Minutes - ICD meeting Thursday July 2, 1998 10-noon, 3-4:30pm Ugh at UTA (with FNAL and LT by phone) send additions/corrections to Elizabeth Gallas In attendance: Lee/Dick/Craig in Arlington. Cliff, Kathleen on the phone from LT. EG on the phone from FNAL. Others at UTA: JL, AW, MS, TE, VR, KD, PD, anyone else ? Collaboration workshop is next week at FNAL: LS, Craig?, EG and JL will attend. Dick can't make it. AW and KD at SUSY98 representing D0. ------------------------------------------------------------------------- Next meeting (in 2 weeks): 1:30pm Thursday 16-July-1998 AW (and KD?) will be in Europe, back the following week. ------------------------------------------------------------------------- ACTION ITEMS: JIA LI must bring the following information to FNAL for the collaboration meeting (meetings with BNL FPS and/or FNAL engineering group) 1. calculate the clearance required in (R,z) to remove an ICD tile module for the FPS group. 2. latest ICD fiber cable routing plan overlayed on the FPS cable routing (use ICD cable drawing with 4 cables per tile module). Get the current FPS drawing from the URL below. 3. drawings showing explicitly the location of the tile module support brackets. be ready for alignment/survey discussion about where you want the weld studs (alignment of modules to 1mm) EG has written to Jack and Herman to see when they are available for this meeting. EG and JL will meet with FPS people tuesday 7-July-1998. KAUSHIK/MARK/VICTOR We wish we had time but it is clear we cannot evaluate/setup 5 different ways to measure the yield (we only had time to discuss 2 methods sketched below). KD stated the first method is probably the best opinion so let's go from there. 1. Kaushik should clarify the method with the formula for the yield as a function of the signal means and widths. Send it to the group by email. 2. Do we measure the yield from only the largest tile and the smallest tile as well ? 3. Do we use one or two fibers per tile (reading out both ends). Two fibers more accurately reflect the current tile module design. 4. Outer edge of the 1/2" tile should be unwrapped and painted with the Bicron reflective paint which should improve the yield from the larger tiles. While the tile is unwrapped, black marker should be applied along the scintillator bridging the isolation groove (this should reduce cross talk between tiles according to CDF). Bicron paint should also be applied to any epoxy shrinkage holes in the isolation grooves, though for future prototypes, EG will fill these by hand with white epoxy (CMS recent discovery). FOR YIELD TEST: JIA LI should try to polish both ends of a few 0.9 WLS fibers. Fiber length 20" (large tile fiber groove length) + 2 x (8-10") = 20" + (2 x 9") = 38". LEE is coming to FNAL for collaboration meeting. He should be prepared to: 1. meet with electronics guys about final designs and cooling (EG will set up) 2. go to Lab 8 with hv fanout prototype to discuss any changes to the production 3. bring lv fanout design to lab8? 4. collect info for his upcoming conference talk. 5. meet with Amber (L2 software). ANDY should order 0.9 mm Bicron BCF 91A WLS fiber and have it shipped directly to: Eileen Hahn, Fermilab M.S. 327 Lab 7, P.O. Box 500, Batavia, Illinois 60510 Let EG know when the order is placed so she can flag Eileen. ------------------------------------------------------------------------- ------------------------------------------------------------------------- JIA LI will come to FNAL for the collaboration meeting to meet with FPS group (Kotcher and Abid) as well at FNAL engineers (Jack and Herm) to discuss mechanical support, cable routing, space available for cable routing (space between EC and CC), alignment and weld stud placement, clearance required for ICD removal for emergency FPS access, and cable and mechanical coordination with the FPS and L0 groups. JIA reports that the 2nd prototype aluminum box looks good though the cover currently on wrong side, but only minor modifications are necessary. This box features an end piece which may be removed and replaced if the location of the connectors are changed. Abid put an updated version of the FPS clear fiber routing scheme on the web from URL directory http://www-d0.fnal.gov/~abid/fps/ Two files: FPScable2.dwg (ACAD 14 drawing, dated: June 28, 1998) FPScable2.eps (postscript version of the same drawing) Jia Li should overlay our current cable routing with this FPS cable routing drawing and bring to FNAL. We need one crate for the prototype test and one for the solenoid test. In principle, we could use the same one but since dimensions are final why not go ahead and build them all. Jia Li will have them made by a local machine shop. ------------------------------------------------------------------------- CLIFF/KATHLEEN/LEE reported on the drawer production status and have a final drawer design. Here is a summary of the mental processes that led to the final design. Factors in ICD PMT drawer design are: 1. light tightness (favors closed design) 2. temperature (favors open) and temp stability (may favor closed design). 3. machining and assembly complexity (favors closed design) - they want to minimize tapped holes so sliding top is good. Jia Li's original design was a closed design. LT thought about using Jia Li's open design using baffles. This design had a baffle which socket wires had to pass through, and too much additional space was required to keep the design light tight. The biggest problem with the closed design was the temperature dependence. The total power dissipation is expected to be 1.2 watts per box. A LT mock up showed that with time, a rise of about 10 degrees is observed, followed by a plateau indicating temperature stability. Cliff's finite element heat transfer simulation confirmed this (air temp rose, but transferred to Aluminum - reaches equilibrium because aluminum transfers heat to outside air by convection). So this leads us to believe that closed design may not be so bad, and a closed box gives more uniform operating temperature. So right now, drawer design is closed. EG, LS will inquire about preamp temperature dependence and calorimeter plans for preamp heat dissipation. If cooling is required, fans would be easiest and chilly air or water cooling very difficult to implement due to access. Cliff reports current dimensions: Drawer is square with motherboard (card) inside: Outside dimensions of drawer --> 10" x 10" x 1 3/4". Drawer interior dimensions 9.5" square interior room after machining. Motherboard dimensions are 9.2" wide by 7.4" long. We need a final mechanical drawing with all correct dimensions for completeness. Kathleen reports they will have a prototype drawer monday. Shortly after, 15 more drawers will be produced in the next couple of weeks. After evaluation, the following 3 sets of 16 drawers can be completed (need production done before quarter begins in september to avoid charges). LEE notes that there are a few minor modifications to the drawer design: - lee says reduce # of screws - make hole for PMT sockets a little wider (allow for shrink wrap on socket wires. - make a double hole on face plate. So we will have 16 drawers for the full prototype test as well as the solenoid test (though only a couple will be instrumented). Temperature sensors are planned for both full prototype and solenoid test. ------------------------------------------------------------------------- Mark/Victor/Andy/Kaushik Module test stand status Kaushik discussed two (of 5) ways of getting pe yield of our ICD tiles. The results from methods used previously by KD are shown in: http://heppc1.uta.edu/kaushik/tilecal/scint/proto1.htm METHOD 1: measure light from both ends of a WLS fiber lying in the tile groove. Fiber should be the length of the groove plus 8-10" on both ends. The light out of each end is read by a PMT, giving 2 measurements from each light pulse (or trigger on cosmics using trigger paddles above and below the test tile). The yield is obtained from the ratio of the means divided by half the difference in their widths, a relation yet to be clarified by KD. Previous tests based on a 4 mm tile, showed uniformity to within 5% with a laser system. Ted is working on getting the laser scanner working again, but modifications may be necessary to fit an entire tile module (setup designed for individual tiles). Results from this test compared to a real tile module will result because a tile will perform differently when wrapped in reflective material and the fiber configuration will be slightly different. EG suggests using 2 0.9 mm diameter fibers per groove to more accurately reflect the current ICD tile module design. Kaushik says that for showers, we want an overall resolution of better than 10%, so uniformity must be kept to better than 5%. KD: how do you directly relate laser uniformity to shower resolution and what are the other factors involved ? To execute Method #1 we need fiber polished on both ends about 38" long for the largest tile. Jia Li says he can polish single fibers at UTA. He should send EG fibers if he is unable to do this --> EG can get access to the single fiber polisher at Lab 5 at Fermilab. We have enough WLS fiber for this test, but we will order more for production (see action item above). METHOD #2 : pulses the tile with an LED signal driven by a variable current. As the current is lowered, output becomes non-linear and then becomes flat. When the ADC output is flat, the system is at the 1 pe limit. We measure resistance in kohms vs output. ------------------------------------------------------------------------- This lead into a calibration discussion, in which I got lost and confused somewhat so whoever wants to add anything, let me know. PAUL /LEE/DICK (unclear about responsibility here) Paul will have a working system to test in the next week based on the current design using LED's and a distribution block to distribute light to PMT's in a crate with fibers potted into the block. Monitoring is one question: the current design includes pin diodes to monitor the light distribution system. Paul can have all pieces made for next week and there are lots of tests to run. In particular, he'll look at light distribution uniformity using 1 LED per block verses 4 LEDs/block. Paul is ready to glue fibers into the distribution block and put it into the backplane. Location of the block is not important for the full prototype test or the solenoid test, since access will not be a problem. In Run II, EG anticipates problems with access to the high abs(z) side and back of the PMT crates due to the cable winder and cryostat supports. Paul can go ahead and epoxy the fibers into the cookies. Once having done so, protect the ends of the cookies from dirt and abrasion since in the current design, polished fibers are unprotected. Andy reports that all pieces are there to plug the backplane fiber cookie/connector/calib_block assembly into the backplane structure, again making sure to protect all fiber ends. We'll use one drawer near the end and one near the middle of the crate, for a total of 12 channels instrumented. Scintillator excess from the cutout of the 1/2" thick tile production is currently in the Swift center and available for use. Kaushik also suggests we evaluate the APD option because he remembers they are cheap and will give a more linear response over a wider range. EG will ship the APD setup to UTA or send it back with Jia Li. Paul will do preliminary evaluation to determine feasibility/cost. The led circuit can pulse at different currents so we should be able to measure some dynamic range of the system as well as the time dependence. Paul needs a lv line for the pin diode (+12 and gnd) for the solenoid test and Run II. Running a very small number of individual lines such as these will not be a problem (only large numbers of cable). Tom Regan (fnal electronics) working on muon system. someone is going to ask him something. ------------------------------------------------------------------------- SOLENOID TEST For the solenoid test, we'll need only calibration fibers (no tile signal fibers or connectors). So more polished single fibers will be needed if we build a new backplane fiber assembly. EG now recalls that Del and Co. will install the pmt crate/iron/backplane assembly when it is delivered to D0 (after the initial field mapping). Location of the block is not important for the solenoid test, since access to all sides will not be a problem. In Run II, EG anticipates problems with access to the high abs(z) side and back of the PMT crates due to the cable winder and cryostat supports. We'll use one drawer near the end and one near the middle of the crate, for a total of 12 channels instrumented. Lee plans to monitor temperature during the solenoid test. Paul will use an off-the-shelf lecroy pulser from the control room to pulse LED's for the solenoid test. ------------------------------------------------------------------------- ELIZABETH reports that we can have our WLS fiber for the pigtail production polished at Lab 7. Lab 7 will do the aluminum sputtering as well. Fiber should be ordered immediately (see action item for Andy above) and shipped directly to Lab 7. If we want other fibers polished, we should put our requests into Lab 7 ASAP in order to get any polishing done while they are still set up to do it. Other fibers include backplane clear and WLS 1mm diameter fiber. ------------------------------------------------------------------------- KAUSHIK reported on Run I box decommissioning. Approximately 30 boxes have been dis-assembled. Before gutting, we insure the box has been tested. Box number and PMT numbers are verified for the database and proper tracking. PMT's are wrapped in a foam packing material. They expect to finish by the end of the summer. ------------------------------------------------------------------------- LEE gave an update of electronics status. He has quotes for assembly of motherboard and bases. K&F will make daughtercards, but does not do assembly. Recent reports from ??? indicate that we could see a drift of component values in components on the surface mount bases. Lee will contact Ray yarema at CDF about their experience. KD emphasizes the need for filtering on the HV supply and warns against switching resistor/capacitor to first stage. He also notes that higher rated lower value capacitors may have better overall filtering performance with time. On motherboard, KD says 1. LV filtration needs to be checked by dean 2. need r for hv rc filter 3. check diode on base for noise Also, fix capacitor on the base itself, not on the motherboard. KD also recommends that we etch onto the coverplate "do not remove with any connector plugged in" since this may blow preamps or cause other damage. There are 3 motherboards on order from K&F and Lee may have one next week to show Dean at collaboration meeting. Lee reports sockets are on order for preamps and HV jumpers. Lee received the 2 prototype HV fanout boards from joann at lab8. They look great except for one cut through ground plane - would like ground plane to be continuous. He will bring the board back to lab 8 next week and report to Joann to start production. Lee reports no final design for lv fanout yet. ANDY wrote up some notes on commissioning plans: ------------------------------------------------------------------------- Initial thoughts on Run II ICD Commissioning ------------------------------------------- I had a discussion yesterday at Fermilab with Dan Owen(newly appointed person in charge of RunII D0 Commissioning). Dan is starting to collect plans/schedules/requirements for commissioning activities over the next two years. Below is my initial list that I came up with talking to Dan, and as a way of starting our discussion on this topic. Note: Commissioning activities are distinct and DO NOT INCLUDE installation work. *** Specifically we should discuss this at the ICD meeting TOMORROW (7/2) *** since Elizabeth will make a short presentation on this at the Workshop *** next week. 1) Solenoid test: This gives us an opportunity to - operate a few (12?) ICD channels in the environment of the D0 detector, but with a partial old DAQ etc. - check the B-field on/off behavior of the PMTs in the iron block - determine the effect of the ICD block on the solenoid field 2) North end ICD - The North end ICD is scheduled to be installed in mid-1999. This gives us the opportunity to test/partially calibrate this part of the ICD using cosmics. This requires that we work with the online/DAQ to let or needs be known - we presumably do this via Dan. This exercise should make the commissioning of the South ICD, which is more time critical, go faster/more smoothly. 3) South end ICD - Scheduled to be installed in late 1999(?). There should be time for cosmic tests. Collisions are scheduled to begin on 4/14/00 and according to Dan there are NO plans for a cosmic commissioning run, just plans for beam tests. We need to consider this in the light of our calibration needs. 4) We need to consider the software (CALIB equivalent etc.) for which we need to supply ICD procedures.