CALICE MAPS Meeting, RAL, 29/07/08
==================================
Present: Jamie Crooks, Paul Dauncey, Anne-Marie Magnan, Owen Miller,
 Marcel Stanitzki, Mike Tyndel, Nigel Watson, John Wilson

Minutes: Paul


Minutes of last meeting: No comments.

 There is no firm return date for the sensor yet; this will be known
 only when they actually start the fabrication process.

 John's EFCA talk has been added to the conferences web page, but it is
 still missing Giulio's 2007 TWEPP talk.


Conferences: PSD08 in Glasgow in early Sept requires the proceedings to
 be written by the start of the conference. This will be difficult as
 this is only four weeks from now and people are away over the summer.
 As we only have a poster contribution and no one is available to
 present this anyway, then it was decided to withdraw.

 Jamie thinks he is available for TWEPP in mid September.

 Paul and Marcel will attend the SiD meeting in Boulder. Paul will give
 a MAPS talk assuming we can get a slot, which Marcel thinks should not
 be a problem.

 There should be a MAPS presentation at the Manchester CALICE meeting
 the week before as it has been a while since we reported. Paul could
 easily give effective the same talk as for SiD but if anyone else would
 like to give it, that would also be fine.


Sensor 1.1: Jamie submitted this on time last week. There were a few
 final fixes and waivers for minor reported errors. As stated above, the
 exact return date is not yet known but is likely to be towards the end
 of Sept.

 He will document the new sensor over the next few weeks so Paul will
 set up a new web page for this.


Sensor 1.0: Jamie showed some slides on the laser measurements; see
 usual web page. He has understood the focus in some detail and the
 focus and timing measurements should be included in the paper. He also
 pointed out that the current sensor is 400mu rather than 300mu thick.

 It is not clear what causes the asymmetry shown in the timing plot. The
 sensor could have been misaligned to some extent, such that the
 position scan was not parallel to the pixel edge. However, a typical
 placement error of ~10mrad would only give ~1mu offset over the range
 of this scan. It was not clear if this would be enough. The timing plot
 shows a longer time of around 40ns when near the pixel centre compared
 with between either of the pixel pairs. This is when the charge has to
 travel an extra 16.5*(sqrt(2)-1) ~ 7mu, which implies a diffusion
 speed of ~0.2mu/ns. Hence, a 1mu difference would only give a ~5ns
 later time, which is much less than the effect seen.

 The next step is to repeat the position scan with the laser better
 focussed so as to see the response in more detail. This will require
 the system to automated and a summer student in RAL/PPD will start
 working on this in a week or so.

 Marcel showed some results from the laser and Fe55; see usual web page.
 The laser results show a long high-side tail which is not seen for
 single pixel laser runs at Imperial, so this may be an effect of
 illuminating multiple pixels, but it is not understood. It would be a
 very useful technique for a per-pixel gain measurement if it could be
 got to work well.

 The Fe55 shows a promising peak for single pixels. This needs to be
 cross-checked against runs without the source but it seems to be
 similar for all pixels so far. Of the few pixels shown, the peak above
 pedestal for quadrant 0 (y<84) seems a little lower than for quadrant 1
 (y>=84) but this needs to be confirmed with more statistics. The peak
 value is around 200TU. Assuming this corresponds to the 5.9keV Fe55
 photon converting within the diode (rather that the epitaxial layer) to
 1600e-, then this gives a calibration of 1TU = 8e-, which agrees with
 the value estimated by Jamie using the laser (see meeting on 29/02/08).
 A fit to the peak (or to the integral plot) is needed to extract the
 best estimate of the peak for each pixel. The pedestal is also needed;
 this can be done from the plots but Paul should also run his masked
 threshold scan on the sensor to measure the pedestals (and noise) when
 the source is removed. Marcel will contact him when this is free; Paul
 leaves for holiday for two weeks on Fri 1 Aug and it would be useful
 to do it before then. This would also give a comparison of Paul's and
 Owen's pedestal estimates, as this sensor has been trimmed by Owen.

 Assuming the calibration holds up, then the 6TU noise seen in the
 quadrant 1 pixels corresponds to ~50e-. This is a factor of two higher
 than the predicted noise at the comparator input. Whether the
 difference is due to the comparator itself, or to the circuit before
 the comparator not acting according to design, is not known.

 Paul showed a few plots on noise; see usual web page. The effect of the
 i12CompBias2 is not understood. The nominal setting for this DAC is a
 value of 830. Jamie suggested turning off various parts of the shaper
 circuit before the comparator to try to isolate what is going on. This
 may also help understand the source of the higher noise.

 Anne-Marie showed some simulation results aimed at the introductary
 section of the paper. It would be useful to add a plot for e.g. 10GeV
 showing how the resolution varies with tungsten thickness. One
 disturbing aspect of the results shown is that the apparent core shower
 density changes significantly depending on the histogram bin size and
 hence presumably is very peaked and not Gaussian-like. For 1x1mm2 bins,
 an average of ~60 tracks are seen per bin, while for 100x100mu2 bins,
 the average is more like ~12 tracks per bin, which is a density of
 ~1200/mm2. This would mean the central pixels for 100GeV showers would
 average ~3 MIPs, completely violating the assumption of low occupancy.
 The actual effect of this on the resolution is not clear; it only
 affects a few pixels in high energy showers of tens of thousands of
 hits so it may be negligible. 

 
Sensor 1.0 paper: Paul went through the paper outline to check on the
 status of the work; see the draft discussed today on the meetings page.
  General: Nigel had got the CALICE ROOT style file and sent it on
   during the meeting (see below).
  Sec 2: Nigel has the data for the machine background but needs to make
   the plot.
  Sec 3: Jamie had given Paul a set of plots during the meeting. He does
   not use latex but can send a flat file of text. The "good" sensor
   photo was used for IEEE and so could give copyright problems so he
   will organise another photo. The DAQ section should have a photo of
   the whole PCB board.
  Sec 4: Jamie has values for the circuit simulation which can be
   compared with the test pixel (for the samplers) and the bulk (for the
   shapers). The charge spread plots need to be remade to make them
   easier to understand (and also not hit copyright issues). The test
   pixel plots can be expanded with the laser results shown by Jamie.
   The bulk pixel section could possibly include the Fe55 per-pixel
   plots from Marcel. Owen will send Paul some plots equivalent to
   fig 20 to compare his method. The data plots for the charge spread
   also need to be remade; Paul will start the data scan when he leaves
   on holiday as it could take around one week.
  Sec 5: It was not clear if we would retain anything on twinning. The
   Fe55 results may be moved to the single pixel section. The Sr90 is
   one significant area where no result has been obtained in the past.
   The beam test results (and "no harm" results in Sec 6) still need to
   be checked with Jamie Ballin, who is on holiday. The cosmics section
   is also an area where no result has previously been obtained so it is
   not clear what results can be obtained.
  Sec 6: The resolution plots need to be redone with the 50e- noise
   value. Similarly, fig 11 should be remade and put here.

  [The ROOT style commands Nigel has got from CALICE are:
    gROOT->SetStyle("Plain");
    gStyle->SetPalette(1);
    gStyle->SetPadTickX(1);
    gStyle->SetPadTickY(1);
    gStyle->SetLabelFont(42,"xyz");
    gStyle->SetTitleFont(42);
    gStyle->SetTitleFont(42,"xyz");
    gStyle->SetStatFont(42);
    gROOT->ForceStyle();
  If there are any comments on these, please make them soon.]


Next meeting: This will be on Thu 21 Aug at 1pm at RAL.