Minutes and Conclusions of the Beetle FE Workshop
Heidelberg, 07. Feb. 2002

Participants:
E. Jans, M. v. Beuzekom, S. Klous
A. Vollhardt, P. Sievers,
D. Baumeister, S. Loechner, C. Bauer, M. Schmelling, M. Agari, U. Trunk

1. Front End Issues
===================

 A. Setups and Procedures
  1. Use a probe or increase Ibuf, when measuring a FE chip. otherwise
     signals are distorted

    - when not using a probe then the current Ibuf has to be increased in
      order to avoid discrepancies. It has to be checked (with a probe) that the
      internal resistance of the buffer is sufficiently low (same pulse shape w.
      probe or cable).

    - it should be checked that measurements with the probe are
      insensitive to the setting for Ibuf. If a correlation is
      found, appropriate measures have to be taken.

  2. Subtract probe/scope noise when measuring noise

    - this procedure is subject to large uncertainties, since the
      values that are subtracted are of similar size. The uncertainty
      should be carefully estimated, since otherwise the results
      would suggest spurious uncertainties.

  3. On FE chips, pulse only one channel to avoid crosstalk and ringing.
     If necessary, connect unused channels to gnd or vdd.

    - whether neighboring channel effects are an issue for the operation
      in the experiment should be studied as well.

  4. Report gain in V/e (Volt/electron).

    - This is the natural unit, while a "MIP" is a sensor-specific quantity

  5. Report Parallel capacitance as Cp=Cload+Ccouple+Cpar with Cpar=3pF

 B. Results & Measurements
  1. Pulse shapes and gain Measurements of the 3 groups look consistent,
     since all discrepancies seem to arise from not obeying 1.A.1 to 1.A.3.
  2. Discrepancies in the noise values also seem to arise from 1.A.1 to
     1.A.3.
  3. The FE "2c" looks most promising. Measurements should concentrate
     on this one (until better settings and setups for the "2e" and "5" sets
     are available).
  4. Discrepancies between Beetle 1.1 and FE chips are not resolved. It
     has to be checked if blocking the internal bias lines resolves the
     problem (simulation & measurement on FE 1.0).

2. Beetle 1.1 Issues
====================

 1. "Sticky Charge": A remainder of the signals in the preceding readout
    burst. Its source has to be found by applying different trigger
    sequences. (S.L)
 2. Comparator results: Still waiting for results (H. Vk.)
 3. "Flat Baseline Settings" strongly depend on the individual chip.
    Comparison of the readout bursts in 128->1 and 32->4 mode can give a
    hint on the culprit.
 4. Fully differential output buffer: No report (H. Vk., U.T.)
 5. DAC calibration: Different slopes can originate from differences in
    the reference currents. Check if poorly controlled components (resistors)
    are used.
 6. There is a 20% gain variation across the channels. Check if they are
    randomly ordered and if they are correlated with the noise. (D.B.)

3. Conclusions
==============

 1. Measurements with improved setups should be continued by all 3 groups.
 2. Consistent measurement results should be available during LHCb week
    (18-22.02.02). Decisions on front end and circuit modification have to
    be drawn from them.
 3. If and when the Beetle has to be reviewed will also be decided during the
    18-22.02.02 LHCb week.
 4. Anticipated submission date is still 15.04.2002, but subject to revocation .

With best Regards

Ulrich