Dear all Below and attached you'll find the minutes of last weeks discussion on the status of the AFE8 and the possibility of using AFE8 for the AFE12. We reaffirmed the decision to purchase a total of 187 AFE8s for the central tracker. We will discuss release of the AFE8 and the order again February 14th. There was also a consensus that the AFE8 could be slightly modified to work effectively for the FPS. A decision on use of the AFE8 will need to be taken in April or May after the AFE12 prototype has been received. Jerry Blazey NIU 815-753-6480 FNAL 630-840-2342 CELL 630-745-0360 *********************************************************************= ******* ************************* Jan 31, 2001 AFE8/AFE12 Meeting Present: Marvin Johnson, Yildirim Mutaf, Fred Borcherding, Jerry Blazey (scribe), George Ginther, Volker Buescher, Paul Grannis, Harry Weerts Notice & Agenda: Dear All Could we meet next week Wednesday 1PM-3PM in the Doghouse to discuss 1) acceleration of AFE12 prototyping and production and/or 2) substitution of the AFE12 with AFE8's or modified AFE8's. with a goal of reaching consensus on how to proceed? Fred has already made a few suggestions regarding the AFE12s, which are attached below. George and Volker have made a suggestion for a use of the AFE8's and a modified AFE8, which is also attached. Of course the full AFE12 is most attractive from the physics point of view. An AFE8 type board with "smart biasing" to accommodate the chip sorting and the needs of the upstream and downstream FPS has some virtues and some shortcomings related to the signal magnitude. Perhaps Fred could lead the discussion on the AFE12 and perhaps Paul on substitution? Jerry Minutes and Action Items: Fred began with a report on the status of the AFE12. The daughter board prototype has been simplified with removal of the virtual (discriminator bits) SVX readout. The simplification reduced the daughter board chip count by nine, but does require modification of VHDL code. An ESE technician is rewriting the code. The resulting simplicity should be sufficient to solve real estate and power difficulties. John Anderson has been working about an hour a day on the AFE12. An estimated two weeks of calendar time is required for completion of the AFE12. Most of Anderson's time has been devoted to the higher priority AFE8 and the group concurred with this approach. A long discussion of the AFE8 status and the purchase order then ensued. There has been progress since last week's discussion including resolution of the large pedestal jumps. However the SVX emulated readout still has missing channels at 53MHz. It operates fine from 26 to about 40 MHz. The loss of the SVX readout would seriously compromise the utility of the preshowers. Operation at less than 53 MHz could increase deadtime in readout to L3. There is also a large spread in pedestal values for AFE8 SVX channels that would preclude zero suppression. If the full SVX complement for preshowers were to be readout, this would make additional load on the readout time. This week or next the LH and RH board noise tests will be conducted. Once the SVX problems are understood or the performance accepted and the noise tests completed the AFE8 boards will be ordered. Likely the AFE8 order will be with Revision C, this requires only a few adjustments for compatibility with the Mixer. The AFE8 order for the central region might occur in two parts: The current order for 146 will be released as soon as tests are completed. This is not likely to occur before the mid-February deadline. The order for the additional 41 boards to complete the central tracker (to allow operation of the full CFT axial and stereo before arrival of AFE12 boards) will be decoupled from the first order. This will permit modification of the card if any final changes are made. Two action items were formulated: (1) Fred was urged to submit the order and to start accumulating parts immediately to avoid any delay of the 41 cards. There was some hope that the 41 cards could be piggybacked on the 146-card order. (2) The group will revisit the situation February 14 to determine the method of acquisition and discuss the need for any revisions. Paul Grannis then discussed the advantages and requirements of AFE12 readout for the preshower detectors and the possibility of substituting AFE8 for the AFE12s. His slides can be found at http://d0server4/users/grannis/afe.ppt. The digitized output requires 80 MIP dynamic range to avoid saturation, and the preshower strips require at least 10% calibration to restore the calorimeter resolution. The trigger output requires two thresholds for efficient trigger at high and low transverse momentum. If the AFE8s were used for AFE12s the dual thresholds would be lost. This would result in a loss of low energy electron triggering but could still support the high transverse momentum program. However for the FPS, VLPCs of different wafers, hence different desired bias voltages, are installed in a given module of the cassettes, so operation with standard AFE8 boards would give many channels with non-optimum (or non-working) signals. Also with standard charge transfers to the SIFTs, the signals would be too large and saturate the dynamic range. The former problem could be fixed by making a simple modification to the AFE8 boards for FPS, providing flexible bias lines that permit operating VLPCs appropriately. Charge input to the SIFT can be reduced by two means: by reducing bias voltage, and by changing the standard input capacitor to the SIFT from 100 pF to 10 - 20 pF. With these changes, appropriate charge levels could be achieved that give reasonable threshold settings for the trigger, and acceptable dynamic range. The rather small signal available for minbias track calibration could be solved using special runs with high transfer gain from SIFT to SVX downloaded. The basic conclusion was that the AFE8=92s could work if there was a modification of the input signal magnitude and of the bias network were modified. Fred and Marvin felt these would be minimal changes. Marvin informed the group that the AFE12 prototypes could be prototyped by April and tested by May. (3) Tentatively the group decided that modified AFE8s would be an attractive alternative to the AFE12s should production of the AFE12s start later than May.