Retested cards used at DESY 201003. =================================== All plots referred to can be found in: http://www.hep.lu.se/eudet/pictures/fectest/20100412/ Cards with found problems ------------------------- FEC comments 2 channel 120 has low rms and high pedestal 3 chip 5 - all channels 0 pulses seen but low amplitudes -> negative "pedestal" 6 memory error in channel 119 7 1.5V is 1.33V, shutdown of PCA16 -> 1.49V channel 11 high pedestal/rms channel 12-15 low pedestal/rms 10 gain1 shaper0 dac2500 rms high: 2.5-6.5 23 channel 67 high pedestal, no signal memory error in channel 47 (location 413, 3FF->3F7, 8->0) 25 1.5V is 1.40V, shutdown of PCA16 -> 1.49V 0.016A channel 111 somewhat high in some settings 34 1.5V is 1.0V, high pedestals 36 memory error in channel 65 38 1.5V is 1.0V, shutdown of PCA16 -> 1.49V 0.50A 56 1.5V is 1.38V, chip 5 channel 15 low ca 0. Location of cards with problem or too low voltage to PCA16 ---------------------------------------------------------- FEC RCU slot GEM module connector 2 1 28 5 159,146,131,118 3 0 4 3 65,64 7 1 0 0 150,120,89,74 10 1 5 0 138,107,79 23 2 9 0 119,104,88,73 25 1 27 5 160,147,132,119 34 2 22 3 44,30,17,3 38 1 3 0 137,106,77 56 0 6 3 68 Would be interesting to know if there was problems on GEM connector side on these!? Power test of cards 7,25,34,38,56 ================================= The test was done bypassing the regularator with extra power supply to force the power to PCA16 to 1.5V. Plots are showing pedestal (x) vs channel (y). Power as before (2.5V in -> 1.36 to PCA): fec7-0-vtest.gif fec7-1-vtest.gif (enlargment) Increase power in to 3.2V -> 1.50V stable, 1.02A fec7-2-vtest.gif fec7-3-vtest.gif (enlargment) Power as before (2.5V in 1.43 to PCA): fec7-0-vtest.gif Increase power to 3.2V -> 1.51V stable fec7-1-vtest.gif Plots show both power 2.5V - high pedestals, and forced to 1.5V - low pedestals in same plot. (total current 0.87A + 0.63A) fec34-0-vtest.gif fec34-1-vtest.gif (enlargment) Plots show both power in 2.5V (1.1V to PCA16) - high pedestals and forced to 1.5V to PCA16 - low pedestals in same plot. total current 0.87A + 0.41A. fec38-0-vtest.gif fec38-1-vtest.gif (enlargment) Plots showing forced to 1.5V to PCA16 fec38-3-vtest.gif fec38-4-vtest.gif (enlargment) 2.5V in -> 1.42 V to PCA16 (0.87A) fec56-0-vtest.gif (enlargment) 2.9V in -> 1.52V to PCA16 (0.97A) fec56-1-vtest.gif (enlargment) The plots show that the bad channels are bad with any power in, while the other channels are probably working if properly powered. Summary of cards with bad pedestal channel(s) card comment 2 channel 120 (chip 7) bad pedestal 3 channels 80-95 (chip 5) pedestals <= zero 5 channels 48-64 (chip 3) pedestals <= zero 7 channels 11-15 (chip 0) bad pedestals 23 channel 67 (chip 4) high pedestal, no signal 25 channel 111 (chip 6) somewhat high in some settings channels 96-111 somewhat higher rms 34 channels 43-46 (chip 2) bad pedestals channels 49-54 (chip 3) bad pedestals 38 channels 6,7 (chip 0) pedestals <= zero channels 21,30,31 (chip 1) pedestals <= zero 56 channel 95 (chip 5) pedestal <= zero General remark -------------- The 1.5 V regulator is designed to deliver 0.75A The PCA16 draws more in normal operation 0.9-1.1 A. Increasing the power in (from 2.5V to ca 2.8V) makes the power to PCA16 more stable at 1.5V, and the current increase from ca 0.87A to ca 0.95A. If the regulator has problems delivering the current, the voltage drops. The voltage drops as the regulator warms up, i.e. this can explain the drifting pedestals after power on. We will buy a regulator that is designed to deliver more (1.25A) and test on a card. Cards brought from Japan ======================== FEC 4,11,80,81 The test was done bypassing the regularator with extra power supply to force the power to PCA16 to 1.5V. plots show pedestals in red -left scale, rms of pedestal blue - right scale. all plots with 1.5V to PCA16, except for card 4 which was working. The plots show pedestals for different settings of gain,shaping,decay time. card comment 4 file: combined-fec4_20100504.pdf card is working 11 file: combined-fec11_20100503.pdf 2.5V in -> 0.96 V to PCA16 (0.87A) forced 1.5V in (0.66A) bad channels: 4,32-47,58-65,67,69,73,74,76,77,94-95,98,101,104-106,109,110,126 all have zero pedestal bad chip channel 0 4 2 32-47 3 58-63 4 64-65,67,69,73,74,76,77 5 94-95 6 98,101,104-106,109,110 7 126 80 file: combined-fec80_20100504_2.pdf 2.5V in -> 1.01 V to PCA16 (0.86A) forced 1.5V in (0.57A) bad channels: 3,7,23,34,35,38,40-47,55,58-62,68,74-76,80,89,92,97,98,101,103,104, 106,107,110,111,114,119,124,126,127 three channels have a pedestal of 1023, the rest 0. bad chip channel 0 3,7 1 23 2 34,35,38,40-47 3 55,58-62 4 68,74-76 5 80,89,92 6 97,98,101,103,104,106,107,110,111 7 114,119,124,126,127 81 file: combined-fec81_20100504_2.pdf 2.5V in -> 1.0 V to PCA16 (0.89A) forced 1.5V in (0.59A) bad channels: 6,12,32-43,51,58-63,67,69,73,74,77-79,88,89,94,101,102,104-106,108, 110,111,120 two channels have a pedestal of 1023, the rest 0. bad chip channel 0 6,12 2 32-43 3 51,58-63 4 67,69,73,74,77-79 5 88,89,94 6 101,102,104-106,108,110,111 7 120 12 file: combined-fec12_20100603.pdf 2.5V in -> 0.87 V to PCA16 (0.86A) 2.5V in -> 1.01 V to PCA16 (0.76A) bad channels: 4,6,7,9,26,33-36,37-47,55,57,58,60-64,66,67,69-73,75-77,93,94,97,100-103, 105,107-110,127 two channels have pedestal 1023, 1 has a high pedestal, the others low but nonzero bad chip channel 0 4,6,7,9 1 26 2 33-36,37-47 3 55,57,58,60-63 4 64,66,67,69-73,75-77 5 93,94 6 97,100-103,105,107-110 7 127 The input was pulsed when forcing 1.5V to PCA16. The responses with zero suppression are shown for gain 0: fec4-0-0-2500.gif fec11-0-0-2500.gif fec80-0-0-2500.gif fec81-0-0-2500.gif fec12-0-0-2500-nosignal.gif fec12-0-0-2500-signal.gif The responses are shown for gain 3: fec4-3-0-2500.gif fec11-3-0-2500.gif fec80-3-0-2500.gif fec81-3-0-2500.gif fec12-0-0-2500-signal.gif fec12-3-0-2500-signal.gif Special study of card 11 for a comparison: (x axis - pulse height, y axis - channel number) fec11-0-vtest.gif pedestals when not forcing PCA16 to 1.5V fec11-2-vtest.gif pedestals when forcing PCA16 to 1.5V No signal visible for bad channels. Remark about damaged cards -------------------------- For the damaged cards to be used do one need to find a way to cure/replace faulty chips, and try to understand why they have become bad to avoid it in the future. Cards not used at DESY/Japan ============================ 5 channels 48-63 has pedestal <=0, but pulses are seen (known fault) 45 OK 46 channel 72 has low pulse height Put a more powerful PCA16 regulator on FEC5 ========================================== 20100511 FEC5 choosen since it has one bad chip, so we could not destroy a fully functioning FEC. Specification : >=0.6V drop., expect unstable voltage out at <2.1V in. power in out 2.5 1.514 2.4 1.514 2.3 1.514 2.2 1.514 2.1 1.514 2.0 1.521 1.9 1.424 Choise: 2.2V warming up effect, studied pedestal of one channel: minutes pedestal after poweron 2 103.3 3 103.9 4 104.0 5 104.2 6 104.4 7 104.7 8 104.8 9 104.8 10 104.9 11 104.9 12 104.9 13 104.9 14 104.9 15 105.0 16 105.0 17 105.0 all pedestals in file: http://www.hep.lu.se/eudet/pictures/fectest/20100412/combined-fec5_20100511.pdf comparison: http://www.hep.lu.se/eudet/pictures/fectest/20100412/combined-fec5_20100504.pdf no significant difference in result. REPAIR OF FEC34 =============== 20100616 -------- The PCA16 chips2 and 3 was replaced, as well as the power regulator for the PCA16s. The new PCA16 chips have numbers: 2 -> 602 3 -> 618? difficult to see number 2.3 V in = 1.612 V to PCA16, and 0.898A 2.0 V in = 1.534 V to PCA16, and 0.847A before "repair", the power to PCA16s was 1.0V It was tested again with 2.2V in => 1.612 V to PCA16, the results are in http://www.hep.lu.se/eudet/pictures/fectest/20100412/ the pedestals for each PCA16 setting is shown in: combined-fec34_20100616.pdf Pulsing all channels (two different gains): fec34-0-0-2500-20100616.gif fec34-3-0-2500-20100616.gif The result show that the card work correctly. One may complain about one channel, in chip2 channel 38 ( = PCA16 chip# 602) has a rather low pedestal, below zero for high gain settings. Cost to replace PAC16 chips are 17 euro/chip (170SEK) to replace power regulator 6.5 euro/chip (65SEK) ILC FEC REPAIR FIRST PROPOSAL ============================= Cards with found problems on PCA16 ---------------------------------- FEC comments REPAIR? 2 channel 120 has low rms and high pedestal (dead) NO? 3 chip 5 - all channels 0 YES pulses seen but low amplitudes -> negative "pedestal" 5 channels 48-64 (chip 3) pedestals <= zero YES 7 1.5V is 1.33V, shutdown of PCA16 -> 1.49V YES channel 11 high pedestal/rms (chip 0) channel 12-15 low pedestal/rms 10 gain1 shaper0 dac2500 rms high: 2.5-6.5 NO? 23 channel 67 high pedestal, no signal (dead) NO? 25 1.5V is 1.40V, shutdown of PCA16 -> 1.49V 0.016A PR? channel 111 somewhat high in some settings 34 1.5V is 1.0V, high pedestals DONE 38 1.5V is 1.0V, shutdown of PCA16 -> 1.49V 0.50A chip 0,1 YES 56 1.5V is 1.38V, chip 5 channel 15 low ca 0. NO? (or YES?) NO? = No need since only one channel bad? Is it all PCA16 error? YES = Needed, In total 5 chips on 4 cards PR? = perhaps enough to change power regulators? not sure which chip to change. NO? + YES = 8 cards and 9 chips. NO? + YES + PR? = 9 cards and 10 chips. This file is: http://www.hep.lu.se/eudet/pictures/fectest/20100412/readme.txt Completely Green PCA16 chips are: 21,34,39,46,48,50,51,53 56,57,62,66,67,68 70-76 78-80 185 506,514,516,525 623 635 (says F.S. = given to Fabio Sauli) 640 (says F.S. = given to Fabio Sauli) 776 786 882 602,618 was used on FEC34. i.e. 33 chips. we need 9 for Lund + 30 for Japan = 39. or 5 for Lund if only very bad chips are replaced. then there are some more that are green and a have number 1 on the list, whatever that means. 20100629 ======== Status of chips with a 1 as comment OK = seems fine to use OK? = some complain NOT = more serious error 22 OK 26 OK 27 OK 35 OK 40 OK 125 OK? Ch 11 low base 141 OK 147 NOT Ch 3 large offset 185 OK 195 OK 196 OK Ch 9 slight tail, ch 7,9 baseline somewhat off 319 NOT Ch 16 large offset 335 NOT Ch 7 baseline off, else OK? 343 OK? Ch 7 baseline off 370 OK 388 OK? Ch 7 baseline off 406 OK? Ch 9 slight tail, high baseline 464 OK 465 OK 500 NOT Ch 15 baseline off rather much 506 OK 514 OK 516 OK Ch 7 slight baseline off 525 OK 612 OK 613 NOT Ch 7 baseline off 617 OK 622 NOT Ch 4 baseline off a lot 623 OK 656 NOT Ch 14 baseline off 675 OK? Ch 8, 16 off 734 NOT Ch 10 baseline off a lot 739 OK? Ch 3,6 off 776 OK 777 NOT Ch 3 off a lot 786 OK 807 OK 819 NOT Ch 12 baseline off a lot 824 OK 830 OK 838 OK? Ch 11 off 839 NOT Ch 11 off a lot 843 OK 848 OK? Ch 3 off 870 OK 882 OK 887 OK 914 NOT Ch 4 high, ch 5 low (short?) 929 NOT Ch 4 off a lot In total: 27 with OK 8 with OK? 13 with NOT