. Jason Mumford has an algorithm to simply ignore muon information if it hurts a good track from the inner tracker. We need an endcap muon trackfinding expert!

6. Other work since last meeting Tim switched the default ORCA geometry to use no ME4/2, and ganged strip readout in ME1/1. Rick simulated more neutron background, and updated his database of neutron background hits from ROOT to ASCII, and implemented it for the barrel muon, too.