Miniproposals

Miniproposal: Date Filed: Title: First Author: Session Leader: Steve Wolfe (shots 1-22)

Engineering Operator Run Comment

continue restart and do ece alignment

Session Leader Plans

Physics Operators Plans

Session Leader Summaries

Entered: Jul 7 2004 03:49:09:773PM

Author: To Be Determined

Run Summary for 940511 Miniproposal 041 - CMOD Restart SL: wolfe PO: granetz EO: gwinn/fairfax Today's plan was to build on yesterday's successes by bringing up a corrected RCUR predictor (i.e. not depending on F23), and then adjust the inner and outer gaps and x-point location. If possible, we intended to get shots for the ECE alignment (MP 040), and maybe even get the piggyback part of Tinios' Zstep experiment (MP053). We actually accomplished the first objective, bringing up the new RCUR predictor. The first shot using the corrected predictor was #6. According to EFIT there were reasonable gaps on both sides; the shot lasted about 800msec, well into ramp-down, and terminated vertically. The other good shot was shot 10, which had RCUR programmed further in, with CLEARIN kept the same. This shot had slightly higher elongation, and a small long term stable oscillation on the Z position which finally went unstable during ramp-down. The run was terminated early when, after a long series of duds and fizzles, we realized that yet another inner wall flux loop, F22, had developed problems similar to F23 and F25. Specifically, it exhibited about a 0.1 to 0.2 Wb offset after the shot. This much error is clearly enough to give several mT of error on Br_0. It was decided to attempt to repair/replace the connector common to these loops. Problems: --------- In addition to the flux loop fiasco, there were a few other problems before and during the run. No ECDC was done because the RF interlock wouldn't close, despite the fact that all temperatures were in range. There was a hard ground fault between B-port and the Machine; this was temporarily "fixed" by unplugging the gage and pump on the gas-feed system. It was noted before the first shot that all shots yesterday had been run with the REF 5-8 switch set to REF_10 instead of REF_IP. There were several power supply faults and delays for fixing them through the course of the run. In some cases, it appeared that large demand signals from the hybrid were responsible for comm failure indications on OH2 supplies. There was also a long delay spent trying to diagnose an IOC on the TF during inversion. Scorecard: --------- Duds 8 (1 due to power supply fault) Fizzles 5 Plasmas 9 (only 3 over 600msec) ---------- total 22

Physics Operator Summaries

Session Leader Comments
May 11 1994 08:52:11:320AM		Steve Wolfe	Pre-run comments for 940511: MP#041 SL:wolfe PO:granetz EO:gwinn No discharge cleaning because of an RF interlock indication. There was a ground fault on B-to-machine; temporarily removed by unplugging the little pump and TC gage on the gas-feed system. Granetz and Horne found that at least some of the magnetics problems are due to a faulty connection in the Amphenol connectors. This is a similar problem to what Goetz had last year, and is apparently a known flaw in the Amphenol connectors! Discovered during set up: all day yesterday, the REF for 5-8 was set to REF_10 instead of REF_IP. This apparently happened at the first shot, as it was correct for the last shot on Friday. It shouldn't have mattered much, as all the wires were on waveforms close to zero. Run plan: Try starting from 940510017, which used all new algorithms except RCUR, and seemed to run well. Put new RCUR algorithm on wire 8, with zero controllers, and compare. Observe flux loops carefully for signs of further damage. Note: yesterday shot 17 had REF_10 on wires 5-8; we changed it to the correct REF_IP. As indicated, bring up new RCUR. Adjust position and x-point locations to taste. Continue clean-up of machine. Make sure ECE gets its alignment shoots, if possible. Test Gerry's zcur step, if we get to it.
May 11 1994 09:06:55:660AM	940511001	Steve Wolfe	Shot 940511001: Start from 940510017, as amended (see pre-run comments). P=1.1e-5 Result: Late breakdown (9.7msec) no current.
May 11 1994 09:24:08:050AM	940511002	Steve Wolfe	Shot 940511002: Br on previous shot (1) looked a bit negative, but leave it for now. Change pressure to 1.0e-5 and go again Result: Short shot; ends at 180 msec. PF breakers pulled out, PS fault at about the same time. There was light on the TV and 500ka, but there are flatlines on the Halpha and Z-meter. Apparently, Ben Welch was firing a test shot and it got a false trigger.
May 11 1994 09:37:04:410AM	940511002	Steve Wolfe	Shot 940511003: -------------- Try again. Result: Dud. Br looks negative. EF2's look funny right around zero.
May 11 1994 09:37:16:330AM	940511003	Steve Wolfe	Shot 940511003: -------------- Try again. Result: Dud. Br looks negative. EF2's look funny right around zero.
May 11 1994 10:00:11:150AM	940511004	Steve Wolfe	Shot 940511004: -------------- Raise Br offset from -.005 to -.004 Raise P from 1.0e-5 to 1.2e-5 Result: Short, about 180msec. PF breakers pulled out again. OH2L has a forced invert at 140-150 msec. That causes the problem.
May 11 1994 10:25:06:700AM	940511005	Steve Wolfe	Shot 940511005: -------------- Try again. The engineers have done something to the margin angle on OH2L Result: Good shot. Ran to 0.8 sec.
May 11 1994 10:42:23:670AM	940511006	Steve Wolfe	Shot 940511006: -------------- Raised the gas puff and put in a pulse for Luke Call Algorithm on RCUR on wire 1. Put offset back at 10000. Result: NIce shot. 800+msec Rcur is further in, but the outer gap isn't much different.
May 11 1994 10:42:46:820AM	940511006	Steve Wolfe	Shot 940511006: -------------- Raised the gas puff and put in a pulse for Luke Call Algorithm on RCUR on wire 1. Put offset back at 10000. Result: NIce shot. 800+msec Rcur is further in, but the outer gap isn't much different.
May 11 1994 10:53:32:870AM	940511007	Steve Wolfe	Shot 940511007: Still more gas puff Draw RCUR at .67 instead of .68 after .237 sec Result: dud
May 11 1994 11:02:41:650AM	940511008	Steve Wolfe	Shot 940511008: No changes. Result: Small breakdown at 10.5msec
May 11 1994 11:14:27:830AM	940511009	Steve Wolfe	Shot 940511009: Make Br offset 0.5 mT more positive Result: Faint flash on TV. Lasted until 50msec Current rise is slow (only 100kA by 50msec) and density is 3e19, compared to 1e19 on shot 6.
May 11 1994 12:27:47:470PM	940511010	Steve Wolfe	Shot 940511010: --------------- Back off on early gas puff. Long delay for a TF fault that occurred after inversion on prev. shot. Result: Good plasma Rxray=.668 Got an oscillation on ZCUR through the whole shot.
May 11 1994 12:45:48:360PM	940511011	Steve Wolfe	Shot 940511011: --------------- Change ZCUR(2) gain programming Bring RXL gain up to final level at 200msec Add gas puff after 150msec. Result: PLasma, but shorter. Disruption at 390msec Vertical instability
May 11 1994 12:53:43:210PM	940511012	Steve Wolfe	Shot 940511012: Put clearout on wire 8 for inspection. No change on the ZCUR situation, will probably do it again Result: Shorter? 180msec. Vertical osciallation again.
May 11 1994 01:16:36:010PM	940511013	Steve Wolfe	Shot 940511013: Move higher ZCUR integral gain later (.388 sec) Result: Same thing. We're getting bad oscillations on ZCUR early.
May 11 1994 01:25:56:290PM	940511014	Steve Wolfe	Shot 940511014: Kill derivative gain on ZCUR(2) until 100msec Get rid of step up in ZCUR (2) proportional gain. Result: dud Br0 looks positive now.
May 11 1994 01:41:04:350PM	940511015	Steve Wolfe	Shot 940511015: Change BR0 offset from -3.5 to -4.5mT Result: Short and xrays?? It still oscillates early, but makes it through that. Then it gets in trouble at 200msec. Note that ZXL comes on at 200msec and its gain is 10.0000
May 11 1994 01:57:35:710PM	940511016	Steve Wolfe	Shot 940511016: Delay ZXL gain until 300msec and reduce to 8.0 Result: Power supply fault before commutation.
May 11 1994 02:18:05:640PM	940511017	Steve Wolfe	Shot 940511017: Delay ZXL gain until 300msec and reduce to 8.0 (repeat of 16) Result: Fizzle. Breakdown at 6.8msec. fields look ok.
May 11 1994 02:31:29:760PM	940511018	Steve Wolfe	Shot 940511018: Drop the gas to 1.1e-5 Result: Another fizzle Late breakdown (18msec) .
May 11 1994 02:40:41:030PM	940511019	Steve Wolfe	Shot 940511019: Drop the gas to 1.0e-5 Result: Fizzle. Got some current, breaks at 11msec, dies at 30msec.
May 11 1994 02:52:34:910PM	940511020	Steve Wolfe	Shot 940511020: Reload shot 10, leave P=1.0e-5 Result: dud. Radial field is about +1mT
May 11 1994 02:59:53:910PM	940511021	Steve Wolfe	Shot 940511021: Change offset on Br0 from -3.5 to -4.5mT Result: another dud
May 11 1994 04:56:55:030PM	940511022	Steve Wolfe	Shot 940511022: Raise fill to 1.3e-5 Result: dud. There is a problem on F22 flux loop. Punt the run so the damned connector can be fixed.

Physics Operator Comments
May 11 1994 08:55:04:540AM	940511001	Robert Granetz	Re-loaded shot 17 from yesterday. Wire #08 was changed to the new RCUR predictor (to get around more magnetics problems). The PID gains are still zero however. The old RCUR is still on wire #01, and will be used for the old shot. Upon loading this shot, Steve Wolfe noticed that the 10 volt wavegen reference was turned on for wires 05-08 instead of the Ip reference. A more detailed examination showed that this occurred on the first shot yesterday, and remained that way for the entire day. No non-zero demands were programmed on those wires anyway, so it didn't matter. We have set wires 05-08 back to Ip reference for today.
May 11 1994 09:09:26:090AM	940511001	Robert Granetz	Fizzle (at 10 ms). Vacuum fields were close to shot 17 from yesterday, except that Br was 0.5-1.0 mT more negative. Next shot, change pre-fill from 1.1 to 1.0e-5. No other changes.
May 11 1994 09:26:14:120AM	940511002	Robert Granetz	Plasma, but short. Ipmax=0.53 MA, duration=0.185 s. No spectroscopy data (due to Ben Welch executing a test shot just prior to the plasma shot). The OH's faulted at the same time as the plasma disrupted, but Gwinn can't verify whether it was the cause. For next shot, no changes.
May 11 1994 09:38:38:560AM	940511003	Robert Granetz	A dud! Because of the bad flux loops, we can't run MFLUX reliably, so it's not so easy to diagnose the null. Next shot, increase BR by +1.0 mT by increasing the offset coeff. Also, raise the pre-fill from 1.0 to 1.2e-5.
May 11 1994 10:08:00:200AM	940511004	Robert Granetz	Plasma, but short. Ipmax=0.55 MA, duration=0.19 s. OH2L stops following its demand at 0.14 s, which caused loss of ZCUR, leading to the disruption. Next shot, no changes.
May 11 1994 10:27:55:480AM	940511005	Robert Granetz	Good shot. Ipmax=0.59 MA, duration=0.80 s. Density is rather low (nl=3e19), so for the next shot, raise the early pulse gas programming. Also change RCUR (wire 01) to the new predictor.
May 11 1994 10:46:03:420AM	940511006	Robert Granetz	Nice shot. Ipmax=0.50 MA, duration=0.815 s. Lots of moly injections. The z-centroid has a growing oscillation starting at 0.7 s, which eventually results in loss of vertical control. Density was still on the low side. For the next shot, increase the pulse gas some more. Also change the RCUR demand from 0.68 to 0.67 m.
May 11 1994 10:53:51:690AM	940511007	Robert Granetz	Dud. Fields look okay. Re-try with no changes.
May 11 1994 11:04:22:820AM	940511008	Robert Granetz	Fizzle (at 11 ms). Next shot: raise BR by 0.5 mT using the offset coefficient (from -4e-3 to -3.5e-3).
May 11 1994 12:17:28:600PM	940511009	Robert Granetz	Very short plasma. Current started to rise, but duration=0.05 s. Might be due to too much gas. For next shot, reduce the pulse gas programming.
May 11 1994 12:30:17:000PM	940511010	Robert Granetz	Good shot. Ipmax=0.61 MA, duration=0.77 s. The vertical position control becomes unstable after 0.7 s. For the next shot, for the slow ZCUR, reduce the P & D gain and increase the I gain. For the fast ZCUR, increase the D gain (P gain was already at 9). Also bring up the gain on RXL sooner, and increase the late gas pulse.
May 11 1994 12:47:27:600PM	940511011	Robert Granetz	Plasma, but it disrupted due to loss of vertical control at 0.4 s. ZCUR oscillation starts growing at 0.33 s. Steve Horne is investigating this problem. For the next shot, put CLEAROUT on wire 08, but don't use it. This is to check its output.
May 11 1994 12:57:44:560PM	940511012	Robert Granetz	Plasma, but even shorter. Ipmax=0.52 MA, duration=0.166 s. Problems with ZCUR control are getting worse. For next shot, delay increase in integral gain until 0.388 s.
May 11 1994 01:17:58:940PM	940511013	Robert Granetz	Another short shot. More problems with ZCUR. For the next shot, tweak the slow ZCUR gains again. Delay derivative gain until 0.1 s, and take away an unnecessary early blip in proportional gain.
May 11 1994 01:26:40:350PM	940511014	Robert Granetz	Dud. For next shot, reduce BR_0 by 1 mT using the offset coefficient (-3.5 to -4.5e-3).
May 11 1994 01:45:11:900PM	940511015	Robert Granetz	Plasma, but short. Next shot: delay ZXL turn-on from 0.2 to 0.3 s, and change gain from 10.0 to 8. (Note: remember we previously ran into a wrap-around bug when the PID gain is set to 10.0)
May 11 1994 02:12:38:680PM	940511016	Robert Granetz	Dud, due to OH2's faulting prior to commutation. Next shot: try again, no changes.
May 11 1994 02:26:53:020PM	940511017	Robert Granetz	Fizzle at 6.5 ms. Null is therefore probably okay, so for the next shot, just decrease the pre-fill from 1.2 to 1.1e-5.
May 11 1994 02:31:56:190PM	940511018	Robert Granetz	Fizzle again. Late breakdown, 17 ms. Next shot, reduce pre-fill from 1.1 to 1.0e-5. Pressure didn't drop between this shot and the previous one.
May 11 1994 02:42:13:700PM	940511019	Robert Granetz	Another fizzle. We're stumped. Next shot: re-load from shot 10 of today.
May 11 1994 02:53:18:590PM	940511020	Robert Granetz	Dud. BR was higher than shot 10 by about 1 mT. Drop the BR offset from -3.5 to -4.5e-3 for the next shot.
May 11 1994 03:01:00:380PM	940511021	Robert Granetz	Dud. Vacuum fields look okay. Next shot: raise pre-fill from 1.0 to 1.3e-5.
May 11 1994 03:25:49:480PM	940511022	Robert Granetz	Dud. It looks like another flux loop, F22, has become flaky. It's also on the same Amphenol connector as F23 and F25. Shut down the run, and prepare to remove, repair, and re-install the connector.

Engineering Operator Comments
Shot	Time	Type	Status	Comment
1	08:54:10:260AM	Plasma	Ok
2	09:09:39:530AM	Plasma	Bad	500 ka, oh2 com fault
3	09:28:25:130AM	Plasma	Ok
4	09:46:15:420AM	Plasma	Bad	550ka, oh2 com fail
5	10:09:39:690AM	Plasma	Bad	550 ka, ef1 com faul after shot
6	10:28:26:740AM	Plasma	Ok	600 ka
7	10:46:43:840AM	Plasma	Ok
8	10:56:26:560AM	Plasma	Ok
9	11:07:15:120AM	Plasma	Bad	ioc tf at inversion, 3&4 took all I for 1/2 cycle
10	12:10:56:440PM	Plasma	Ok	good 600 kA shot. no faults
11	12:30:53:790PM	Plasma	Ok	600 kA disruption at .4 sec
12	12:48:44:360PM	Plasma	Ok	Loss of plasma vertical position control; 500 kA disruption
13	01:06:49:680PM	Plasma	Bad	com fault oh2 u /l 550 ka ip
14	01:20:45:210PM	Plasma	Ok
15	01:31:45:660PM	Plasma	Bad	oh2 com fail, 570ka
16	01:45:22:980PM	Plasma	Bad	Appears that PF breaker tripped before commutation
17	02:12:46:990PM	Plasma	Ok	Fizzle. Reason for PF breaker trip on shot 16 remains a mystery. No fa
18	02:25:39:480PM	Plasma	Ok	Fizzle. All power supplies OK.
19	02:33:59:250PM	Plasma	Ok	Fizzle. TF crowbar at end of inversion.
20	02:46:38:950PM	Plasma	Ok	Dud.
21	02:56:03:930PM	Plasma	Ok	Dud.
22	03:06:14:340PM	Plasma	Ok	Dud. Another bad flux loop found. stopping run to effect repairs.