CALICE MAPS Meeting, RAL, 21/08/08
==================================
Present: Jamie Crooks, Paul Dauncey, Anne-Marie Magnan, Marcel
 Stanitzki, John Wilson

Minutes: Paul


Minutes of last meeting: No comments.


Conferences: Marcel went through the list of conferences to the end of
 2008. These are:
  - LHeC, CERN, 1-3 Sep. There was not much enthusiasm for this so it is
    unlikely we will give a talk.
  - TWEPP, Naxos, 15-19 Sep. Jamie will give this talk.
  - IPRD08, Siena, 1-4 Oct. No one currently assigned.
  - IEEE, Dresden, 19-25 Oct. We need a volunteer for this; possibly
    Renato would be interested?
  - LCWS08/ALCPG, Chicago, 16-20 Nov. Some interest from Anne-Marie.

 Paul will give a status report at the SiD meeting in Boulder, 17-19 Sep
 and either Paul or Anne-Marie will give a similar talk at the CALICE
 collaboration meeting in Manchester, 8-10 Sep.


Sensor 1.1: The date this will be completed is 23 Sep, so that the
 sensors may be at RAL a few days following this. Jamie was not sure
 about the dates for the hi-res versions, but thinks these may be
 returned in early Dec.


Sensor 1.0: Jamie showed some slides on laser measurements; see usual
 web page. He sees a similar difference in the gains for the two shaper
 quadrants to Paul. Having calibrated the laser on the test pixel to
 the Fe55 signal size, the results from bulk pixels seem to indicate a
 much higher threshold than the direct Fe55 peaks on the bulk; this
 needs to be understood as it could indicate the laser is not giving
 the effect we think. The most critical measurement to be done is to
 find the actual calibration value from the laser to cross-check the
 Fe55 and Jamie will do this next.

 Jamie had looked into the i12CompBias2 effect shown in previous
 meetings. He thinks increasing the DAC setting would cut down the
 frequency response and so would decrease the noise, as observed. The
 fact that laser signal is still seen, although Fe55 hits are not, would
 then only be explainable if the Fe55 time response (which is
 effectively instantaneous) is significantly different from the laser
 response (which lasts for ~25ns as this was seen on the Imperial
 system). It is not clear if this would explain the observed effects.

 Jamie did 25k configuration loads and readbacks without observing a
 single bit error. Each configuration load transfers 168x168x5 bits,
 or 141kbits, so this was a total of 3.5Gbits. This implies an upper
 limit on the BER of 7x10^-10.

 Marcel showed some plots of threshold scans on single pixels; see 
 directory linked from usual web page. Many of these show long tails
 to high thresholds but these often start at ~250-300TU, which is not
 understood. The comparator saturation would not be expected to kick
 in until higher values. The important measurements here is to fit to
 the Fe55 peaks (or more likely the integral distribution) and get
 values for the gain per pixel for the two quadrants, to confirm the
 laser values. This requires correcting for the pedestals (see below).

 Paul showed some plots from Owen of the "50% edge" distribution and
 compared them to his pedestal distribution found doing a threshold scan
 with small numbers of channels enabled (42 per region), giving a "no
 crosstalk" scan. This was done for no trim, Owen's trim, and a trim
 using the pedestal value found. All plots are in directories linked
 from the usual web page. The last of these gave a much narrower
 distribution. It was decided that the pedestal determined from a "no
 crosstalk" scan (which takes a few hours) is probably the most accurate
 and should be used for all pedestal subtractions. However, for the Fe55
 measurements, as they will need a per channel correction, then the trim
 is not so important.

 Anne-Marie showed some "ideal case" comparisions of resolutions for
 analogue and digital ECALs, including varying the number of layers and
 the sampling fractions; see the usual web page. The resolution vs
 energy plot will be used for the paper. The more realistic resolution
 plots for the last section of the paper also need to be remade. The two
 for 10GeV (rather than 20GeV, for comparison with the "ideal case"
 plots) are straightforward but a further plot to do the resolution as
 a function of energy will require significant simulation effort,
 possibly via the Grid. Anne-Marie, Nigel and Marcel should agree how to
 handle this.


Next meeting: This will be on Fri 12 Sep, at 1pm as usual.