| Miniproposals | ||||||||||
|
| Operators | |
| Session leader(s): | Steve Wolfe |
| Physics operator(s): | Earl Marmar |
| Engineering operator(s): | Bill Parkin,Bill Byford,Ed Fitzgerald |
| Engineering Operator Run Comment |
| MP#555 PF Error field evaluation and compensation |
| Session Leader Plans |
| Entered: Aug 6 2009 01:18:27:267PM |
| Author: Steve Wolfe |
Session Leader run plan for Friday, Aug 7, 2009 MP#555 PF Error field evaluation and compensation SL: wolfe PO: Marmar This experiment will use the acoil in two different configurations; engineering ops should be prepared to make the necessary tap changes. We will begin with the taps in the standard +Dtop -Dbot -Jtop +Jbot connection. Depending on results, we will probably change to +Bbot +Dtop -Gbot -Jtop about 1/3 of the way through the run. This experiment will require low density but will probably have lots of disruptions. While I would like to avoid runaways, I'm more concerned with disruption recovery, so I recommend D2 pre-puff in B-top and substantial PG3 puff in seg1. Analysis of single coil data from 1090702 indicated significant changes in the non-axisymmetric fields from several of the PF coils. Unfortunately, the analysis is somewhat ambiguous with respect to both the magnitude and orientation of the resultant field for typical shots. Two possible solutions have been loaded into the \analysis::top.non_axi.model nodes in analysis trees 1090702333 and 1090702888. The former indicates we should expect smaller amplitude (2/1) error field, but with orientation rotating from 2nd to 3rd quadrant during the flattop, which makes the feedback compensation problematic. Possibly the present standard acoil configuration (+Dtop -Dbot -Jtop +Jbot) will work later in the shot, if we can avoid a locked mode in the early phase. The latter model predicts similar or larger amplitude to the old calculation, but rotated by nearly 90 degrees, so that our normal compensation configuration will be less effective. A better match would seem to be the "rotated helical" configuration (+Bbot +Dtop -Gbot -Jtop). The major difference between these models is in the contribution of the EF3L coil; in one case this contribution is opposite to that of EF3U, so that the net error is small, while this cancellation is not present in the other model. The plasma phase of this miniproposal will provide additional data to help resolve this discrepancy. 1) Check intrinsic threshold [1-5 shots] We will begin by checking the intrinsic locking threshold on a 1MA, 5.4T, LSN ohmic shot. Previously we have found the theshold for locking to be in the range of nebar~1 to 1.5e20/m^3 (nl04~0.6 to 0.9e20). If the 1090702333 calculation is correct, we should expect to see a considerably lower threshold, by about a factor of two, or more. Start with any recent 1MA LSN shot, setting the initial target density to nl04=1e20/m^2, and program negative density after 0.7sec to allow the density to fall. The a-coil should be disabled for these shots. If a locked mode comes in early, then increase the early density and repeat. If no locked mode appears at the lowest accessible density, then we should probably declare victory, leave the acoils off, and quit, but we won't. Instead we'll try raising the current to 1.2MA and then to 1.35MA. Both models predict that ef3 and ef4 should generate significant error-field, so higher current should provoke a locked mode, and we can proceed to step 2. Most likely, the intrinsic threshold will be about where it used to be, 6< nl04 < 9e19 at 1MA. This could indicate either that nothing has really changed, despite the results of the measurements of 1090702, or it could indicate the 1090702888 model is valid. To find out, we need to turn on the acoil 2) [2-5 shots] Test feedback compensation using original 2003 model, standard A-coil configuration. Repeat the falling density shot with A-coil energized and B21_PROJ feedback turned on, with waveform at zero as usual. Observe whether locking density goes up or down. If the locking density decreases, then program a flat density just above the locking value and repeat using the same feedback. Observe locking density and time if any. Determine minimum flat density which does not lock using this feedback. 3) [2-5 shots] Test feedback compensation using 1090702333 model If the intrinsic locking density was significantly lower than 9e19, then test the 1090702333 model; otherwise this model is probably ruled out, and we can skip to step 4. With the acoil still in the old standard configuration (+Dtop -Dbot -Jtop +Jbot) set up the B21_PROJ feedback with keywords SHOT=1090702333,ACOIL="+DTOP -DBOT -JTOP +JBOT". Set the target waveform to zero (as it normally is). See whether the locking density goes up or down. If it goes down, run a flat density just above the locking threshold and observe the time history, then reduce density and repeat. Note that it may be necessary to keep the density higher early in the shot before the feedback turns the acoil on. 4) [2-5 shots] Test feedback compensation using 1090702888 model If the intrinsic locking density is nl04 < 4e19 then this model is probably ruled out, but may as well try the alternate orientation anyway. Change A-coil tap configuration to [+Bbot +Dtop -Gbot -Jtop], and re-program B21_PROJ feedback, using keywords SHOT=1090702888, ACOIl="+BBOT +DTOP -GBOT -JTOP" . Be sure engineers update the patch panel configuration in the Engineering Control widget. Then repeat falling density shot and observe change in locking density. If the locking density is lower than the intrinsic value and the minimum observed with the earlier feedback programming, then program a flat density just above the observed locking value and repeat with the same feedback. Determine the flattop density at which locking is avoided for the full flattop. 5) [2-5 shots] Assess variations to feedback Using the best feedback programming from steps 2-4 (which might involve a change in the A-coil patch panel), program non-zero target values for the B21_PROJ. Start with targets of +.0002T and -.0002T, and vary to reduce the minimum locking density, if possible. 6) [3 shots] Assess compensation algorithm at higher current. If not already done in step 1, raise current to 1.2 or 1.3MA and repeat falling density shot with best feedback. |
| Physics Operators Plans |
| Entered: Aug 6 2009 04:58:38:670PM |
| Author: Earl Marmar |
| Engineering setup for Friday, August 7, 2009
MP#555 PF Error field evaluation and compensation SL: Wolfe PO: Marmar Run begins at 09:00 and ends at 17:00 Power systems as on: 1090806031 Acoil: +Dtop -Dbot -Jtop +Jbot (standard) Hybrid Enabled Gas setup: fill B-Top with 6 psi D2 Hybrid enabled (PG4) fill B-side lower with 1 psi Ar Hybrid DISABLED (PG1) leave B-side upper as is Hybrid DISABLED (PG2) fill B-main (C-side) with 40 psi D Hybrid enabled (PG3) leave NINJA as is Hybrid DISABLED Enable gatevalves and shutters: ECE, VUV, Z-bolo Torvac gatevalve toggle (yes/no): no Boronization(yes/no): no Overnight ECDC (yes/no): YES ICRF(yes/no): no LH(yes/no): no DNB(yes/no): no Cryopump (yes/no): no Vessel temperature: 60/60/60 overnight ------------------------------ ECDC Parameters ------------------------------ gas and pressure: D2 at 2e-4 Torr sweep: 44/45/103 cm scan: 20/120 s ______________________________________ During the run, we are likely to want to change the A-coil configuration at least once. See SL_PLAN entry for details of the run plan. |
| Session Leader Summaries |
| Entered: Aug 7 2009 06:38:32:633PM |
| Author: Steve Wolfe |
Run summary for Friday, August 7, 2009
MP#555 PF Error Field Evaluation and Compensation
SL: wolfe
PO: Marmar
EO: Parkin, Byford, Fitzgerald
The goal was to evaluate the possible changes to the error field
compensation feedback implied by the analysis of the single coil
tests performed on 1090702 versus the old prescription, based on
the results of similar tests performed 1030717 ("r3d" model).
Two candidates were tested ("333" model and "888" model), based on
different truncations of the SVD fitting of the 1090702 data, as
described in my logbook entries for that run made earlier this week.
We successfully completed most of the run plan, and conclude that
the original compensation algorithm is still the best of those tested.
We also discovered a long-standing bug in the B21_PROJ routine which
leads to a negligible (< 10%) error in the original configuration but was
causing a significant error in the testing of the "888" model.
The first part of the experiment confirmed that the intrinsic locking
threshold density is very similar to the 2003 value for 1MA LSN
plasmas, namely nebar~1.2e20/m^3. This result was inconsistent with one
of the new models, which predicted a lower threshold density, but
consistent with the other which predicted error field of similar magnitude
but different orientation, as well as different sources, i.e. contributions
from different PF coils.
The results of the active tests were that the original feedback prescription
is superior to either of the newer candidates for standard 1MA plasmas,
based on the minimum achievable density. We did not manage to test the
compensation to as low densities as have been run in the past at 1MA,
and so could not determine whether some modification would be indicated.
We did succeed in avoiding locked modes down to nebar~8e19/m^3 (shot#8)
with the original programming, and later down to nebar~6e19/m^3 with the
same programming but targeting a B21_PROJ=+.0001T. These bounds were not
set by appearance of locked modes but by our inability to get the density
down to lower levels. The tests using compensation based on the other models
resulted in locked modes at higher densities. In the case of the "333"
model the feedback did not turn on the a-coil and we essentially hit the
intrinsic threshold. The other algorithm using the +Bbot +Dtop -Gbot -Jtop
a-coil configuration gave substantial a-coil demand and resulted in
locked modes with fields that the "888" model predicted to be < 1G.
Raising the current to 1.2MA did result in a locked mode with the
old compensation algorithm (Shot#27), at the rather high density of
nebar~8.5e19/m^3 and an inferred |b21|~1G based on the r3d model. This
locked mode appeared around 0.55sec, just into flattop. A variation (Shot#29)
of the compensation to target a projected b21=+.0002T, i.e. higher A-coil
current, resulted in no locked mode, although the shot died from an injection
at .86sec; the density at 0.55sec was similar to #27, but then came
back up to 1.0e20/m^3 for the remainder of the flattop. This may or may
not represent an improvement. Some modification to the model predictor
may be indicated, but it is not yet clear which terms should be modified.
In the course of testing the "888" model, which required changing the
a-coil configuration to +Bbot +Dtop -Gbot -Jtop, which has a toroidal
phase of 76degrees, we discovered a bug in the B21_PROJ algorithm.
The error was forming the predictor using a weight for the acoil current
accounting for only the real_part (cosine component) of the acoil field,
which is amounts to a factor of .93 for the normal 21 degree orientation,
but is a factor of four down at 76 degrees, causing the feedback to
drive the acoil to very high currents and over-compensate the predicted
error. It took me an hour delay to figure out what was happening,
and we corrected it by manually editing the B21_PROJ predictor (shot#17)
to use the approximately correct weight 3e-7 for the acoil current.
I have now modified the nonaxi_smw procedure, but not yet uploaded it
to CVS or downloaded the revision to the live directory. I'll take
care of that before the next run, but since it makes a negligible
contribution in the standard configuration, it shouldn't be necessary
to re-call the predictor for this reason.
|
| Physics Operator Summaries |
| Entered: Aug 7 2009 05:29:38:673PM |
| Author: Earl Marmar |
| The machine ran well today. Most shots had 1 MA disruptions, and the last 2 plasmas (27 and 29) were 1.2 MA disruptions. In spite of this we had only 1 fizzle. Early hards were a problem from time to time through the day.
There were many changes to the a-coil programming, which was the point of the miniproposal; physops need to be aware of this if they plan to import any of these shots for future use. There was one shot lost to alternator vibration during check, and one to no gas (no high voltage during pulse). There was one small delay due to data acquisition, and a couple of delays due to apparent high temperature readings on EF3L (most likely due to the Ross relays). OH coaxes behaved well. Scorecard: Shot,Duration(s),Ipmax(MA),Result Shot,Duration(s),Ipmax,Result | 01 0.159 0.553 Plasma | 02 1.627 1.004 Plasma | | 03 1.604 1.003 Plasma | 04 1.540 1.040 Plasma | | 05 1.988 1.009 Plasma | 06 1.999 1.012 Plasma | | 07 0.000 0.003 Dud | 08 1.576 1.012 Plasma | | 09 0.324 0.805 Plasma | 10 0.665 1.090 Plasma | | 11 0.610 1.006 Plasma | 12 0.686 1.063 Plasma | | 13 0.727 1.081 Plasma | 14 1.175 1.056 Plasma | | 15 0.484 0.998 Plasma | 16 0.747 1.046 Plasma | | 17 0.754 1.057 Plasma | 18 0.803 1.025 Plasma | | 19 1.173 1.083 Plasma | 20 0.971 1.082 Plasma | | 21 1.324 1.039 Plasma | 22 0.314 0.800 Plasma | | 23 0.890 1.037 Plasma | 24 0.000 0.048 Dud | | 25 1.564 1.012 Plasma | 26 0.791 1.055 Plasma | | 27 0.716 1.304 Plasma | 28 0.120 0.055 Fizzle | | 29 0.876 1.271 Plasma Date 1090807: 29 shots. 26 plasmas. 1 fizzles. 2 duds. 0 System Error. (Plasma = Ip gt 60.0 kA Dud = H_alpha lt 0.40 V) Total plasma duration: 26.21 s (exceeding -55 kA) |
| Session Leader Comments | |||
| Aug 7 2009 09:02:18:497AM | 1090807001 | Steve Wolfe | Shot#1: Intrinsic threshold determination very short runaway disruption. |
| Aug 7 2009 09:23:23:123AM | 1090807002 | Steve Wolfe | Shot#2: Intrinsic threshold determination plasma into rampdown, density doesn't drop very far Locked mode in rampdown at 1.56sec, nebar=1.4e20 (nl04=7.5e19) Nebar only gets down to ~1.5 in flattop. Next: increase clearin 2mm, let the walls unload to try to get more density drop in flattop. |
| Aug 7 2009 09:34:05:163AM | 1090807003 | Steve Wolfe | Shot#3: Intrinsic threshold determination increase clearin 2mm, let the walls unload to try to get more density drop in flattop. Very similar, runaways down, still not much density drop. locks at 1.52sec, 1.36e20/m^3 Next: drop initial density to 8e19 |
| Aug 7 2009 09:50:15:550AM | 1090807004 | Steve Wolfe | Shot#4: Intrinsic threshold determination drop initial density to 8e19, since threshold is below that. Locked mode at 1.28sec, 1.2e20/m^3, in the expected nominal range. Clear signatures on C-H LM loop diff and bp26de-jk at 1.25sec Nominal error field is 5.4G, Model 333 gives 1.6G, Model 888 gives 5.1G I think this probably rules out the 333 model, but we'll proceed to test anyway. Next: step 2, turn a-coils back on, normal settings, standard taps. |
| Aug 7 2009 09:59:18:163AM | 1090807005 | Steve Wolfe | Shot#5:test 2003 feedback compensation algorithm. plasma full length, hards to 2e13. density never gets below nl04=7e19 oscillation on halpha after 0.7, ece sawteeth change period, but no obvious locked mode. Old model says |b21|< 1G, model333 claims overcompensated up to 4G, Model888 says up to 6G. This is looking like the original model is doing better than either of the new ones. Need to get the density down. Drop the initial density to 6e19. |
| Aug 7 2009 10:10:59:800AM | 1090807006 | Steve Wolfe | Shot#6:test 2003 feedback compensation algorithm. Drop the initial density to 6e19. plasma full length density at 6e19 but doesn't drop during the pulse. No locked mode. Need to drop density; where's the cryopump? |
| Aug 7 2009 10:23:30:653AM | 1090807007 | Steve Wolfe | Shot#7:test 2003 feedback compensation algorithm. Drop the initial density to 5e19. Alternator alarm inhibit. No power shot. |
| Aug 7 2009 10:57:40:727AM | 1090807008 | Steve Wolfe | Shot#8:test 2003 feedback compensation algorithm. Drop the initial density to 5e19. disrupts at 1.57sec, injection at 1.15sec. No locked mode, density gets down to 8e19, then comes back up to 1e20m^-3 Having trouble getting down far enough to see anything; maybe should have ecdc'd in He. Next: try again, lower target |
| Aug 7 2009 11:12:14:970AM | 1090807009 | Steve Wolfe | Shot#9:test 2003 feedback compensation algorithm. Drop the initial density to 3e19. early disruption, early hards 8e13 looks like an injection. Try again. |
| Aug 7 2009 11:26:25:937AM | 1090807010 | Steve Wolfe | Shot#10:test 2003 feedback compensation algorithm. initial density to 3e19. Lots of runaways disrupts at .668sec possible locked mode between 0.4 and 0.45sec, looks like it stays locked; may actually start before 0.4sec. Magnetics signature at 0.380, nebar~6e19/m^3. Not a good data point, since we're not up to current yet and it's got lots of runaways. Try for one at nl04=4e19, then move on. |
| Aug 7 2009 11:37:26:273AM | 1090807011 | Steve Wolfe | Shot#11:test 2003 feedback compensation algorithm. Try for one at nl04=4e19 disrupts at 0.6sec, hards way down No locked mode, maybe something at 0.4sec that goes away Move on to step 3 without finding the minimum density with the old model. See if the 1090702333 model does worse. |
| Aug 7 2009 11:52:55:597AM | 1090807012 | Steve Wolfe | Shot#12: Test 1090702333 model feedback compensation Try for one at nl04=4e19, reduced P to 4 on NL04. disrupts at .68sec Locked mode at .39 to 0.4sec, similar to shot#10 Got no acoil current this early, so it's not really a valid test The 333 model is claiming the phase is ~110deg, so no demand in this direction (demand is negative). Next: raise nl04 to 5e19, change D gain |
| Aug 7 2009 12:05:06:310PM | 1090807013 | Steve Wolfe | Shot#13: Test 1090702333 model feedback compensation Try for one at nl04=5e19, raise D gain to 5 on nl04 More runaways, got to .7sec, ran at 4e19 Locked mode at 0.56 or so, but no acoil demand (still negative). nebar=6e19/m^3. Next: back to original B21_PROJ feedback, see if it will run at this density, which is lower than #8, without locking. |
| Aug 7 2009 12:22:38:057PM | 1090807014 | Steve Wolfe | Shot#14: back to original B21_PROJ feedback, see if it will run at nebar~6e19, which is lower than #8, without locking. plasma to ~1.2sec, nl04=4.7e19 There is an injection followed by a possible locked mode right at the end, 1.13sec on the injection. Through 1.1sec I can set a bound of nebar>8e19 for the locking density. According to the old model |b21| < 0.5G, according to 1090702888 model |b21| ~ 5G@-22deg Go ahead and change the acoils to +BBOT +DTOP -GBOT -JTOP, and change B21_PROJ feedback to 1090702888 model. |
| Aug 7 2009 12:55:59:287PM | 1090807015 | Steve Wolfe | Shot#15: Move on to step #4, "BBOT +DTOP -GBOT -JTOP" change B21_PROJ to use 1090702888 model. Same density program. plasma to .45sec, acoil is on, not a locked mode. Injection kills it. The applied field is 8G, which I don't understand. The b21_proj claims it's got an error signal of < 1G, but the total field is 6G at 78 degrees, i.e. at the applied field angle. This implies that the intrinsic error is nearly ortho, but it's at -118deg, although it is over around -180 earlier in the shot. The B21_PROJ gain and the acoil enable both come on at 0.2sec, when the model claims the angle is -170degrees, which is in fact almost perp to the applied angle. That won't work. Delay P and I gains to 0.4sec, when the angle is at -115deg, which is a little better aligned. |
| Aug 7 2009 01:30:51:330PM | 1090807016 | Steve Wolfe | Shot#16: Move on to step #4, "BBOT +DTOP -GBOT -JTOP" change B21_PROJ turn on time from .2 to 0.4 sec. Same density program. plasma to 0.7sec, lots of hards, locked mode at ~0.62sec based on sawtooth change, but I don't see any magnetic signature. Acoil is nearly full on, b21_proj error is down below 1G again, but model-based total field is still high, around 8G, although the intrinsic field is coming out around 3G around -90deg. I'm doing something wrong! Somehow the coefficient on the acoil seems to be wrong. |
| Aug 7 2009 02:03:24:523PM | 1090807016 | Steve Wolfe | Shot#16: The problem is apparently with the Acoil coefficient that's getting picked up from the 109070288 tree. It is coming out as 7.47e-8 and it should be around 3e-7. I think it was probably created with the wrong patch panel configuration. Yes, the patch panel settings are taken from the engineering tree, and the value in engineering_1090702888 for \ENGINEERING::TOP.POWER_SYSTEM.CORRECT_COIL:PATCH_PANEL:CONNECTIONS is [0.,0.,1.,-1.,0.,0.,-1.,1.] instead of [ 0.00000, 1.00000, 1.00000 , 0.00000 , 0.00000 -1.00000, -1.00000 0.00000 No, I changed it and it didn't help; still getting a 7.47e-8. I think there's a bug in B21_proj. It's taking real_part of acoil_coef instead of abs(acoil_coef). That made a negligible error in the 21 deg case, but it's big for this version. Need to edit the nonaxi_smw routine. Not sure what the routine is doing is correct. Try manually editing the acoil coefficient to 3e-7 and run it again |
| Aug 7 2009 02:17:25:323PM | 1090807018 | Steve Wolfe | Shot#17: Retry with the 888 model, but with the coefficient on the acoil set to 3e-7 manually. I think the b21_proj algorithm has a bug. plasma to .75, locked mode at around .56 to .61sec, nebar=6.5e19 Now the model net field is under 1G, as it should be. The net field according to the old model is around 3G, which may make sense at this density as a locking threshold. Move the start time back to 0.2sec, to give it a fair comparison. It seems pretty clear that this compensation is not as good as the old one, although I'm still not sure the algorithm is correct. |
| Aug 7 2009 02:25:54:820PM | 1090807018 | Steve Wolfe | Shot#18: Testing 1090702888 model, with corrected (?) coefficient Start feedback at 0.2sec, programmed nl04=5e19. Still disrupts at 0.8sec Locks at .63sec, magnetic signatures on C-K and H-K, and on bp26_de-jk 888 model says total b21 is ~1G, old model says 3G. nebar=7.6e19. Next: raise nl04 program to 6e19 |
| Aug 7 2009 02:40:01:713PM | 1090807019 | Steve Wolfe | Shot#19: Testing 1090702888 model, with corrected (?) coefficient Raise nl04 to 6e19, see if it will run clean. plasma to 1.16sec, locked mode at 0.83sec, nebar=8.9e19/m^3 model888 b21 is < 1G, old model b21=3.5G This model also fails. Next: target b21_PROJ to -.0001T (step 5 for this model). |
| Aug 7 2009 02:54:37:540PM | 1090807020 | Steve Wolfe | Shot#20: Testing variation on the 1090702888 model, with corrected (?) coefficient. Target to -.0001T plasma to 1sec, injection at .4sec, locked mode at .55sec, but may be influenced by earlier injection. nebar=9.5e19, b21(888)=1.2G, b21(r3d)=3G : this doesn't fit either way, repeat and see if injection disappears. |
| Aug 7 2009 03:14:14:100PM | 1090807021 | Steve Wolfe | Shot#21: Testing variation on the 1090702888 model, with corrected (?) coefficient. Target to -.0001T plasma to 1.3sec, still lots of runaways locks at 0.78sec, has an injection at 0.75sec. I think the injection causees the lock, but move on with the opposite variation anyway. nebar=9.7e19, b21(r3d)=3.3G, b21(888)=1.1G Next: target to +.0001T |
| Aug 7 2009 03:24:18:467PM | 1090807022 | Steve Wolfe | Shot#22: Testing variation on the 1090702888 model, with corrected (?) coefficient. Target to +.0001T early disruption, 0.3sec, injection and runaways. No locked mode. Try one more time, then change taps back to +Dtop -Dbot -Jtop +Jbot |
| Aug 7 2009 03:41:58:403PM | 1090807023 | Steve Wolfe | Shot#23: Testing variation on the 1090702888 model, with corrected (?) coefficient. Target to +.0001T plasma to .9sec, locks around .85sec nebar=9e19, b21(r3d)=4G, b21(888)=1.3G Next: back to standard taps, standard feedback control, leave target at 0.0001. Use the keyword for ACOIL, since the EO hasn't updated the engineering tree yet. |
| Aug 7 2009 04:03:35:153PM | 1090807024 | Steve Wolfe | Shot#24: Testing variation on the original (r3d) model, including the bug that's been there since 2003. It makes a little less than 10% difference in this configuration. target=+.0001 No gas. PLC glitch, no HV enable. |
| Aug 7 2009 04:19:22:180PM | 1090807025 | Steve Wolfe | Shot#25: Testing variation on the original (r3d) model, including the bug that's been there since 2003. It makes a little less than 10% difference in this configuration. target=+.0001 plasma to 1.55sec No locked mode, nebar down to 5.7e19 by eof b21(r3d)=1.5G, b21(888)=7G, b21(333)=5.3G. next: -.0001 target |
| Aug 7 2009 04:32:31:207PM | 1090807026 | Steve Wolfe | Shot#26: Testing variation on the original (r3d) model, including the bug that's been there since 2003. It makes a little less than 10% difference in this configuration. target=-.0001 plasma to .75sec, short-lived locked mode at 0.65sec, nebar=1e20 injections starting at .6sec could be implicated. Move on to step 6: 1.2MA, target back to zero. |
| Aug 7 2009 04:45:59:577PM | 1090807027 | Steve Wolfe | Shot#27: step 6: 1.2MA, target back to zero. plasma to 0.7sec, locks between 0.5 and 0.55sec nebar~8.5e19, b21(r3d)=1G. |
| Aug 7 2009 04:49:04:300PM | 1090807027 | Steve Wolfe | Shot#27: step 6: 1.2MA, target back to zero. plasma to 0.7sec, locks between 0.5 and 0.55sec nebar~8.5e19, b21(r3d)=1G. next: B21_PROJ target to +.0002 |
| Aug 7 2009 05:11:12:580PM | 1090807029 | Steve Wolfe | Shot#29: 1.2MA, B21_PROJ target to +.0002 plasma to 0.86sec, nebar=1e20, no locked mode. b21(r3d)=2.5G, @35deg. Is this better than 27 or was that just random? This one has a big injection at .85sec. END OF RUN |
| Physics Operator Comments | |||
| Aug 7 2009 09:08:07:423AM | 1090807001 | Earl Marmar | start from 1090807032
nl_04 target to 1e20, then negative after 0.7 seconds increase nl_04 prop gain to 8 (from 2) extend nl_04 derivative gain to end of shot plc disable A-coils enable argon puff (PG1); program 50 ms @100, starting at 0.3 seconds bouncy startup, runaways, dies at 550 kA baseline offset f25, 1.2e-2 |
| Aug 7 2009 09:34:28:147AM | 1090807002 | Earl Marmar | prefill puff to 26 ms (from 24 ms)
plasma disrupts in rampdown still a bit of a bounce in startup |
| Aug 7 2009 09:34:55:897AM | 1090807003 | Earl Marmar | increase clearin to +0.001 (from -0.001)
plasma same rampdown disruption |
| Aug 7 2009 09:46:39:780AM | 1090807004 | Earl Marmar | nl_04 target to 8e19 (from 1e20)
plasma disrupt at end of flattop |
| Aug 7 2009 09:57:50:023AM | 1090807005 | Earl Marmar | prefill puff (deuterium) to 27 ms (from 26 ms)
plc enable A-coils plasma ramps all the way down |
| Aug 7 2009 10:11:54:910AM | 1090807006 | Earl Marmar | nl_04 target to 0.6e20 (from 0.8e20)
plasma early hards are up, but suppressed at about .1 s ramps all the way down |
| Aug 7 2009 10:42:30:667AM | 1090807007 | Earl Marmar | prefill puff to 28 ms (from 27 ms)
nl_04 target to 0.5e20 (from 0.6e20) alternator alarm during check; no power shot |
| Aug 7 2009 10:56:32:897AM | 1090807008 | Earl Marmar | delays before this shot for alternator and then data system
no dpcs changes plasma early hards disrupts just after end of flattop |
| Aug 7 2009 11:09:35:127AM | 1090807009 | Earl Marmar | nl_04 target to 3e19 (from 5e19)
plasma disrupt just after .3 s, preceded by a large impurity injection hards in the mid to high 10^13's |
| Aug 7 2009 11:24:22:717AM | 1090807010 | Earl Marmar | prefill puff to 31 ms (from 28 ms)
plasma plenty of hards disrupts in flattop, 0.66s |
| Aug 7 2009 11:43:36:950AM | 1090807011 | Earl Marmar | prefill puff to 34 ms (from 31 ms)
nl_04 target to 4e19 (from 3e19) plasma hards much reduced (low 10^12's) disrupts at .6 s, preceded by injection |
| Aug 7 2009 11:51:13:703AM | 1090807012 | Earl Marmar | nl_04 prop gain to 4 (from 8)
moving to step 3 in the run plan B21_PROJ keywords: SHOT=1090702333,ACOIL="+DTOP -DBOT -JTOP +JBOT"; target waveform still 0 plasma early hards ok, rising after 0.5 s disrupt in flattop, 0.68s |
| Aug 7 2009 12:05:11:310PM | 1090807013 | Earl Marmar | nl_04 deriv gain to 5 (from 3)
nl_04 target to 5e19 (from 4e19) plasma disrupt at 0.72 s |
| Aug 7 2009 12:14:46:347PM | 1090807014 | Earl Marmar | B21_PROJ back to earlier configuration (keywords left blank)
plasma disrupt at 1.17 s, preceded by impurity influx hards ~1e12 for most of the shot, rising just before the end |
| Aug 7 2009 12:50:52:920PM | 1090807015 | Earl Marmar | moving to step 4 in the run plan
A-coil taps changed to: +Bbot +Dtop -Gbot -Jtop B21_PROJ feedback keywords to: SHOT=1090702888,ACOIL="+BBOT +DTOP -GBOT -JTOP" plasma disrupt at 0.48 s, preceded by increase in radiated power |
| Aug 7 2009 01:06:46:543PM | 1090807016 | Earl Marmar | B21_PROJ prop and integ gain turn-on moved to .4 s (from .2 s)
plasma disrupt at .74 s, not preceded by any obvious impurity problem |
| Aug 7 2009 02:10:24:593PM | 1090807017 | Earl Marmar | delay before this shot while session leader considers the results and path forward
Edit B21_PROJ, correct_coil:bus_r_cur to 3.0e-7 plasma disrupt at 0.75 s |
| Aug 7 2009 02:24:24:487PM | 1090807018 | Earl Marmar | B21_PROJ prop and integ gain on at 0.2 s
plasma disrupt at 0.8 s |
| Aug 7 2009 02:59:10:730PM | 1090807019 | Earl Marmar | nl_04 target to 6e19
plasma early hards up disrupt at 1.17 s |
| Aug 7 2009 02:58:51:650PM | 1090807020 | Earl Marmar | moving to step 5
no change in a-coil patch panel B21-PROJ target to -0.0001 T, from 0.1 s through 2.0 s 4 minutes in init because of data system hang plasma early hards similar to shot 19 disrupt at 0.97 s |
| Aug 7 2009 03:12:25:897PM | 1090807021 | Earl Marmar | no changes
plasma disrupt at 1.32 s |
| Aug 7 2009 03:23:13:887PM | 1090807022 | Earl Marmar | B21_PROJ target to +.0001, from 0.1 s through 2.0 s
plasma disrupt at .31 s |
| Aug 7 2009 03:38:00:557PM | 1090807023 | Earl Marmar | no changes
plasma disrupt at 0.89 s |
| Aug 7 2009 04:03:42:233PM | 1090807024 | Earl Marmar | return the a-coil configuration to the "standard" settings +Dtop -Dbot -Jtop +Jbot
prefill puff to 35 ms (from 34) B21_PROJ keyword ACOIL="+DTOP -DBOT -JTOP +JBOT" leave the B21_PROJ target at +0.0001 T dud; no gas |
| Aug 7 2009 04:19:41:963PM | 1090807025 | Earl Marmar | no dpcs changes
plasma disrupt at 1.56 s |
| Aug 7 2009 04:35:38:663PM | 1090807026 | Earl Marmar | B21_PROJ target to -0.0001 T
plasma disrupt at 0.79 s |
| Aug 7 2009 04:50:48:550PM | 1090807027 | Earl Marmar | B21_PROJ target to 0
Ip to -1.2 MA plasma disrupt at 0.71 s |
| Aug 7 2009 04:56:02:180PM | 1090807028 | Earl Marmar | B21_PROJ target to +0.0002 T
fizzle, almost no plasma current |
| Aug 7 2009 05:11:04:440PM | 1090807029 | Earl Marmar | prefill puff to 33 ms (from 35 ms)
plasma disrupts at 0.87 s no hards |
| Engineering Operator Comments | ||||
| Shot | Time | Type | Status | Comment |
| 1 | 08:59:57:590AM | Plasma | Ok | |
| 2 | 09:12:33:567AM | Plasma | Ok | |
| 3 | 09:25:20:420AM | Plasma | Ok | |
| 4 | 09:38:07:523AM | Plasma | Ok | |
| 5 | 09:50:55:097AM | Plasma | Ok | |
| 6 | 10:03:42:123AM | Plasma | Ok | |
| 7 | 10:20:56:010AM | Plasma | Ok | |
| 8 | 10:49:25:203AM | Plasma | Ok | |
| 9 | 11:04:09:717AM | Plasma | Ok | |
| 10 | 11:16:59:133AM | Plasma | Ok | |
| 11 | 11:29:41:673AM | Plasma | Ok | |
| 12 | 11:42:34:947AM | Plasma | Ok | |
| 13 | 11:55:19:693AM | Plasma | Ok | |
| 14 | 12:07:58:950PM | Plasma | Ok | |
| 15 | 12:35:02:473PM | Plasma | Ok | |
| 16 | 12:57:39:507PM | Plasma | Ok | |
| 17 | 02:04:32:417PM | Plasma | Ok | |
| 18 | 02:17:09:610PM | Plasma | Ok | |
| 19 | 02:29:42:700PM | Plasma | Ok | |
| 20 | 02:42:24:793PM | Plasma | Ok | |
| 21 | 03:06:20:343PM | Plasma | Ok | |
| 22 | 03:18:53:620PM | Plasma | Ok | |
| 23 | 03:31:25:020PM | Plasma | Ok | |
| 24 | 04:00:10:540PM | Plasma | Ok | |
| 25 | 04:12:43:630PM | Plasma | Ok | |
| 26 | 04:25:14:717PM | Plasma | Ok | |
| 27 | 04:37:47:383PM | Plasma | Ok | |
| 28 | 04:50:35:423PM | Plasma | Ok | |
| 29 | 05:03:06:247PM | Plasma | Ok | |
| System Availability | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| Aug 7 2009 08:59:32:040AM | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||