CALICE MAPS Phone Meeting, 04/06/08
===================================
By phone: Jamie Crooks, Paul Dauncey, Anne-Marie Magnan, Owen Miller,
 Matt Noy, Marcel Stanitzki, Nigel Watson, John Wilson

Minutes: Paul


Minutes of previous meeting: No comments.


High-resistivity epitaxial layer: Marcel gave a summary of the issues
 of trying wafers with a high resistivity epitaxial layer during the
 next sensor fabrication; see slides on usual web page. There is clearly
 a significant likelihood that the hi-res epi sensors will not work but
 the cost of these is small and is not a big issue. The main point is
 that we can only submit one design for both the standard and hi-res epi
 options and the current design will need to be modified. There is some
 risk and some effort associated with this redesign.

 The design changes are to add deep p-well under the memory areas and
 to add a guard ring around the pixels. The guard ring is needed around
 each region separately as well as around the test structures. There is
 no room to add a guard ring between the two quadrants within a region.

 In terms of effort, Jamie estimated this is one day of his effort. It
 involves no schematic changes. He will not do this before the FDR, when
 it should be clear how much of the contingency is available.

 For risk, the guard ring will be connected to an external pin and so
 its potential can be controlled externally. There is a small risk that
 it could cause crosstalk between columns and rows but this is unlikely.
 Jamie considered the overall risk to be small, given the possible
 benefits. It was decided to go ahead with the redesign if there is
 sufficient contingency following the FDR.

 The hi-res epi wafers will be purchased from a different vendor than
 the foundry. If these are found to work well using a standard CMOS
 process then this implies there is less issue with multiple vendors
 and cost for a future large production. The order for these wafers
 has not yet been submitted; Renato had reported they had an 8-10 week
 lead time so it may be too late.
 
 It was decided to go with 12mu epitaxial thickness so that the changes
 from the standard wafers are minimised. The fact that the wafers come
 from a different source is already not ideal but nothing can be done
 about this.


Sensor 1.1 design: Jamie showed progress on the design for the next
 fabrication; see slides on usual web page. The rerouting of the diode
 contacts increases the input capacitance, and hence input noise, by 6%.

 With seven configuration bits, the n-well in the storage for these is
 closer to the edge of the deep p-well zone. It would be worth
 considering enlarging the deep p-well area by another 0.5mu or so, as
 long as it does not come too close to the diode.

 The separation of the comparator and monostable power meshes is 
 possible through reusing a power grid from the samplers.

 The reason for the significant change in S/N with the gain was not
 well understood. This implies the downstream noise (unaffected by the
 gain) is ~2.2mV and the input noise will change from ~3.3mV to ~2.3mV,
 i.e. so the input noise is no longer dominant. Increasing the gain
 resistor to increase the gain can only help with this issue so there is
 no danger in having "too large" a resistor. Jamie estimates the work to
 lay out the extra part of the resistor is around one day but there is
 some risk due to it becoming close to digitial circuitry. It was
 decided that Jamie should lay out the extra resistor part but be
 prepared to remove it if it is thought too risky at a later date.


Sensor 1.0 testing: Owen showed some results from his and Marcel's
 studies; see slides on the usual web page. He sees significant changes
 in the pedestal distributions when unmasking single pixels or whole
 columns. The crosstalk resulting from the active column seems to be
 in addition to the pedestal distribution of the single pixel itself.
 The laser results are encouraging although it was not understood why
 some channels seem never to see a falloff at higher thresholds. It
 would be useful to see if this is true at lower laser intensities also.

 Paul showed some results on crosstalk and on using the laser to see
 charge spread. The dependence on the laser focus is not understood.
 John suggested reducing the laser intensity to see if the focus
 dependence was reduced; if the effect is proportional to the electric
 field, rather than the intensity, then the shape should change
 depending on the pulse energy.

 Nigel reported that he had connected the trigger cables as described by
 Matt and the system no longer crashed as before. Paul will try to check
 the software tomorrow remotely to see where the remaining issues are.


Conferences: John will give a 15 minute talk at the ECFA meeting in
 Warsaw next week. He will give a practise run through at 11.00 on
 Fri 6 Jun by phone; Paul will distribute connection details for this.

 It was thought that John should show the physics simulation results
 shown previously at LCWS and EPS in 2007 but now showing the level of
 agreement of the charge spread simulation with the data measurements.
 This means we can claim the simulation is verified to the level of
 accuracy needed for the simulation results.

 John can also talk about the good performance of a single pixel and
 that both the spread of pedestals and the crosstalk have meant we do
 not now think the beam test data will yield useful results.


Next meeting: This will start at 9.30 on Wed 25 June; note unusual time.
 Before this, John will give a practise talk at 11.00 on Fri 6 Jun and
 the FDR will be at 10.00 on Fri 13 Jun.