CALICE MAPS Meeting, RAL, 16/07/07 ================================== Present: Jamie Crooks, Paul Dauncey, Owen Miller, Vladimir Rajovic, Marcel Stanitzki, Claire Stephens, Renato Turchetta, Mike Tyndel, Nigel Watson, John Wilson Minutes: Paul Claire Stephens joined for her first meeting; she is a summer student working with Marcel for six weeks from today. Minutes and matters arising: Jamie has a new draft of the sensor users' manual which he will give to Paul to put on the web site soon. He has not yet updated the documentation from the reviews. There was no news on whether we can publically state the foundry used to produce the sensors yet. Jamie sent some information on power supplies to Birmingham. However, they have bench power supplies which can handle the ~6A required for running four sensors and so do not need to buy supplies for this round of the sensor fabrication. Paul found that there are no tungsten sheets available from the French groups. Nigel has got Fluka running but questioned whether it would be easier to use Mokka for the sensor study simulation as it would not require reimplementing Anne-Marie's digitisation. Vladimir will probably stay on for an extra three weeks rather than the two months mentioned in the last meeting. Sensor production: The wafers sent were not the splits required. Instead of six wafers of the usable ones and two wafers each of the other three splits, we were sent three wafers each of all four splits. However, when Renato chased this up, the foundry sent another three wafers of the usable sensors, giving us six of these plus extras of the others. The foundry does very basic tests of simple transistors and resistors so as to identify bad wafers, but the results for our wafers have not been released yet. Giulio wants to verify the epitaxial layer thickness directly. He has some of the sensors with no deep p-well layer and will try to cleanly snap one to see if the epitaxial layer is visible. Sensor simulation: Marcel reported from Giulio that he has started doing the full 3x3 pixel simulation about one week ago and estimates it will take three or four weeks to complete. He is thought to be doing the 21 steps as before, even though the sensor pixel is not quite four-fold symmetric. Sensor testing: Vladimir reported on the status of the sensor PCB; see slides on the usual web page. The PCBs eventually were fabricated by Exception PCB (manufacturer #2 in his talk). The PCBs will be 1mm thick and the bowing is expected to be minor and should not be a problem for the sensors if the amount of bow does not change after the sensor is mounted. The company was asked to send the PCBs to Paul at Imperial, although it is possible they were posted before this instruction was received, in which case they will go to Jamie at RAL. Vladimir will contact them to clarify this. The possibilities for assembly are to do (at least) one board at Imperial by the end of this week (Fri 20 Jul), or all three boards commerically by around Thu 26 Jul. Vladimir is due to return to Belgrade on Fri 27 Jul but hopes to get a visa and return a week later, i.e. Fri 3 Aug. Jamie said he would be able to test the PCBs in the week when Vladimir is away. It was decided to go with commericial assembly. Note, the company should not put on the zero ohm resistors used to set the power-up level of the threshold DAC; this can be done at RAL later. Paul asked when the other 12 (or more) of the sensor PCBs should be manufactured. The assumption is that the sensors cannot be removed and the PCBs reused with a new sensor, although Mike questioned whether this is really true. It is certainly a limitation if we have a low sensor yield and so would be worth investigating options for conductive glue which would allow a non-operative sensor to be removed. Paul reported for Matt Noy that the USB_DAQ adapter PCB has been returned and was being assembled at Imperial; see photo linked from the usual web page. Matt is back on Wed this week and the assembly should be finished before then. The extension memory card for the USB_DAQ is now being laid out by Vladimir. Jamie thinks the power-up state of the DACs on the sensor PCB should be close enough to the nominal ranges that tests can be done even without Matt's boards. However, the RAL NI card is not now available for programming the DACs via the SPI interface, in case this is needed. Marcel thought it might be possible to use an FPGA development board. Marcel reported that the laser-DAQ socket communication had been defined; the document is on the web under the MAPS DAQ pages. He has been working on some of the LabView problems related to getting this working. One question is how the laser interface could be tested. RAL will not permit a socket from outside the site to connect so the test has to be done internally. It should therefore be done when the new PCs are installed at RAL. Paul will need access to these (he has a csf account to get into RAL initially) and from there could test the socket communication. Marcel reported the PCs themselves had not yet arrived, although they should have been delivered today. He expects them any day and will get the OS installed quickly on each when they arrive. Paul would like to install the DAQ software before he goes on holiday for two weeks from the coming weekend; this may be tight. The mechanical structure for holding the cosmics stack is on hold until the exact dimensions of the PCBs, wire bond heights, etc, are measured. Marcel had got prices for tungsten plates. The most useful option seems to be plates of 100x100x5mm^2, which is ~1.4X0 per plate. With eight of these, a variable stack could be built up to ~11X0, which is enough to get to shower maximum in the DESY electron beam, which has a maximum energy of 6 GeV. The plates would be around £100 each, which is higher than estimated but affordable. There is a ~two month lead time so we should order these before the end of Aug. However, Konstantin thinks it might be possible to borrow some plates so this should be checked before ordering. Finances: Paul showed a spreadsheet of the state of the finances for MAPS so far; see usual web page. Black shows definite costs and red shows estimates. We have spend significantly more on the PCBs for the sensor readout than forseen, but the overspend is not at a level we have to be too concerned yet. Physics analysis: Marcel reported on PandoraPFA work; see slides on the usual web page. Anne-Marie will need to get involved to fix the assert which appears when both Marcel and JamieB run the digitisation. Marcel would like to investigate speeding up the simulation as it takes one day to do ~400 tt events. If doing a production on the web, then this would not be such an issue but it is inconvenient for debugging work as is being done now. Various alternatives to subdividing the pixels into 5x5mu^2 cells were discussed but it is possible these either will take similar times (as they store effectively the same information) or will be an approximation to what is needed. For the latter, we would need to have a careful study showing the approximations are reasonable before they could be used. A change was made some time ago to LCIO to allow SimCalorimeterHits to store separately the position and energy of every hit in the cell. This may give some improvement to the simulation speed and might be worth testing. Nigel showed some plots from Yoshi on results on linearity and resolution for 50x50mu^2 and 100x100mu^2 pixels, with and without various digitisation effects; see slides on the usual web page. The residual plots of non-linearity are hard to interpret as all points (including the ones showing non-linearity at high energies) are used for the fit and the plot is lin-log, distorting any straight line in the residuals. It would be good to remake this plot using only the lower energy points in the fit. It was also not clear if the large non-linearities found contradicted the results shown by Nigel around 1.5 years ago, where the effect seemed much smaller. Conferences: Nigel will give the MAPS talk at EPS on Thu 19 Jul. He will give a practise talk tomorrow (Tue 17 Jul) at 4pm; he will circulate the phone meeting details. It was decided that none of the more recent physics results since LCWS were mature enough to be shown. With only a 15 min talk including questions, this should not be too difficult to fill anyway. We have a talk at TWEPP (see web site for a link) in Prague in early Sept. Giulio is going and no one else put themselves forward so he was accepted as the speaker if he is prepared to do this. The two abstracts for IEEE were also accepted. Jamie and Marcel are able to go and again, with no other candidates, they were accepted to give these talks. Renato had no news from Vertex07. Detector collaborations: Phil Burrows is organising SiD-UK meetings and the next is a phone meeting at 10am tomorrow morning. Paul and Marcel will call in for this. There is also likely to be an SiD-UK general meeting in Sept. Mike is considering putting in a bid for engineering resources so the UK can contribute to the LoIs and beyond. It was considered important for such a bid to go in before the Programmatic Review is complete, so as to show there is significant interest in the ILC within the UK. The timescale for this review was not known. Mike considers something like 2SY to cover both concepts would be reasonable. This would require both SiD and GLDC to agree and put in a joint bid. This would be better than two separate ones but will require more coordination. It would also require the needs of the two groups, in terms of the type of engineering support required, to be similar. Mike is thinking mainly of mechanical engineering for the MAPS calorimeter in SiD although there is a well-developed mechanical design group in LDC already. It was clear from the recent LCUK-CC phone meeting that the UK needs to keep in contact with both detector concepts up to the EDRs in 2010. Within CALICE, we need both the MAPS and DAQ projects to be represented in both. RAL are likely to be the only group in the UK who join both and Marcel certainly has stronger ties to SiD. The MAPS therefore needs at least one of Birmingham or Imperial to join GLDC. Nigel indicated that Birmingham would probably go this way. Next meeting: This will be at 1pm on Tue 7 Aug in a RAL/PPD meeting room, as usual.