Miniproposals

Miniproposal: 147 Date Filed: 7/15/1996 Title: Enhanced Startup Techniques (Higher Pre-charge and Parallel EF4) First Author: Ian Hutchinson Session Leader: Steve Wolfe (shots 1-37)

Miniproposal: 159a Date Filed: 7/29/1996 Title: Restart Operation: August 1996 First Author: Steve Wolfe Session Leader: Steve Wolfe (shots 1-37)

Engineering Operator Run Comment

Enhanced Startup Hutch/Wolfe/Silva

Session Leader Plans

Physics Operators Plans

Session Leader Summaries

Entered: Jul 7 2004 03:49:15:850PM

Author: To Be Determined

Session Leader Summary for Tuesday, 960806 MP#159A and 147 (higher pre-charge) SL: wolfe PO: hutch EO: Silva/Daigle The PEI was fixed this morning, and two hours of ECDC were carried out prior to the run. The plan for today was to continue trying to get plasmas at the standard pre-charge, then go to higher pre-charge as called for in MP147. However, after two fizzles, we abandoned the standard pre-charge and went to high pre-charge. It turned out that we did not need to change the timing of the power supply starts. All of the supplies ran fine at the higher pre-charge, once the PLC current limits were straightened out, and the commutaion bank voltages were increased for the higher current. That was the good news. The run was a real struggle. After lots of tweaking (see physop report), shot 12 gave a 600kA shot similar to 960805020, but starting from the higher pre-charge. However, the next five or six shots were all duds, with various amounts of tweaking. Reloading shot 12 (at shot 19) gave a fizzle. More fizzles followed. Shot 25 was a 300kA plasma, very disruptive. Shot 26 was a dud, at slightly lower pre-fill. The duds were hard to understand, as the nulls looked ok, and we tried raising and lowering the pressure. Raising the fill again gave fizzles, then shot 32 was a 500kA plasma. The next shot was a fizzle (no changes) On shot 34 we changed to parallel ef4 operation. The circuit has been changed to move the ground point to the other side of the coils, in an effort to eliminate the TF crowbars we had on Friday. This was successful. However, only duds were obtained as we never got the ef4 current program right (shots 34-36). Ian noticed a discrepancy in the mflux plots, which he attributes to a problem with the processed flux values at FLUX_15 and FLUX_14. These were being calculated from sums of several full and partial loops, because F16 had been flakey. Granetz changed the definitions to use the F16 loop, which is now believed to be ok. We are clearly not ready to proceed with runs that require plasmas. Therefore, tomorrows run will be changed to a continuation of MP159A amd 147. Overnight ECDC will be carried out tonight, and hopefully cleaner plasmas will be obtained tomorrow.

Physics Operator Summaries

Entered: Jul 7 2004 04:36:56:380PM

Author: To Be Determined

Physop Summary 960806. MP 147 Enhanced Startup PO Hutch SL Wolfe Summary _______ Despite a depressing day, suffering from a very dirty machine because of the lack of cleaning, significant parts of the enhanced startup tasks were accomplished. (1) Three plasmas up to 500kA were obtained (and many fizzles) with higher ohmic precharge. This is as good a record as we have had with the previous precharge this campaign. Robust current rise still requires far more cleaning for both scenarios. It is not clear if there is more field tuning required, but there may be, since breakdown was not very robust. Use shot 32 (or 33) as a model for the future. It was one of 5 in a row that had a breakdown. NO changes in precharge timing proved necessary because EF1 voltage capability proved to be at least 250V in real life. Good! Timing of cross-over and crowbar could not be addressed because the plasmas were so resistive and short. (2) Parallel EF4 with the higher precharge scenario was successfully run, albeit without a flash. This establishes that with the ground on the other side of the coils, the TF trip problem does not occur (3 shots out of 3). Therefore the way is open to complete the development of the parallel EF4 scenario. The trim resistor proved to be needed to help balance the currents. It lowered the imbalance from about 70A to about 10A out of 1440. The controller was adjusted to use the average current. Shot 36 should be used as a future model but needs the EF4 voltage programming to be lowered to account for the new configuration. (3) Some inaccuracies in the MFLUX reconstructions were found in the loop F16 region. These were substantially decreased by using the real F16 loop in processed flux as opposed to the one that was used for most of the run, which synthesized F16 from partials. (4) Breakdown occurred only when B_z0 was within about 1 mT of +.01T and perhaps benefitted when the field stayed near that value for a little while. Pressure range only .035 to .07 mtorr (cross-calibrated to MKS) has been explored (\edge::top.gas.ratiomatic:F_side). We are still not properly back into routine operation. More cleaning and startup tuning is needed tomorrow. Plan Was ________ ECDC started at 6am in deuterium. Plan: Start with the same settings as Monday. Get a sensible plasma. Once we are satisfied, set up with adjusted resistors: OH1 .033 ohms OH2 .006 ohms EF1 .070 ohms These are targetted to give slightly less total voltage with the planned new currents: OH1 -20.75kA, OH2 -27.75kA, EF1 -6.13kA, EF2 -2.06kA, EF4 -1.475kA. Therefore the commutation voltages should be unchanged: OH1 800V, OH2 1100V, EF1 1000V. PLC current limits should read: EF1 U/L [-7.5,5.] Assuming this now means -7.5 precharge. OH1 [-27,20] OH2 U/L [-30*,26] * Note the increase. EF2 U/L 5.5 Assuming that this supply is configured negative. EF3 12 EF4 [-3.5,4.*] * Note increase, in preparation for new configs. We will start with timings unchanged. It is highly likely that these will have to be changed to allow the supplies to reach demand in the precharge. EF1s are likely to be the biggest problem. In the past we have had to play games to let these reach demand. Scorecard _________ Shot,Duration(s),Ipmax(MA) Shot,Duration(s),Ipmax 1 0.000 0.016 2 0.000 0.036 3 0.000 0.016 4 0.000 0.016 5 0.000 0.016 6 0.000 0.013 7 0.000 0.021 8 0.000 0.013 9 0.000 0.018 10 0.000 0.016 11 0.000 0.013 12 0.216 0.682 13 0.000 0.016 14 0.000 0.021 15 0.000 0.014 16 0.000 0.013 17 0.000 0.016 18 0.000 0.013 19 0.000 0.018 20 0.000 0.013 21 0.000 0.021 22 0.000 0.021 23 0.000 0.016 24 0.000 0.013 25 0.220 0.300 26 0.000 0.016 27 0.000 0.016 28 0.000 0.016 29 0.000 0.005 30 0.000 0.016 31 0.000 0.018 32 0.157 0.558 33 0.000 0.018 34 0.000 0.016 35 0.000 0.013 36 0.000 0.011 Finished summary for 960806 Total of 36 shots. 36 shots. 3 plasmas. 13 fizzles. 20 duds. Details _______ Setting up new PCS entries: EF1 Raised several points. Increased early gain. Copied upper new settings to lower. EF1 voltage to -150/-70 from -90/-50 during precharge. OH1 Raised two points. OH2 Raised two points. Had to enter a new factor because of "exceeding wavegen capacity" Set Factor to .00035 from .0004. EF2 Raised two points. Copied Upper to lower. EF4 Raised two points. B_r feedback gain ramped on slowly. Saved as shot 101. Cold start: $ cmod$models:[hybrid_tcl] @hybinit_dispatch. $ type /tail=100 mds$root:[syserr]shape_bitbus_server.log look for: rtnode -- data loaded, A matrix responds rtnode -- data loaded, M matrix responds rtnode -- data loaded, PID matrix responds Gas: PG4 B-top is prepulse: -.517 to -.500 PG3 K-side (alias A-side) is around 35V during first .2s. 1 Fiz B_r seems a little high, +.001T. p_0=.055 mtorr. Lower prepuff to 15ms from 17. 2 Fiz Got .05 mtorr. 35kA crash. Nearly went. B_r problem. Change resistors. Load new traces. Try for higher precharge. Measured resistances. Oh1 32.78, OH2U 5.74, OH2L 5.80, EF1U 69.5, EF1L 70.6. 3 Dud Currents obtained: EF1L/U -6052,-6018 EF2 -2074,-2126, EF4 -1434, OH1 -18932, OH2 26781, 27838. OH1 badly low but consistent with set gains and voltages. Got -250 V from EF1s. OH2 failed to commutate. Was driven up to 31.6kA but held there. OH1 gain increased to 2. OH1 programmed voltage to -100 from -50V. 4 Dud OH1 limited by PLC limits. 27,20 meant 20 limit in precharge. Set OH1 limits to 27,27. 5 Dud OH1 gives 20672A. Some B_r problems. EF1s to -6200A. B_r offset to .006 from .007. Prepuff to 19ms from 17. 6 Dud OH1 current very low. -11kA. Switch fault. Program OH1 V to +200V at 3ms. 7 Fiz OH1 switch worked but faulted. EF1s to -6300 OH1 V to +200V at 1ms. 8 PSD No EF2 or EFc. 9 Fiz Null looks good from viewpoint of field gradient. B_z too low and rising too slowly. Lower OH2 voltage to -5V from +6.5V at 5 and 10ms. Lower OH1 voltage to +150V from +200V at 5 (and to 120V at 10ms). EF4 to -1440 from -1475 A. 10 Fiz Earlier breakdown 5.5ms. OH2s to -10V, OH1 to +125V, prepuff to 15ms. EF4 to -1420A. 11 Dud Looks as if the EF4 correction was too much. EF4 to -1440. 12 Pla Plasma diverts. OH2s rail Lower OH2 to -20V at 5/10ms, and OH1. Add some gas at 40ms. 13 Dud Not obvious why. Perhaps a little too much B_z Add a lower point at 20 ms on EF4 to help it. Set EF4 V to -200V at -.1 to 0s. Should decrease EF4 by as much as 50A 14 Dud B_z effect was a bit too much, as anticipated. Set EF4 demand to -1420. prepuff to 17ms. 15 Dud No change in the EF4 current obtained. EF4 to -1410. Remove OH1V spike at 1ms. 16 Dud B_z oscillation is reduced. Raise puff to 20ms. EF4 to -1430, Voltage programmed ot -200 throughout. 17 Dud Puff to 13ms. OH1 V to 0 at 1ms. 18 Dud Depression really sets in. Reload 12 19 Fiz Carbon II =.06! z-meter .4, B_z is higher (~.011) at breakdown than the previous shots that were duds. add gas puff at 40ms. 20 Dud This shows that the breakdown is very flakey at this setting. EF2s to -2100 from -2060A. 21 Fiz CII .05, Breakdown at 6.1 ms. OH1V to 0 at 1ms. 22 Fiz Similar. CII .035. Density .04e20 OH1V to 75V OH2s to -20V at 5ms 23 Fiz CII .02, bd at 6.5ms EF4 V to -150V, P to -1410A 24 Dud Got 1414A. Keep PG3 at zero till 10ms to prevent early chop. EF4 V to -140V, P to -1440A. 25 Pla CII .03, long hesitation. Still getting an early chop. Ghastly resistive plasma till 0.2s. Fill .05 mtorr. Delay PG3 even more. Reduce puff to 13ms from 14. 26 Dud EF4V to -130V. puff to 14ms. 27 Dud Got .05mtorr. Noticed that there seems an error in mflux. In the f16 area. Wolfe says no obvious errors in the flux loop data. Need to check the matrix. prepuff to 19ms. 28 Fiz Got .065 mtorr. CII .035 29 PSD no power supplies. 30 Fiz Repeat. 31 Fiz Repeat. Slightly more current. 32 Pla 500kA. Faster rise: 450kA at .1s. Then goes totally resistive. 33 Fiz Repeat. CII .03, Zmeter .35. Parallel the EF4 for a test. 34 Dud No TF trip. PF trip. EF4L -1494A, EF4U -1421A. This was with no resistor. Remarkably open looking reconstructions. But B_z is clearly lower because the average EF4 is higher. Add resistor to the EF4s. Change the definition of the F16 processed signal to use the actual loop. It was using a construction that added the partials across the port. This may account for the the error in the signal at f16, which could be seen in the reconstructions and in a plot of the signals versus loop number. Difference is of order 30mW. 35 Dud No TF trip EF4L -1413, EF4U -1425A Edit EF4 predictor to use 0.5 each of U and L. Set EF4 demand to -1390A. 36 Dud Hardly any change. EF4V to 70V, to compensate for the parallelling. Change definition of f15. Two failures on slow sequence. Gave up.

Session Leader Comments
Aug 6 1996 07:56:57:060AM		Steve Wolfe	RUn 960806 MP#159A and 147 Start machine and higher pre-charge SL: wolfe PO: hutch EO: Silva/Daigle We will continue the program from yesterday, bringing the plasma up. Once this is accomplished, begin the higher pre-charge development of MP147. The PEI supply is now fixed. ECDC began at 06:00.
Aug 6 1996 09:17:16:600AM	960806001	Steve Wolfe	960806001- Same as last shot from yesterday (960805023) fizzle. All supplies worked. bd at 6.5msec.
Aug 6 1996 09:38:44:890AM	960806002	Steve Wolfe	960806002- reduce pre-puff. fizzle. crashes at 15msec
Aug 6 1996 09:52:42:250AM	960806003	Steve Wolfe	960806003- Go to higher pre-charge (IHH's decision). See physop report for new resistor and pre-charge currents. Timing unchanged. Looks like oh2's didn't commutate. dud.
Aug 6 1996 10:10:30:650AM	960806004	Steve Wolfe	960806004- Increase gain and add some feedforward voltage to OH1, which was 1400A short of target. dud. OH1 was hitting PLC limit, because the plus and minus direction are still confused. Change both limits to be |27|ka.
Aug 6 1996 10:25:22:900AM	960806005	Steve Wolfe	960806005- Repeat, with higher PLC limits. Also, raise OH1 comm bank voltage to 1000V from 800. dud. oh1 got to current this time. A lot of positive index, pretty awful null.
Aug 6 1996 10:34:26:610AM	960806006	Steve Wolfe	960806006- Change EF1 currents to -6200 from -6130. Change Br0 from .007 to .006 raise pre-puff by 2msec. Low oh1 current, only 13kA. Switch never closed?
Aug 6 1996 10:56:01:520AM	960806007	Steve Wolfe	960806007- Try it again (same as 6) Change EF1 currents to -6200 from -6130. Change Br0 from .007 to .006 raise pre-puff by 2msec. Got a flash. Got faults on the OH1 comm. circuit. It looks like it opened and crowbarred ok. Faults may be due to conduction failure circuit.
Aug 6 1996 11:04:28:460AM	960806008	Steve Wolfe	960806008- MOre EF1 PF crowbar. OH2U was first to go. No TMX supplies. No light on the PF breaker; bulb was burned out.
Aug 6 1996 11:23:08:110AM	960806009	Steve Wolfe	960806009-repeat 8 MOre EF1 got a flash . Null looks good, not enough vertical field.
Aug 6 1996 11:35:56:960AM	960806010	Steve Wolfe	960806010- reduce oh1 and oh2 voltages; decrease IC_EF4 by 35A. flash, not much current. bd at 6.5msec, cii at .04mW/cm2/ster still not enough bz.
Aug 6 1996 11:44:58:150AM	960806011	Steve Wolfe	960806011- more of the same. less gas. dud. To much vertical field. overdid the ef4.
Aug 6 1996 12:00:45:730PM	960806012	Steve Wolfe	960806012- back to -1440a on EF4. plasma. 600ka, .2sec. EF2's go to the current rail. very disruptive in seg2
Aug 6 1996 12:15:43:500PM	960806013	Steve Wolfe	960806013- tweak current rise and gas. dud.
Aug 6 1996 12:28:43:460PM	960806014	Steve Wolfe	960806014- Add some feedforward voltage to EF4 dud
Aug 6 1996 12:40:50:860PM	960806015	Steve Wolfe	960806015- Add some gas, more positive Bz. dud. fields don't look that bad. maybe a little bit too negative Br
Aug 6 1996 12:51:10:060PM	960806016	Steve Wolfe	960806016- tweak ef4 again. remove voh1 lurch another dud.
Aug 6 1996 01:01:28:390PM	960806017	Steve Wolfe	960806017- change ef4 voltage program. Add more gas no change to Br, which looks too negative to me. dud. null looks good. pressure to 1e-5 gage.
Aug 6 1996 01:07:54:810PM	960806018	Steve Wolfe	960806018- decrease pre-puff by 5msec! tweak oh1 voltage. Still a dud.
Aug 6 1996 01:18:48:710PM	960806019	Steve Wolfe	960806019- reload shot 12. fizzle. at last, some light!
Aug 6 1996 01:23:42:920PM	960806020	Steve Wolfe	960806020- Add gas at 40msec. another dud. It's real marginal
Aug 6 1996 01:38:14:750PM	960806021	Steve Wolfe	960806021- increase ef2 current. I have no idea why. fizzle. Null looks better. Still need faster Bz ramp.
Aug 6 1996 01:48:28:220PM	960806022	Steve Wolfe	960806022- more negative on OH1V at 0.0 fizzle.
Aug 6 1996 02:03:05:330PM	960806023	Steve Wolfe	960806023- more oh voltage tweaks. fizzle, slightly later (7msec) and lower Cii light
Aug 6 1996 02:11:53:700PM	960806024	Steve Wolfe	960806024- tweak some EF4 voltage, as before. dud.
Aug 6 1996 02:23:51:410PM	960806025	Steve Wolfe	960806025- Got a plasma back. Only 300kA., makes it to 200msec, but current is dropping with disruptions from 80msec on.
Aug 6 1996 02:34:59:130PM	960806026	Steve Wolfe	960806026- drop the pre-puff. delay pg3. dud.
Aug 6 1996 02:45:29:470PM	960806027	Steve Wolfe	960806027- put back pre-puff. change ef4 voltage. another dud.
Aug 6 1996 02:58:59:680PM	960806028	Steve Wolfe	960806028- raise pre-puff. Ian thinks the flux reconstruction is screwed up. I don't see any problems with the flux loops. fizzle. bd at 6.75msec, cii at .03.
Aug 6 1996 03:03:39:740PM	960806029	Steve Wolfe	960806029- repeat. magnet resistance fault. no 13.8kv supplies.
Aug 6 1996 03:13:04:230PM	960806030	Steve Wolfe	960806030- repeat. fizzle.
Aug 6 1996 03:21:51:450PM	960806031	Steve Wolfe	960806031- repeat. fizzle.
Aug 6 1996 03:37:30:540PM	960806032	Steve Wolfe	960806032- Got the plasma back. 500kA. 150msec. PF crowbar.
Aug 6 1996 03:41:19:820PM	960806033	Steve Wolfe	960806033- No changes fizzle.
Aug 6 1996 04:22:26:600PM	960806034	Steve Wolfe	960806034- Change to parallel ef4. (why not?) dud. pf crowbar. Got an 80amp imbalance in EF4u vs ef4l
Aug 6 1996 04:41:11:040PM	960806035	Steve Wolfe	960806035- put back the ef4 balance resistor. Granetz has changed the f15 definition in the tree to use the actual loop. dud. ef4 balance was much better.
Aug 6 1996 04:55:38:950PM	960806036	Steve Wolfe	960806036- Reduce ef4 demand to -1390. make ef4 predictor as average of ef4u,l. another dud.

Physics Operator Comments

Engineering Operator Comments
Shot	Time	Type	Status	Comment
1	09:10:31:050AM	Plasma	Ok	no faults
2	09:22:28:420AM	Plasma	Ok	ok no faults
3	09:46:57:980AM	Plasma	Ok	no faults
4	10:05:19:520AM	Plasma	Ok	no faults
5	10:17:00:340AM	Plasma	Ok	no faults
6	10:30:04:460AM	Plasma	Ok	oh1 switch no close
7	10:45:30:250AM	Plasma	Ok
8	10:59:40:830AM	Plasma	Ok	faults
9	11:14:05:340AM	Plasma	Ok	no faults
10	11:29:41:400AM	Plasma	Ok	no faults
11	11:41:37:000AM	Plasma	Ok	no faults
12	11:48:57:400AM	Plasma	Ok	no faults
13	12:09:58:250PM	Plasma	Ok	no faults
14	12:22:30:030PM	Plasma	Ok	no faults
15	12:35:05:670PM	Plasma	Ok	no faults
16	12:44:46:520PM	Plasma	Ok	no faults
17	12:54:58:770PM	Plasma	Ok	no faults
18	01:05:50:340PM	Plasma	Ok	no faults
19	01:12:56:560PM	Plasma	Ok	no faults
20	01:21:28:250PM	Plasma	Ok	no faults
21	01:30:58:310PM	Plasma	Ok	no faults
22	01:40:40:190PM	Plasma	Ok	no faults
23	01:49:53:550PM	Plasma	Ok
24	02:09:34:460PM	Plasma	Ok	No Faults.
25	02:17:38:590PM	Plasma	Ok
26	02:28:57:440PM	Plasma	Ok
27	02:36:47:790PM	Plasma	Ok
28	02:53:34:790PM	Plasma	Ok
29	03:01:37:800PM	Plasma	Bad	Magnet resistance fault.
30	03:07:54:100PM	Plasma	Ok
31	03:20:14:910PM	Plasma	Ok
32	03:27:28:490PM	Plasma	Ok	comm fault
33	03:39:05:540PM	Plasma	Ok
34	04:10:02:140PM	Plasma	Ok	ef4 par pf breaker out oh 1 comm fault
35	04:35:58:290PM	Plasma	Ok
36	04:44:44:850PM	Plasma	Ok
37	04:57:49:360PM	Test	Bad