CALICE MAPS Meeting, RAL, 12/09/08 ================================== Present: Jamie Crooks, Paul Dauncey, Barnaby Levin, Michael Lynch, Owen Miller, Marcel Stanitzki, Renato Turchetta, Nigel Watson Minutes: Paul The two RAL summer students, Michael and Barnaby, were introduced. Minutes of previous meeting: No comments. Sensor 1.1 status: Jamie reported the sensors will be shipped from the foundry on Tue 23 Sep and should be at RAL by the end of that week. Proof-of-life tests should only take a few weeks if the sensors works and so we assume assume they will become available for the other test systems by mid or end Oct. The hi-res wafers should be shipped to the foundry next week. Because the masks exist, the fabrication on these wafers should take only 6-8 weeks. Hence, we could get the sensors back by mid to end Nov. We will only use a few of the 100 hi-res wafers purchased, so the rest would be available for SPIDER work if desired. The Python foundry test structure chip will also be available. Jamie has written a user manual for V1.1 and Paul will put this on the web. Jamie should check the existing V1.1 documentation to make sure everything is there. Sensor 1.0 tests: Jamie showed some results of gluing on the passivation side of the sensor; see document on usual web page. The passivation layer is easily strong enough mechanically. There is a concern of the glue causing problems to the sensor operation so Jamie is testing a sensor before and after gluing some material to its top side. Jamie also showed some slides of the laser charge diffusion data, replotted assuming symmetry of the pixel so as to give strips through the sensor; see usual web page. Barnaby showed some work by himself and Michael on laser studies at RAL; see slides on the usual web page. These results are all from sensor #6. Paul showed some crosschecks on the noise measurements; see slides on the usual web page. The RMS of a threshold scan seems to be a reliable measure. Hence, the absolute noise value and the pixel-to-pixel spread of the noise are both not understood. The noise values obtained are repeatable for the same sensor. Nigel suggested comparing the spread vs location for different sensors; this has not been done although Paul may have accurate data for two sensors available. Marcel showed the status of the Fe55 calibration; see slides on the usual web page. He is using the pedestals determined using Owen's method, which finds the 50% high-side falling edge point. This will be systematically above the true pedestal and so will underestimate the signal size. Marcel's Fe55 peaks do seem low so it was hoped this is the cause; otherwise it indicates the noise is even higher than we estimated up until now. The pedestal from a threshold scan without pixel pickup is more reliable and should be used instead. Conferences: Renato reported that Norbert Wermes had asked for slides to show in his overview talk at PSD08 in Glasgow and Renato had sent him some. His talk is on the web and is linked from the usual web page. We did not send a speaker to the LHeC conference last week. For the upcoming conferences: o IPRD08, Siena, 1-4 Oct, 25min talk; Marcel will give this. o IEEE, Dresden, 19-25 Oct, poster; Owen should have first choice but otherwise John Wilson or Rebecca Coath may be able to attend. o LCWS08, Chicago, 16-21 Nov, assumed talk; Anne-Marie will give this. Nigel has been asked to represent CALICE and give a SiW ECAL talk at the CLIC meeting at CERN, 14-17 Oct. This could cover both analogue and MAPS ECALs. It was not thought prudent to ask for a separate MAPS talk as we have little information on CLIC applications and performance. The first meeting in 2009 would be ACFA but this is not a priority so we should not bother asking for a talk there. Sensor 1.1 testing: Assuming it works as foreseen, the testing of the V1.1 sensor should be straightforward as we have most of the tests already in place and working. We should be ready to start testing by mid Oct. The priority is to check the basic performance and the changes from V1.0. These tests are: o Noise: This should be the same as seen now, as the resistor changes should not have increased the gain of the preShaper. The noise can be measured at Imperial and RAL/TD. o Gain: The absolute gain has to be measured at RAL/PPD using Fe55. The relative gain should be measured using the laser systems at RAL/PPD and at Imperial. o Pickup: It is important to see if pickup occurs and how many pixels need to be unmasked to induce it. These measurements can be done at Birmingham and Imperial. o Trim bits: The code to load six trim bits, rather than four, has been modified but cannot be tested. This should be done at Imperial so Paul can debug it. Further tests at a lower priority are: o Memory corruption: This can be studied through the "twinning" method at RAL/PPD and the hit override at Imperial. o Test pixels: These will be studied at RAL/TD. We need to aim for a beam test early in 2009. This would have two main aims; to measure the MIP efficiency vs threshold and to see real EM shower hits and so get data on hit densities at various depths within the showers. The first study would be done in "tracking" mode with no tungsten. Ideally, five sensors would be used, with only the central one having the threshold varied. There may be issues with triggering five sensors which Paul will discuss with Matt. At least two PMTs will be needed with at least one at the front and one at the back. The EM shower study will be done in "calorimeter" mode, with various amounts of tungsten in the stack. This requires two PMTs at the front. The preparations for the beam test need: o At least six sensors all fully characterised, i.e. with trims and bad channels masked. o The PMTs and triggering to be understood and reliable. o Monitoring software tools working. o The stack simulation set up and working. This should all happen through a system test at Birmingham which could start with the V1.0 sensors and hence work on this should start asap. The aim should be to have a fully reliable system by Dec. Nigel reported that currently the Birmingham system is missing triggers and so crashing due to timeouts. Paul and Matt will need to visit to help sort this out and this should ideally happen within the next few weeks. In addition to at least six standard sensors, we should plan to take at least one non-deep p-well sensor, one 5mu epitaxial layer sensor and possibly one hi-res sensor, if these are understood in time. All sensors should be potted. Next meeting: This will be on Mon 6 Oct, starting at 10am; note this is not the usual time. Renato has reserved the R76 meeting room.