Miniproposals

Miniproposal: Date Filed: Title: First Author: Session Leader: Unknown (shots 1-19)

Engineering Operator Run Comment

Finish Power System Tests and Start Plasma Operation/Cleanup PO:Wolfe EO

Session Leader Plans

Physics Operators Plans

Session Leader Summaries

Physics Operator Summaries

Entered: Jul 7 2004 04:37:19:003PM

Author: Steve Wolfe

Physics Operator Summary for Thursday, Oct 9, 2003 Power Tests and Plasma Startup PO: wolfe EO: Cochran,Dekow,Parkin Engineering setup ----------------- overnight ECDC in D2, stopping 1-2 hours prior to the run. Run begins at 9:00 and ends at 17:00. Power system setup : Initially running Power Supplies under PLC control through completion of power testing. For first plasma, setup as on 1030715001 Gas setup - fill B-top with 6 psi of D2 Hybrid enabled fill B-side lower with 1 psi of Ar Hybrid enabled LEAVE B-side upper AS IS (should have He3) fill C-side with 30 psi of D2 Hybrid enabled The following gate valves should be enabled, assuming no vacuum problems: ECE, VUV, DNB RF setup: Transmitters 3 and 4 in RUN ON by 7:00AM Run plan: --------- After finishing the Power System tests, assuming no problems, we'll try for plasma. Starting shot is 1030717001 (1MA, 5.4T, SNB). Since the RF group plans to condition J-port in the morning, we will want to do 30minutes of ECDC before first plasma attempt (could be done between test pulses). If we get successful startup, we'll continue running boring cleanup plasmas while monitoring H/D, etc. It may be necessary to back off on the current to 800kA. RF Group may start conditioning of D&E if things are going really well. Results: -------- More problems than usual for a campaign-start day, but we managed to get one short ugly disruptive plasma on the last shot. Tomorrow, it's Bob's problem. The alternator had some sort of problem that delayed their coming up to speed, so we started with TMX tests. The EF2's ran fine. The correction coil supplies have been re-configured to have one master (the original) and one as a slave, using the gate drives from the master supply. The first test was run with only the master connected to the coils and the slave running into a 380Ohm resistor, so that it's voltage could be monitored. First shot was with PLC control. This was successful, although it later was noted that the voltage and current monitors had been given the wrong gains in the tree (factor of two). During this test, the cap bank charging was also successfully tested. When the alternator came up we switched to commutation tests. The oh's commutated correctly (shot#4). On shot 5 ef1l commutated but ef1u did not; the switch never closed. This repeated on shot 6, so we went on to test the TF's; the standard three steps all ran successfully, alhthough there was a delay after the first to diagnose and replace a bad digitizer (tr16). Returning to the EF1U commutation problem, it turned out that there was a loose limo connector from the j221 to the switch close trigger. EF1's both commutated correctly on shot 10. The first try at running the c-coil under hybrid control (also shot 10) failed because the PLC voltage was set to zero instead of 300V. For this and subsequent tests the 2000A waveform was copied into seg1 and moved to -1.5sec, so we could test with fizzle detector at 0.5 or 0.08sec. This worked (master only) on shot 11, and it was noted that the slave voltage was reading twice as high as it should. The shot was repeated on 12 with Bill watching the meter to verify that the supplies were giving the same voltage. The current came out at 1050A for a 2000A demand; voltage was 150V. On shot 13 the slave tmx was put in series with the master, and as expected the voltage on each dropped to ~75V; the current actually went up a bit to 1160A. It was then decided not to try raising the current because Bill was nervous about hitting the 3200A IOC trip, although the supplies are nominally not capable of going over 3000A; the PLC set point had been being put at 4000A because of the offset problem, but for the remainder of the run it was reduced to 3000. The first plasma attempt was shot 14. The hybrid did not put out proper demand on the OH2 supplies. The symptoms seemed consistent with the wavegen input to the a-matrix not getting through, but the digitized acquisition output was fine. After a lot of test shots and running back and forth to the interface room and swapping cables, I found that the ribbon cable from the A3 output to the A6 summation input was loose. Reseating this connector fixed that problem. However, the subsequent shot 15 was also a dud, and it turned out that the sparker was turned off and disconnected. For shot 16 the sparker was put back in service, and we got a fizzle. After a gas tweak and a br tweak, 17 was another fizzle; on this one the CII and Halpha signals were delayed by .26 sec, a timing glitch that I remember having seen before. Shot 18, after another drop in gas and increment on Br was a dud. For shot 19 I undid the Br tweak, ending at +.5mT (having started at 0). There was a long delay because cryo was seeing very unbalanced resistance temperatures on the EF4's (-150,-198) and the EF1U was reading -157. There were also some other funny looking readings. At 17:00 we decided to bypass cryo and take a last shot, which resulted in a good startup, but disrupted at 0.43 sec and caused the oh1 self-powered circuit to fire on the disruption spike. If this behavior repeats tomorrow it may be advisable to drop the program from 1MA to 600kA until the machine cleans up some. Shot 19 never got above 800kA and current was falling for lack of voltage before the disruption. The cryo resistance signals need to be checked. One speculation is that the ross relay contacts need polishing. I am also suspicious that the connections did not all get put back correctly after the single coil tests on 1030717. RF did not get a chance to do any J-port conditioning due to a combination of problems with the tuner and instrumentation. They will be trying again tomorrow. Ian's vacbrbz routine was failing during testing, and Tom discovered that the code included two instances of mdsclose calls with no arguments. Until this summer, contrary to the documentation and to the behavior on other platforms, this syntax on linux used to close only the most recently opened tree; this was fixed this summer to follow the documented behavior of closing ALL open trees, with the result that the vacbrbz routine was closing the analysis tree before it could mdsput its outputs. Tom edited the IDL routines (in Ian's directory) to make the mdsclose calls specific. Scorecard --------- Tests 13 Duds 3 Fizzles 2 Plasmas 1 ------ -- Shots 19

Session Leader Comments

Physics Operator Comments
Oct 9 2003 07:03:56:740AM		Steve Wolfe	Setting up for first (plasma) shot Load From 1030715001 Load/clean Load at 9-Oct-2003 07:02:33.00 Open tree /home/wolfe/pcs_scratch -1 Open tree done The first ten shots or so will be for completion of power system tests begun yesterday.
Oct 9 2003 09:14:38:937AM	1031009001	Steve Wolfe	shot#1 : ef2U -3kA ok
Oct 9 2003 09:22:40:830AM	1031009002	Steve Wolfe	shot#2 : ef2L test -3kA ok
Oct 9 2003 10:04:33:517AM	1031009003	Steve Wolfe	shot#3: Correction coil test; masrter only this time set for 300V, 1000A got 300A and 40V. Still has an offset issue. Signal on hybrid demand is wrong (tree mistake) Should probably continue these under hybrid control. Alternator is ready, so move on to commutation and tf tests. Come back to C-Coils later.
Oct 9 2003 10:13:05:203AM	1031009004	Steve Wolfe	shot#4: oh commutation test 10kA commutation test looks ok to me
Oct 9 2003 10:23:49:843AM	1031009005	Steve Wolfe	shot#5: ef1 comm test. ef1U did not commutate, and seems to have a small 60Hz oscillation Ian's vacbrbz routine is failing. Seems to have a problem doing the puts. It runs on alcserv2_submit. Code doesn't seem to have been changed. emailed twf and jas >>>>>>>>>>>>>>>>>>>>> 1031009003 >>>>>>>>>>>>>>>>>>>> Thu Oct 9 09:54:15 EDT 2003 IDL Version 6.0 (linux x86 m32). (c) 2003, Research Systems, Inc. Installation number: 101682-4. Licensed for use by: alcserv2.psfc.mit.edu % Compiled module: DO_VACBRBZ. % Compiled module: MDSOPEN. % Compiled module: MDSVALUE. % Compiled module: MDSCHECKARG. % Compiled module: MDSISCLIENT. % Compiled module: MDSIDLIMAGE. % Compiled module: VACBRBZ. Getting matrices from tree IAN % Compiled module: MDSSETDEFAULT. % Compiled module: MDSCLOSE. Getting data for shot 1031009003 % Compiled module: DERIV. % Compiled module: MDSPUT. % MDSPUT: %SS-W-NOMSG, Message number 0x00000000 % MDSPUT: %SS-W-NOMSG, Message number 0x00000000 % MDSPUT: %SS-W-NOMSG, Message number 0x00000000 % MDSPUT: %SS-W-NOMSG, Message number 0x00000000 % Compiled module: MDSGETMSG. % MDSVALUE: %TREE-W-TreeNOT_OPEN, Tree not currently open % MDSPUT: %TREE-W-TreeNOT_OPEN, Tree not currently open % MDSCLOSE: %TREE-W-TreeNOT_OPEN, Tree not currently open % Compiled module: MDSSETEVENT. Elapsed 0:04.41 User 0.55 System 0.17
Oct 9 2003 10:42:26:030AM	1031009006	Steve Wolfe	shot#6: Try again after power cycling same behavior, switch on ef1u never closed. Move on to tf tests.
Oct 9 2003 10:57:27:093AM	1031009007	Steve Wolfe	shot#7: tf 50kA test tf current ok, alternator ok as far as the traces I can see next shot: 100kA
Oct 9 2003 11:29:21:280AM	1031009007	Steve Wolfe	shot#8 is delayed to check on a bad digitizer in the TF crate
Oct 9 2003 12:09:31:330PM	1031009008	Steve Wolfe	shot#8: 100kA tf pulse ok
Oct 9 2003 12:22:59:733PM	1031009009	Steve Wolfe	shot#9: 150kA tf pulse looks good alternator seems happy Now they need to fix the ef1u commutation problem. It has been determined that the trigger is not getting through to it.
Oct 9 2003 12:59:20:687PM	1031009010	Steve Wolfe	shot#10: EF1U commutation problem seems to have been a loose cable from the J221. Try again after Gary plugged it back in. Also, fire correction coil at -1.5 off before 0, hybrid control (seg1) master only, slave into dummy load ef1's commutate ok. C-coil didn't go. PLC voltage was set at 0, needs to be 300V next shot: c-coil only
Oct 9 2003 01:18:44:907PM	1031009011	Steve Wolfe	shot#11: fire correction coil at -1.5 off before 0, hybrid control (seg1) master only, slave into dummy load OK, got 1050A (asked for 2000) but it was in regulation. next shot: There was a snafu on the slave voltage calibration. Bill will watch the meters to be sure it is doing the right thing
Oct 9 2003 01:26:46:860PM	1031009012	Steve Wolfe	shot#12: enable gas system; repeat of c-coil test Big gas puff, we had the f.d. still at 0.5sec Bill reports identical voltages on the meter. next shot: series the tmx slave for the c-coil. put the f.d. back at 0.08
Oct 9 2003 02:18:25:610PM	1031009013	Steve Wolfe	shot#13: series the tmx slave for the c-coil. put the f.d. back at 0.08 gas looks more reasonable. Got 1160 A at 75V on the master Hold off on the high current test for now. Bill is worried about what will happen if it hits the IoC. next shot: Go for plasma
Oct 9 2003 02:38:06:767PM	1031009014	Steve Wolfe	shot#14: Remove the c-coil program in seg1 and Go for plasma Turn the check magnetics back on in magnetics tree This will be like 1030715001 except the taps are in +dtop -dbot -jtop +jbot and plc limit at 3000A Dud - no oh2's
Oct 9 2003 02:44:25:640PM	1031009014	Steve Wolfe	shot#14:Dud - no oh2's Looks like the hybrid had a problem.
Oct 9 2003 03:08:37:373PM	1031009014	Steve Wolfe	shot#14:Dud - no oh2's Looks like the hybrid had a problem. It looks like the P10 wavegen output is not making it into the a-matrix. I see it on the acquisition output, but the a-matrix output indicates it isn't there. push on the connectors and try a test shot #43; same thing. Actually, it's worse. On shot 43 it looks like the aouts for oh2u, ef1u, ef1l and the ef2's are also missing. They're all there on the acq outputs. Try replacing the ribbon cable. The P outputs to the a-matrix go through an attenuator, so I can't just swap it.
Oct 9 2003 03:17:49:640PM	1031009014	Steve Wolfe	shot#14:Dud - no oh2's Looks like the hybrid had a problem. Digitizing the output of the wavegen attenuator #44 It looks like the attenuator outputs aren't attenuated? No, they're down by a small factor but they're all there. Could be a cable issue. Try connecting a new cable to the hybrid input
Oct 9 2003 03:44:18:820PM	1031009014	Steve Wolfe	shot#14:Dud - no oh2's Looks like the hybrid had a problem. Put in a new cable, acq cable back where it started- shot 45 Still screwed up. Put one of the new t cable on and digitize with it connected to the hybrid - shot 46. Now there are no sensible outputs from the hybrid, but the wavegen outputs are there. The digitizer looks ok, except that now it is down by a factor of ten. I think the attenuator may just be a resistor. Try again - I found the a3 output cable was loose That looks good. Put the cables back the way they were. shot#48- That works. Lock it up and take a shot.
Oct 9 2003 03:58:11:907PM	1031009015	Steve Wolfe	shot#15:Another try for plasma Dud. Got the power supplies this time. Got more gas than I expected, but that shouldn't cause a dud. Br is a little negative, rcur is driving in. No sparker. No voltage no current; must be off. next shot: no changes if the sparker can be turned on.
Oct 9 2003 04:05:54:557PM	1031009016	Steve Wolfe	shot#16:no changes except the sparker is back on and reconnected to c-port fizzle; moving in, br still too negative. next shot: drop gas, tweak br
Oct 9 2003 04:21:04:800PM	1031009017	Steve Wolfe	shot#17:drop gas from 31->26, tweak br from 0 to 0.5mT fizzle. The halpha and CII spikes are at .26sec - timing problem. next shot: drop gas and raise br again
Oct 9 2003 04:33:07:787PM	1031009018	Steve Wolfe	shot#18:drop gas from 26->22, tweak br from 0.5mT to .001T dud! Br back to 0.5
Oct 9 2003 05:22:21:533PM	1031009019	Steve Wolfe	shot#19: Br back to 0.5 bypass cryo because the ef4 resistance readings look crazy and the ef1u is reading -157 and won't go lower after a 30min delay; could be dirty contacts or screwed up inputs. short plasma, about 430msec oh1 self-power circuit fired after disruption Happens right after crossover. nice current rise, H/D=0.2; junk starts coming in around 0.3sec, current only got to 800 and turns around starting at 0.4. Probably just needs to clean up. Maybe should try a few at 600kA instead of trying for 1MA.

Engineering Operator Comments
Shot	Time	Type	Status	Comment
1	09:10:05:093AM	Test	Ok
2	09:19:03:983AM	Test	Ok
3	09:52:22:780AM	Test	Ok
4	10:09:59:047AM	Test	Ok
5	10:20:11:657AM	Test	Bad	EF1U Switch Never Closed
6	10:39:37:937AM	Test	Bad	EF1U Commutation Problem
7	10:48:32:030AM	Test	Ok
8	12:01:02:157PM	Test	Ok
9	12:14:42:623PM	Test	Ok
10	12:47:28:313PM	Test	Ok
11	01:00:22:170PM	Test	Ok
12	01:21:50:470PM	Test	Ok
13	02:05:38:313PM	Test	Ok
14	02:33:46:140PM	Plasma	Bad	Hybrid Problem
15	03:46:23:480PM	Plasma	Bad	No Sparker
16	04:02:13:207PM	Plasma	Ok
17	04:15:16:713PM	Plasma	Ok
18	04:28:31:860PM	Plasma	Ok
19	05:04:57:477PM	Plasma	Bad	OH 1 self powered circuit fired due to disruption.