CALICE MAPS Meeting, RAL, 25/06/08 ================================== Present: Rebecca Coath, Jamie Crooks, Paul Dauncey, Anne-Marie Magnan, Owen Miller, Renato Turchetta, Marcel Stanitzki, Mike Tyndel, Nigel Watson, John Wilson Minutes: Paul Minutes of previous meeting: No corrections. Jamie introduced Rebecca Coath as it was her first meeting. She is working with Jamie on the updates for the sensor V1.1. Information from foundry: Renato reported that he and Jamie had discussed the high-resistivity production with the foundry. The foundry will put their standard test structures on all the wafers, including the hi-res ones, so these will be tested by them to characterise the process. These are added between the sensors we require and do not cost us anything. The foundry also recommends a split of at least two wafers of 5mu epitaxial layer, non-deep p-well as this is their standard characterised process. This will act as a control to allow them to check the overall production. This would mean the total list of six splits we would have would be: 6 wafers of 12mu epitaxial, deep p-well 3 wafers of 12mu epitaxial, non-deep p-well 3 wafers of 5mu epitaxial, deep p-well 3 wafers of 12mu epitaxial, deep p-well, hi-res 3 wafers of 18mu epitaxial, deep p-well, hi-res 2 wafers of 5mu epitaxial, non-deep p-well which implies a significant testing load for us. The foundry also suggest we add a stripped down version of their full reticle test device, which would occupy one seat. This contains a version of many possible components in the design rules. This was thought useful for the future and would be produced for every split. The cost of this will be picked up by CFI. The effort to test these devices will be significant and would have to be done outside the current project; again this could be done through CFI. An alternative would be to assign a technical student to do these tests; Konstantin had a student allocated to him before he left and this person might be able to work on this testing. Marcel has two summer students in Aug and Sep this year but the devices will not be returned in time for them to contribute. There was no new information about the requirement for a guard ring around the outside of the sensor, although Renato thinks this should not be needed. Renato mentioned that we should be aware of latch-up, which is where a particle causes a very high current to flow in one diode, potentially damaging the sensor permanently. This was not seen in the beam test but the regulators may have prevented such a current from tripping the power supplies, so it may not have caused damage either. The alpha source would have caused the biggest ionisation and no odd effects were seen from that either. Sensor V1.1 design: Jamie showed some slides on progress with the design since the FDR; see the usual web page. He has implemented the majority of the changes from the FDR and is checking these. The foundry have moved the submission date to be a week later so it is now Mon 21 Jul, although Jamie intends to finish two weeks before this to allow time for an early submission, including the foundry checks. The assumption in the FDR that the true V1.0 resistor value was 2.7Mohm is now thought to be wrong and the resistor is probably indeed close to the design value of 4MOhm. The former value did not include the "DeltaW" line width correction. Hence, if the gain is indeed low, it is not due to the resistor being smaller than expected directly. The DeltaW correction itself could be wrong but it would have to be very different from the expectation to explain the possible discrepancy by itself. The resistor track width has now been increased from 0.18mu to 0.26mu which should reduce the sensitivity to DeltaW a little. Jamie will include a transistor test structure, which could be used in future as an alternative resistor to polysilicon, in the test structures for the V1.1 sensor. He is including several polysilicon resistor test structures also. These can be bonded to existing pads on the PCB (if the resistors to the DACs are removed) and so can be measured on bonded sensors. The total number of I/O pads for test structures will be 32. (the "NWell-NWell isolation" takes 5 rather than 3.) The two new 3x3 arrays of test pixels are either side of the original test pixel location, so the centres of the active test pixels are +/-100mu displaced. This should still be comfortably within the 1mm PCB hole, so they should still be accessible by the laser. The testing document for the PDR is still outstanding; the RAL section has now been completed but it requires the Birmingham section to be updated. Mike also suggested it should contain a list of the standard tests which should be done, so that it is clear what the order of them is. It should also specify which of the splits should be examined as a priority. Sensor testing: Paul showed some slides on various sensor tests; see usual web page. Owen pointed out that although the trim range quoted by Paul was ~80TU, he has seen values between 50 and 150TU on other sensors. This may be due to Jamie having changed the external resistors for these boards; Owen and Paul should report a list of which board has which range to Jamie to check this. Owen also reported he had seen that the crosstalk appeared to be less severe for the samplers than the shapers. The cause is thought to be due to the shared power supply of the monostables and comparators and the samplers have a separate power supply for the second monostable, so this observation could be consistent with this assumption. Nigel reported the cosmics system is now running, after some effort on finding grounding problems, etc. The trim values are being finalised and the running should be able to start by the end of the week. Nigel pointed out that there is some error or warning printout which makes the log files very large and Paul will look into this. In Marcel's absence, Owen had no news on any analysis of the Fe55 data which had been taken some weeks ago. Sensor V1 paper: Paul went through the outline paper he had circulated previously. Renato thought it would be better to split into two or even three papers, separating the sensor design and characterisation from the results and physics studies. Also, it was not clear if the design bugs should be discussed unless they are directly relevant to the results being shown. Mike thought it would be good to have a more detailed writeup of the results, beyond the detail required for a journal paper. The outline could act as such a writeup, with the paper(s) being produced from stripping down the results. The writeup would be an internal document and could be a CALICE Note, for example. The overall timescale for the paper is to have an (at least semi) final draft in time to hand to the PPRP in Sep. This means we should aim to have it close to finished by the end of Aug. The next step is to find out what exists for each section, so we can identify people to produce the missing items. To gather existing plots, etc, then the following people were assigned to each section: Sec 2: Anne-Marie Sec 3: Jamie, Renato Sec 4: Paul, Marcel Sec 5: Paul Sec 6: Nigel, Anne-Marie These people should report back at the next meeting, given that time is tight. Next meeting: This will be back to the usual time of 1pm and will be on Wed 16 July.