Update on October access (10/23/01): ------------------------------------------------------ Work done is listed under the individual problems. **Access so far** indicates that work is ongoing. **Theory** indicates the buest guess at what is going on. **Things to do** indicates the next steps that I believe we should take. Comments are more than welcome. Plans for October access (updated 10/16/01) ----------------------------------- 1.High Voltage work. 2.Electronics work. 3.Insulation work 4.Survey (possibly move?) work ******************************************************************************* 1. High Voltage work: ---------------------------- {Note: Normal capmeter reading for sense is 5.6 to 5.8nf and for potential is 2.5 to 2.7nf. In all cases, nearest neighbor cells should also be checked BEFORE and AFTER making the modifications. Twice we found that nearest neighbor ribbon cable connections had been jostled. } Note that this is a high prority list, consisting of known or possible high voltage breakdown and/or high voltage connection problems. SL1-106-11, SL4-227-05 are known missing high voltage connections that are also known breakdowns. SL7-138-05 is a known breakdown with likely good connections. SL1-064-10, SL8-445-01 and SL8-445-11 have connection problems (YMon "low occ" and Dave-pulse "DTH = 150"), but no known breakdowns as yet. The remaining problems are either "high occ" or "high mult" in YMon, but look fine in Dave-pulse. In the past these have always turned out to be strictly readout problems, and the lack of "Dave-pulse" was likely due to a difference in "cuts" or some intermittent behavior. These include SL1-151-01 ("high mult" = 3.5), SL3-061-01 ("high occ"), SL7-233-0 ("high mult" = 3.2). Note that SL8-105/105-09 is a special case where both S9's were disabled, but only one caused the breakdown. We hope to determine the bad one near the end of the shutdown when Ar/Ethane flow is resumed. -SL1-064-10 (B,324,14-06-0) (USE): -->YMon "Low Occ" and Dave-pulse "DTH = +150". Signature of missing HV connection, so check with capmeter. The cap reading was previously checked "ok" in May, so it may be intermittent. **Access so far**: -->Capmeter: All sense wires normal (5.7-5.8nf) in cells 63, 64, 65. Pot wires normal (2.7) cell 64. -->Spy: Compared to neighbors: Pulse wider on 64-10 and could not see noise when using capmeter. -->Unplug ASDQ DB: Capmeter: All sense wires near 8nf expcept 64-10 = 5.7nf (unchanged). Morris measures all 25kohm series resistors on ASDQ DB to be fine. -->Tried measuring resistance to ground from ASD MB socket for ch 64-10 (9th from OD). Saw nothing, indicating no contact through socket to cold capacitor leg. -->Add various reistors in series with "cold" cap leg on a good channel (11): Capmeter: infinity: 8.2nf, 5Mohm: 8.2nf, 470kohm: 7.9nf, 330kohm: 7.4nf, 120kohm: 6.1nf, 47kohm: 5.9nf, 5kohm: 5.8nf (same as ground). -->Use dental pick to measure resistance from cold capacitor leg to ground: 0.3ohm (dead short) -->Use high magnification to examine footprint of cold leg: either solder or "artwork" bridge between cold leg and ground (reason for dead short). Decide to punt. **Theory**: -->It looks like this channel has a coincidence of a bad socket and likely bad artwork on the ASD MB. We could cut the cold leg of the cap and "jumper" it to the input of the ASD DB, but this seemed like a future high voltage problem in the making. We decided to punt (that is, leave channel dead in readout, but stable for high voltage) and just re-install the ASD DB and microcoax. Ayana attached the analog monitor cables which were also unplugged. **Thing to do**: -->Check connections for cells 64 and 65 in Spy or Dave's program. -->Check HVDB with capmeter for 63, 64, 65 (WEST SIDE). -SL1-106-11 (C,473,16-03-1) (LSE): -->There is a trip in quad c (cells 106-108). Cell-106-11 has YMon "low Occ" and Dave-pulse "High DTH" indicating a disconnection, so we should check this with the capmeter and hopefully fix the problem. The quad d S11 trip as yet has no candidate. **Access (so far):** -->Capmeter: All sense wires normal (5.7-5.8nf) on 106, 107, 108; all pot ok on cell 106. -->Spy: Compared to neighbors: Pulse wider on 106-11 and could not see noise from capmeter. -->ASDQ DB: Measure .7v at input resistor for 106-11 and neighbors (as expected). -->Unplug ASDQ DB: Morris measures that 25ohm series resistor on ASDQ DB is near 400kohm for 106-11. Capmeter: 106-11 = 8.0nf, like nieghbors (normal for unplugged ASDQ DB). -->Plug in new ASDQ DB: Capmeter: All cell 106 sense wires measure 5.7 to 5.8nf. Spy: 106-11 is now same width as neighbors. **Theory**: -->It's most likely that the 106-11 25ohm series resistor "blew" due to the observed channel 11 breakdown in the cell region 106-108. Since all connections were/are good, we don't have a candidate for the cause of the breakdown. There is an inconsistent measurement. Morris measured that the channel 11 series resistor was 400 kohm after the ASDQ DB was unplugged, but we should have seen a 7 to 8nf capmeter reading with this resistance instead of the normal 5.8nf. Perhaps the resistor continued to "break" during the board extraction and measurement? Perhaps another "carbon path" (see alleged carbon path in 1-64) was knocked loose? **Things to do**: --> Perhaps the most reasonable way to proceed is to disable 106-11 at the HVDB (remove 300ohm resistor and tie to P12). Here we are using the blown series resistor as a signature for the breakdown, and since there is no known cause, disable the wire. Check values of all 300 ohm resistors on HVDB. --> The microcoax for cell 104/105 was unplugged during the ASDQ DB swap. We still need to check connections on these channels (the connections for 106/107 were checked). -SL1-151-01 (D,691,19-20-3) (LNE): --> YMon "High Mult" (3.5), but Dave-pulse says ok. First look with Spy. **Access (so far)**: -->Spy: Looks quiet with normal looking pulses. Intermittent?? **Things to do**: -->Leave it. -SL3-061-01 (B,194,02-12-2) (UNE): -->YMon "High Occ", but Dave-pulase says ok. First look with Spy. -SL4-227-05 (D,579,18-17-3) (LNE) : --> Bad cap connection (lots of other HV problems with cell). BACKGROUND (from September access): Before removal, HVDB (SN ?) caps measured S0 = 8.2, S2 = 8.5, S3 = 8.6, S4 = 8.5, ***S5 = 7.6***. After removing the HVDB, all wire measured about 615nf (including known +40nf offset) = ASD cap + 40nf wire capacitance, EXCEPT S5, which measured 80 - 40nf offset = 40nf. This means that S5 is disconnected at the ASD motherboard, either at the hot cap leg or the ground leg. We did not come across this when previously working with this cell (intermittent??). Given that we have no power to check things out after working on the ASD end before the EAST plug closes, we decided to just proceed with the planned work and hope that things last until the October shutdown. Tieing S5 to the P6 voltage should dramatically reduce the trips, since it was infrequently tripping at the normal S5 voltage. Anyway, the 300ohm resistor was removed from S5 and it was jumpered to P6. After reinsertion, S0 = 8.15, S2 = 8.45, S3 = 8.52, S4 = 8.53, S5 = 7.96, P0 = 2.91, P2 = 2.88, P3 = 2.91 and P4 = 2.90. The reading for P6 was normal, since S5 does not see the ASD capacitance. ACTION: In addition to the above, a new problem shows up in YMon: 227-8 is completely dead, but this is a TDC problem. Something is really odd with this cell. Anyway, the first thing to do is determine the source of the 227-S5 missed connection (ASD MB or ASD DB). My guess is that we will decide to crack open the ASD MB. Aseet suggests that if we need the stringing engine, a temporary seal can be made with tacky tape until it can be set up. **Access so far**: -->Capmeter: Verified that the above capmeter readings are still valid. -->Capmeter: Remove HVDB and measure (+13 pf offset): All channels near 600pf except S5 = 52nf (true = 52 -13 = 39pf). Confirms that "disconnect" is at the ASD end. -->Remove ASDQ and measure "25 ohm" series resistors: All near 25 ohm except 227-S5, which measures "open" (> 20 Mohm). -->Plug dummy ASD DB into 226/227: Capmeter: 227-S5 now measures normal (about 600pf). **Theory** -->Breakdowns of 227-S5 "blew" the 25 ohm series resistor on the ASDQ DB. Since the resistor is completely open, it's not really safe to leave it this way, even with the sense wire tied to the neighboring potential wire high voltage (cold leg of cap with charge up and possibly breakdown on the ASD MB). **Things to do**: -->Since the cell is half disabled, it still seems useful to crack open the motherboards (one at a time) and try to find the problem - may require use of stringing engine. -->Either replace blown series resistor on ASDQ DB or install a spare. -SL7-138-05 (B,233,03-18-1) (USW): --> Trips in cell 138. Tie S5 to P6. **Access so far** -->Capmeter: All wires normal in cells 137, 138, 139. -->Removed HVDB for cell 138 **Things to do** -->Check values of all 300 ohm series resistors on HVDB. -->"Disable" channel 5 by removing 300 ohm resistor and ties to P6. -->Reinstall HVDB and check capmeter for cells 137, 138, 139. -SL7-223-00 (C,375,05-21-3) (USW): -->YMon "high mult" = 3.2, but Dave-pulse ok. First look with Spy. -SL8-104-09 and SL8-105-09 (A,178,12-05-2) (UNE): -->Trip previously disabled. BACKGROUND (from September access): Cell 104: HVDB (SN 2332) caps measured ok before removal. S9 had it's 300 ohm resistor removed and was jumpered to P10. After re-insertion, P10 = 2.8nf and S9 = 8.1nf, which is ok. Cell 105: HVDB (SN 2351) caps measured ok before removal. S9 had it's 300 ohm resistor removed and was jumpered to P10. After re-insertion, P10 = 2.8nf and S9 = 8.3nf, which is ok. ACTION: Toward end of access. Just before Ar/Et flow, reinsert resistor to (say) cell 104. See if we have trips at full voltage. If do, change 104 to 105. We know that this trip occurs without beam, so this should be a conclusive test. -SL8-445-(01 and -11) (D,708,19-06-0) (LNE) : -->YMon "low occ" and Dave-pulse "DTH = 150". Check with SPY and capmeter for connection problem. **Access so far**: -->Capmeter: All sense and potential channels in cell 445 measure normal except ch 11: about 7nf. It's important to note that ch 1 measured normal, 5.8nf. -->Spy: Both chs 1 and 11 have wide pulses compared to neighbors in Spy. Unfortunately, we forgot to check for noise when measuring the capacitance. -->Unplug ASDQ DB: Capmeter: Ch 1 = 8.2nf, Ch 11 = 8.4nf. Morris measures 25 ohm series resistor to be 12.5kohm for ch 11, 31ohm for ch 10, 21 ohm for ch 9 and normal 25 ohm for the rest., including ch 1. -->New ASDQ DB: Capmeter: ch 11 = 7.8nf, ch 10 = 5.8nf, ch 1 = 8.2nf, ch 0 = 5.8nf. VERY SURPRISING, especially for ch 1. -->Unplug HVDB: Capmeter (with +10pf offset): ch 11 = 606pf (normal), ch 10 = 606pf, ch 1 = 47pf, ch 0 = 616pf. Note that about 40pf is what is expected for a disconnection at the ASD end. -->Ch 11: Found 300 ohm series resistor on HVDB "blown" to 400 kohm, which accounts for the 7.8nf reading with the HVDB plugged in. Therefore, breakdowns on this wire blew the series resistors on both the HVDB and ASDQ DB. -->Ch 1: We could make the ch 1 connection to the ASD end come and go (ie, capmeter readings go from 600pf to 40pf) by wiggling the cap on the ASDQ MB or sometimes by pressing on the ASDQ MB near the hot leg of the cap. Measurements using the dental picks on the cold and hot legs were inconclusive. The solder joints of both the cold and hot legs were "reflowed", but the signal would still come and go (sometimes on it's own and sometimes by wiggling the cap or pressing the board). After cutting the cold leg of the cap, we still could not see a connection to the other side when poking the dental pick to the stub of the low insertionn force socket. We decided to also cut the hot leg, remove the cap, and put corona dope of the stub of the hot leg. The channel is obviously still dead in readout, but should be safe for high voltage. -->Replace original ASDQ DB with spare. Plugged in microcoax (both for 444/445 and 446/447) **Theory**: -->Ch 11: Here it's pretty clear that breakdowns caused the 300 ohm series resistor on the HVDB to blow to 400 kohm, and the 25 ohm series resistor on the ASDB DB to blow to 12.5 Kohm. The cause of the slightly changed values of the 25 ohm resistor on channels 10 and 9 is not obvious, but Aseet thinks it may be later breakdowns due to the the changed RC of ch 11. There have been several trips on SL8-S11 which have not been pinned down to a quadrant or cell. This is certainly a good candidate for those trips. -->Ch 1: It' not clear that the flakey connection at the ASD end has caused any trips, since the series resistors on both DB's are normal. There are no SL8-S1 trips in my records of the last couple months. **Thing to do** -->"Disable" channel 11 on the HVDB: Remove 300 ohm resistor and tie to P12. Check values of all 300 ohm resistors on HVDB. -->Check cells 444, 445 and 446 with Capmeter after reinstalling HVDB -->Check signals on 445 - 447 with Spy or Dave's readout. ******************************************************************************* 2. ELECTRONICS WORK: ----------------------------------------- -SL6-LNW QDR short to -3v due to washer. Dave said that "null-noise" indicates that possibly both TREFE and TREFO are affected. He believes the washer is now trapped between the CV board and the extrusion. -SL2 microcoax miscabling: Swap cells 24/25 with 26/27 (A,046&047,01-16-2&3) (UNW). -->MIcrocoax miscabling. Fix and undo the "compensating" swap at the TDC. **Access DONE**: -->We found the microcoax swapped and put them in their correct position. The flat cables were then put in their normally correct locations at the TDC input, and a note was written on the cable indicating that this was done. All channels were checked to be ok in Spy. In addition to the above, the following is a list of "cell level" readout problems which stick out like a sore thumb in the YMon. All have corresponding problems in Dave-pulse. -SL2-071-(08 & 11) (B,325,04-06-1) (USW): -->Dead in Dave-pulse, "low occ" in YMon. -- Nothing in old records. **Access so far** -->Spy: Both channels: Levels ok, 1/2 pulseheight -- indicative of "microcoax" problem. -SL6-096-(8-11) ( (B,236,03-14-0) (USW): -->Hot in Dave-pulse and "high occ" in YMon. At HV ribbon connection. **Access DONE**: -->Spy: Observed noise while working on ASDQ end (least likely). Pressing microcoax had no effect and angle brackets and screws secure. -->Spy: Noise disappears when better ground HV ribbon bundle to F-cage. -SL7-050-(8-11) and -051-(6-11) (A,089,01-03-1) (UNW): -->Hot in Dave-pulse and "high occ and high mult" in YMon. **Access DONE**: -->Spy: Loose screw on angle bracket. Noise gone after tightening. -SL7-372/373-(5-11) (D,658,08-02-2) (LNW): -->Hot in Dave-pulse and "high occ" and "high mult" in YMon. **Access DONE**: -->Spy: OD angle bracket unsoldered. Noise gone after resoldering. -SL8-116 to118-most channels (A,192,13-19-0&1) (USE): -->Extremely hot cells (worst in chamber). These have been worked on extensively with no luck. -->BACKGROUND: Put ferrite core on flat cable at TDC input with no effect. -->5/14/01: Microcoax well seated and latched, ASDQ angle brackets well solder and screws tight, lateral staggering of ASDQ pins had no effect on noise, HVDB and HVDB Fcage grounds fine (adding Cu tape had no effect). Last gasp is adding ferrite core to SL8 quad A HV ribbon, but we have never noticed any improvement in upper quadrants by doing so in the past. Removing ASDQ Fcage seemed to have no effect on the noise. This is the noisiest area on the YMon list, and it does not want to budge. -->5/15/01: Tried ferrite core on quad A SL8 HV ribbons only - no effect on noise. Removed core. -->TRY?: This is at the end of one of the longest DAC voltage runs. I believe one can bump up the DAC voltage a bit by removing (or reducing) a series resistor? Also could try replacing the ASDQ. **Access so far**: -->Spy: Observed noise while checking HVDB end. Screws all tight and Faraday cage resecured. Noise stll bad (it was worse when F-cage removed). -->Spy: Checked ASDQ DB angle bracket and screw connections for 116/117 -- ok, but tightened screws. Note that noise increased when F-cage installed (still there at lower level without F-cage). -->Installed a spare ASDQ DB in 116/117, but things looked about the same. Here again saw significantly less noise when F-cage off. -->At Morris' suggestion, "Bumped up" effective DTH for 116/117 by tieing a 1 kohm resistor from +3v to chip side of 10 ohm series resistor from DTH: 3v - 1kohm - chip - 10 ohm - DTH. This effecitively gives an extra 30 mv boost to DTH (225 mv to 255 mv) for cells 116/117. -->Spy: With the boosted DTH, no noise was observed on 116/117. I could see a general reduction of pulseheight with the narrow pulses we are using. I briefly looked at cell 118 and could see no noise on this cell with the F-cage off (I had not previously checked it during this access). **Things to do** -->Check that the reduced pulseheight in cell 116/117 is consistent with raising DTH from 225 to 255 mv by doing so explicitly in a "normal" cell. -->Check to see if we can get away with less than a 30mv jump in DTH (unfortunately alot of work). -->See what is necessary to do for cell 118 and 119 (possibly nothing). -->Likely leave F-cage off ASDQ end (We've done this in two other spots to reduce noise). -SL8-197-(8-11) (B,320,15-18-0) (USE): -->Hot in Dave-pulse and "high occ" in YMon. **Access so far** -->Spy: All channels quiet and normal looking -- TDC problem? -SL8-199-(9,10,11) (B,321,15-18-1) (USE): -->Dead in Dave-pulse and YMon. -SL8-271-(9,10,others) (C,429,16-16-1) (LSE): -->Noisy in Dave-pulse. "high occ" in YMon. **Access DONE** -->Spy: Unfortunately, F-cage was removed before I looked at the channels. At that point most channels in 271 were noisy. The HVDB screws were tight, and the F-cage was put back and secured after rebending tabs. All channels in 270 and 271 were then very quiet. -SL8-363-(0,1,3) (D,583,18-09-3) (LNE): -->Noisy in Dave-pulse. "high occ" and "high mult" in YMon. **Access DONE** -->Spy: Pressing on HVDB F-cage made noise come and go. Rebent tabs and resecured. Noise gone at the moment, but I'm not sure this is solid. ******************************************************************************* 3. "INSULATION WORK" -------------------------------------- BACKGROUND: East side: The inner cylinder area is such a rat's nest, that it was difficult to see any way to effectly add insulation. We decided to punt until October. Aseet was able to see 3 of the 4 ID temperature monitors (all but the one on the bottom north) and came up with a way to better assure that they are measuring the endplate temperature and not the ambient. First, the 592 was attached to the plate using CA (superglue), then it was covered with spray polyurethane foam. Slightly more foam made it onto the 592 than Aseet's fingers. West side: Aseet was again able to see only 3 of 4 592 temperature monitors on the COT endplate ID, so the upper north monitor was not insulated. The other 3 were insulated using the polyeurathane foam. These 592s were securely attached to the endplate with RTV, so CAing was unnecessary. ACTION: I recommend skipping the mylar trube, but attempt to install the bubble insulation of the endplate ID. Because things are such a rat's nest, we will need that advice of silicon cabling experts as to what can and cannot be moved. ******************************************************************************* 4. SURVEY and possibly move COT: ------------------------------------------------------------- In order to get the SVT "in range", the SVX has to move up by 2 to 4mm. If it is impossible to move the SVX, we may have to move the COT. Sounds like move the enitre central detector is completely impractical. ******************************************************************************* Possible Insulation projects: -------------------- 1. Al/bubble wrap around ID. Kourosh bought a new roll, 2' x 25'. ID of COT inner cylinder: 31.756", pi*ID=99.764"=8.3'. This would essentially "complete" the insulation of the COT can: 1) Extrusion + FCage + ASD heat on endplate surface, 2) Al/bubble wrap to ISL and annular pieces, 3) Al/bubble wrap and heaters around ISL). 2. Mylar tube to block convection (Kourosh?). Need OD of ISL (with wrap): 64cm + 2 * 1/4" = 25.2" + .5" = 25.7", ID of COT: 31.76". 1/2(31.76-25.7)= 3.0". If we oversize by 1/2" --> tube diameter would be 3.5". Butt and tape (wide Kapton) two 5' long tubes together during installation. 3. Pipe wrap (Mike?) - where possible and easy. -- have to unwrap when doing COT work below. Phone numbers --------------------- Ken (8630, LRP 722-0582, Home: 365-2241) Kourosh (8630, 5215, 978-1958) Karen (3377, 8630, Cell: 508-0117, Home: 365-5585) Michelle (3377, Home: 393-0861) Wanda (3377) Ines (3377) Dave Erikson (3366, 8630, LRP:266-8728): Wojciech (3366, Home: 630-783-9462 ) Tools and equipment required: -------------------------------------------- - Capmeter (2) and DVM (2) (check) - Two pliers for HVDB ribbon connectors (check) - Hex screwdriver for HVDB screws (check) - Spy for alternative connection checks (check) - Kapton tape (both 4" and 1") and copper tape?? - Ordinary needle nose (2), scissors, cutters (2), flashlights (2)??-- - Al/bubble wrap and mylar tubes (check). - Baggies and duct tape (check)