| Session Leader Summaries |
| Entered: Apr 27 2006 05:05:25:010PM |
| Author: Amanda Hubbard |
A mixed run today. The main difficulty seemed to be getting full length 8 T plasmas, rather than getting H-modes.
Full length shots were on 2,3,6,7,10,11, and 13; all but shot 6 had H-modes.
The others disrupted at typically 0.4 s, well before RF, due to large impurity injections. There were also injections in the flat top. The source of these, and their relation to 8 T operation, remains to be determined.
On the other hand, we did get some useful pedestal comparison points. Shape development took only a few shots, with essentially the target shape and q95 by shot 6.
Shot 7 was a good H-mode with target nel=1.e20 m-2; got neped ~ 2e20 m-3 which was the same as 1010731008 at 5.2 T. Higher Teped (740 eV) and therefore lower nu*.
Shot 10 had reduced target nel, giving lower neped (1.3e20 m-3) and higher Teped (830 eV). We thus have close to the desired ne and Te for dimensionless scaling to the 5.2 T point (2e20 m-3, 910 eV) but not at the same time; this would need a little more power. Interestingly, got a couple of what look like discrete ELMs early on this shot, with distinct drops in edge Te. Later there are more extended d-alpha rises; rather reminiscent of some of our ELM runs in JFT2M shape and bear further analysis.
Shot 11 had slightly elevated target ne, and reduced Tped. Probably not enough, in retrospect. All of the H-modes today had ne rising, along with Prad. I speculate that this is due to the low nu* (typically ~ 0.5, down to 0.3 shot 10). A worthwhile direction for a future run would be to raise ne, by the target and/or puffing into the H-mode. But, this would take some time, and probably more RF power due to the higher threshold. We did see fluctuations on PCI on parts of some shots (eg, shot 3, 1.2 s) which look to me like higher f QC modes though Liang Lin is not convinced.
For shot 13, reduced Ip to 1 MA as an alternate comparison point for AUG, in case they have trouble matching this shape at full B and Ip. Did get H-modes, with 3.6 MW RF; some trips. ne95 close to 2, Te95 ~600 eV. Liang sees a weak QC mode on this one.
Given only 7 full length shots, 6 with H-modes, we did learn some interesting things and get some starting points for future comparisons. But, cannot consider this expt complete -need to understand and fix the operational difficulties (mainly injections) with these 8 T shots and have more RF available; note that it was determined early in the run that J-port has transmission line arcing somewhere and was limited to only ~ 1 MW.
Note that some shots (10,11 at least) reached 170 kJ; perhaps our highest energy at 8 T? Most were over 150 kJ.
|
| Physics Operator Summaries |
| Entered: Apr 27 2006 06:13:26:230PM |
| Author: Steve Wolfe |
Physop summary for 1060427
MP#382 C-Mod/JET/AUG pedestal identity experiments
SL: Hubbard
PO: Wolfe
EO: Cochran, Parkin, Dekow
A full day of 7.8 Tesla operation resulted in 13 shots,
7 full length plasmas, 5 early disruptions, and a fizzle.
Some progress on the MP today, and I think a good start
which provides a basis for JET and AUG to proceed; AUG may
actually get some time soon, but JET is likely delayed due
to TS problems.
The shape, based on 1010731008, was matched by shot#6 (see
figure in logbook entry under PHYSOP). The key move was
dropping ZCUR with lower Xpt coordinates fixed to match the
top. Should have thought of that on Tuesday for Martin/Craig's
run, instead of fighting with the upper x-point.
The main operational difficulty today was injections, which resulted
in 5 disruptions before the RF. This was also the case yesterday.
I tried changing the early shape evolution and density rise to
try to prevent whatever interaction was causing the injections,
but without success. It's possible that my changes on shot#5 caused
a control problem around crossover which resulted in the disruption
on this shot. Video views seem to indicate dust showers and/or explosions
in the divertor near C-port were the culprits on most of these short shots.
An optimist might think that things were improving at the end of the
day, but I wouldn't bet on it.
The other operational problem which cost us a lot of runtime was trying
to track down the reason for the TF breaker pulls after successful shots.
The engineers found, based on freshly installed instrumentation in the
TF cabinet, that a Type 2 fault was occurring between 1.5s and 1.6 sec,
just before PLC inversion time; this has the effect of causing a
forced inversion (which wasn't noticed because it was happening within
5 msec of the programmed time for this to happen anyway) and then opening
the breakers 3 seconds later. While all the current was out of the magnet
at the time the breakers opened, there was a large negative voltage spike
which, at least on one shot, produced a transient current pulse in the
supply of several tens of kA; this actually ended up producing an IOC
fault, which generated a type 1 indication. By moving the test probes around
in the fault chain shot-to-shot, the engineers finally got data on the last
shot of the day that indicated the source of the original Type 2 is a
calculated junction temperature fault. This makes some sense, since we
are running slightly longer TF pulses than most recent 8T runs (by 0.1sec),
and the fault was showing up right at the end of the TF. Gary's MATLAB
program indicates that the junction temperatures should be well within
bounds at this time, but it could be that we have never pushed quite as
high as this, so we wouldn't have encountered this fault indication. Bill
and Gary will check further, and decide whether they should adjust the
trip level or tell us to restrict the pulse length.
The remaining problem today was the ICRF, specifically J-port which continued
to fault repetitively at powers above 400kW (each). Acoustic monitoring
was installed on J3 and J4 transmission lines near the phase shifters, and
indicated the arcs were occuring in or near the J3 phase shifter again. This
is the unit that was removed and repaired previously. Since it is now pretty
clear that it is broken again (still?), we have decided to cancel tomorrow's
run and get started on dismantling it for a more thorough, and hopefully more
effective, repair, aimed at getting us up to high power operation next week.
On the plus side, we did get good heating at 8T resulting in H-modes
at the original target q95=4.1 and also, on the last shot, at higher q95=5.2,
which was requested by AUG as a possible alternative point. At q95=4.1
(Ip~1.3MA) we got data at target densities of .8, 1., and 1.1e20,
which provided some range in nu_ped. Wmhd got as high as 170kJ, which I think
is the highest we've had at 8T. He3 heating seems to work reasonably well,
contrary to some of our earlier experience.
On the last two shots we allowed Brock and Earl to shoot pellets into the
rampdown. I'd like to blame the lone fizzle on shot 12 on the 600kA disruption
after the pellets on #11, but since we had been disrupting at higher current
all day I really can't. On the other hand, I don't really understand why #12
didn't go.
I do think that the startup today, with lower current in the EF1 coils, is
better than we were running yesterday. Most didn't bounce, and runaways
were down. Breakdowns were all nice and early.
The sparker was running at high currents all day. It seems to like high
field. Early in the day I observed multiple large spikes/glitches on the
magnetics inputs to dpcs, which disappeared after t=0. These were found
to be associated with arcs in the sparker (spikes in the sparker current).
See notes under DPCS topic for more details.
All the dpcs spikes lined up exactly with sparker arcs, although not
all the sparker arcs show up. Since the spikes that are visible typically last
only one or two samples (100usec per), it is likely the others were
just too short to register. The spikes were also seen, less frequently,
on the same signals digitized in the old hybrid CAMAC. They were not
present on the signals in the magnetics digitizers, which indicates they
are coming in from the iso-amps. Later in the day the arcs went away and
so did the interference spikes; the sparker continued to have high current
(over 1mA) but no arcs. Irby has put in a tech request for better shielding
of the sparker circuit, and I put one in to enclose the iso-amp cards
in a shielded (and ventillated) crate. Meanwhile, this primarily only a
cosmetic issue as long as the sparker arcs stop at t=0, since before breakdown
the magnetics signals are only used in controlling the BR0 and the
fast spikes are not enough to affect OH2 and EF2 the coil currents
on this time scale.
Note that if they were present during plasma I would need to filter
the zcur error signal to use it for a gasjet trigger, but I don't think
it would be a problem for the chopper feedback.
|
| Session Leader Comments |
| Apr 27 2006 09:12:34:383AM | 1060427001 | Amanda Hubbard |
First shot disrupted ~ 0.4 secs. Had just reached 0.78 T, 1 MA.
The big injection was at 0.35 secs.
|
| Apr 27 2006 10:17:51:757AM | 1060427002 | Amanda Hubbard | Much better. Nearly full length, 1.35 MA, 7.85 T.
Got 1.8-2.8 T, not bad for 1st shot after boronization. And, enough to get H-mode (with target nel 10^20 m-2, nebar 1.9e20 m-3.
Long H-mode, with 2 short back-transitions or perhaps short ELMy periods.
q95 in right range, 4.1-4.3; likely some changing of shape.
Note that sawteeth start very late, 0.6 secs.
Pedestal parameters rather erratic given RF variations.
Very roughly Teped ~ 570 eV, neped up to ~2.7e20 m-3. Which makes a pretty high pressure, 1.6e23 eVm-3, and Wmhd 157 kJ, apparently best yet at 8 T.
|
| Apr 27 2006 11:00:55:430AM | 1060427003 | Amanda Hubbard | Full length shot, with some shape tweaks.
But, J-port still tripping. D+E ramping up to 2 MW, clean.
Following shot, RF crew are going to do some vacuum RF tests on J to see what's wrong.
Late H-mode at 1.1 secs. 159 kJ
Pedestal parameters similar to last shot, Te~704 eV, neped ~ 2.4-2.8e20m-3.
q95 has dropped a bit, to ~4.05.
beta_N to 0.6
Interestingly, a mode is seen on PCI from 1.2-1.28 secs. This is the part of the H-mode when ne rises, and rolls over and Te is dropping, which would be consistent with nu* threshold for QC. Jerry's scope shows nu*ped 1 to 1.5 in this period, vs ~0.7 minimum, (when there is no mode). Mode f 200-300 khZ, much higher than typical at lower B.
|
| Apr 27 2006 11:24:36:210AM | 1060427004 | Amanda Hubbard | Short plasma, disrupts 0.4 s (like shot 1).
It has been determined there is likely arcing in the J-port transmission line, in the power room. This will prevent completing the high power goals of the run. But, there is still room to complete part of the run. Propose to complete the shape development, and to scan target density and Ip to some degree (limited by L-H threshold). This will give us a much better basis on which to estimate the Tpeds and dimensionless parameters to expect at higher powers.
|
| Apr 27 2006 11:41:35:633AM | 1060427005 | Amanda Hubbard | Disrupts at 0.44 secs.
|
| Apr 27 2006 12:20:53:443PM | 1060427006 | Amanda Hubbard | Much better - got the shape match "nailed". q95 a bit low, ~ 3.9
RF up to 3 MW, but no H-mode; not sure why since nel was about the same (9.6e19 ,-2, nebar 1.75e20) Might well have been due to injection at 0.967 s; Te8 was still rising before this, and never recovered. Te95 was ~ 340 eV, apparently below threshold.
For next shot, square up and increase power on D and E, to get H-mode.
Leave J alone; it actually produced up to ~ 1.1 MW before tripping.
Also reduce Ip to match q95=4.1
|
| Apr 27 2006 02:57:05:100PM | 1060427007 | Amanda Hubbard | Full length plasma, with 2-3 H-modes. Injections and RF trips complicate life somewhat (no truly steady periods), but we did get up to 3.9 MW RF, and 3.5 MW for an extended period. 163 kJ.
Looks like Te_95 up to 760 eV, ne_95 ~2e20 which is about where previous 5.2 T H-modes were; an interesting comparison point at ~ same PRF. nu* is accordingly lower, down to ~ 0.5.
Next shot, decrease target ne; this should raise Teped.
Jim notes drift in TCI; ne in second H-mode is not really higher, we think.
|
| Apr 27 2006 02:42:53:727PM | 1060427008 | Amanda Hubbard | Disrupts early again, ~ 0.37 secs.
|
| Apr 27 2006 03:45:33:743PM | 1060427010 | Amanda Hubbard |
Good one! CLEAN RF trace at 3.3 MW, including 1 MW J.
Lower target ne of 8e19 m-2 (nebar 1.2e20 m-3) did give lower neped and higher Teped (~1.36e20 m-3, 830 eV at psi=0.95). nu* down to ~ 0.3.
First H-mode had two clear ELMs, seen as D-alpha spikes plus clear dips in edge ECE (eg GPC_9, FRCECE_28) Later in this H-mode, and in the second one, get longer rises in D_alph, maybe brief L-Modes?
First, hottest, H-mode was ended by an injection at 0.95 secs.
second H-mode had teped ~600 eV and neped ~ 1.0-1.2 e20.
No QC mode, as expected.
This shot bears MHD stability investigation.
Next shot, try higher target ne, 1.1 e20m-2
(given limited full length shots today, have to do 1-point parameter scans.)
|
| Apr 27 2006 04:26:43:380PM | 1060427011 | Amanda Hubbard |
Another full length shot, with H-modes. Target nel was raised to 1.1e20 m-2, though actual nel before RF was 9.7e19 (nebar 1.7e20 m-3). A little higher than eg shot 7, though perhaps I should have raised more.
J-port power was raised, resulting in trips. Ptot up to 3.7 MW - D and E look happy.
Did get slightly higher neped in first H-mode (ne95 up to 2.3e20 m-3). Teped up to 750 eV, more typically 600 eV.
Still plagued by injections, eg at 900 ms, 980 ms. More generally, these shots are not steady, with ne and Prad rising until back-transitions. It would be nice to keep raising nebar to reduce nu*, see if we can get steady EDA. But, I don't think we'll have time today.
Next shot: reduce current to 1 MA and target nel to 1.e20 m-2 (Wolfgang has requested a lower current in case AUG has trouble matching shape at low q, high B).
|
| Apr 27 2006 05:05:51:230PM | 1060427013 | Amanda Hubbard |
For shot 13, reduced Ip to 1 MA as an alternate comparison point for AUG, in case they have trouble matching this shape at full B and Ip. Did get H-modes, with 3.6 MW RF; some trips. ne95 close to 2, Te95 ~600 eV. Liang sees a weak QC mode on this one.
|
| Physics Operator Comments |
| Apr 27 2006 07:30:14:850AM | | Steve Wolfe |
Setting up for MP#382/PEP-7: "C-Mod/JET/AUG pedestal identity experiments"
Import seg2 from 1060426018 (1.6MA, 7.8T, 1.3e20)
Decrease Ip from 1.6 to 1.35MA, aiming for q95=4.1
Decrease NL04 from 1.3 to 1.0e20
INcrease RXL from -.005 to +.005 (from .91sec)
Increase ZXL from -.01 to -.007
Increase ZXU @1.543s from .02 to .047
Leave TF rampdown starting at 1.4sec, but remove backporch
In seg1
Decrease EF1U/L from -5900 to -5800A
this should help compensate the extra index from the TF,
and give ~1mT more positive Bz
Decrease PG4 from 22->20ms
Decrease PG3 from 40 to 30 between .05 and .10s
Load at 27-Apr-2006 07:29:25.00
Open tree /home/wolfe/pcs_scratch -1
Open tree done
|
| Apr 27 2006 11:30:17:460AM | 1060427001 | Steve Wolfe |
Shot#1: Try for 1.35MA, starting point for tuning shape and q95
RF will condition up as we go along.
Early disruption, .38sec, this looks familiar
PF breaker pulled out.
Bouncy startup, had hard x-rays. P0 was under .05
This shot it looks like it's touching outboard around
.25sec; I don't think it was doing this yesterday,
but I can keep it a little further in. Current is just up to 1.0MA
at the time of the injection; there is a small Ip gain switch then.
|
| Apr 27 2006 11:29:54:757AM | 1060427002 | Steve Wolfe |
Shot#2: Slow down current rise program a little to match 1.0MA at .35
Move early RCUR points in to .657 from .6585
Drop seg1 pg3 puff from 30 to 25 (after .05sec)
plasma to 1.8sec but almost lost it to an injection at .2sec
157kJ during the RF. J-port doesn't look good.
Yijun wants reduced outer gap.
|
| Apr 27 2006 11:30:42:600AM | 1060427003 | Steve Wolfe |
Shot#3: Increase rcur by 4mm (decrease outer gap by 8mm)
Increase zxu by .02 from .047 to .067
plasma to 1.75sec, I should check on the rampdown.
H-mode gets to 158kJ
One bounce in startup, no runaways, no serious early injection
The top of the discharge hardly came down at all.
I've succeeded in moving the inner upper xpoint out of the vessel
but it still has too much pull from ef2U
Try leaving RXU inside, instead of swinging back toward zero
after 1sec
|
| Apr 27 2006 11:31:01:150AM | 1060427004 | Steve Wolfe |
Shot#4: Try leaving RXU inside, instead of swinging back toward zero
after 1sec.
The rampdown disruption seems to be EF2's going to zero, led by
ef2L. Slow down the RXL and clearin ramps after 1.5sec
Increase pg4 from 20 to 23ms; Change seg2 V_chopper voltage from 100V to 0
Drop ZCUR from -.009 to -.015
And another early disruption, injection at .35sec
Before any of my changes except the ZCUR drop, which started at .2sec.
Looks like it was tracking this ok.
|
| Apr 27 2006 11:46:29:460AM | 1060427005 | Steve Wolfe |
Shot#5: I don't believe the injections are due to anything in the
control, but try anyway. There seems to be some indication it's
happening just after the strike goes into the slot:
Start ZXL negative, slow down clearin coming off the limiter.
Early disruption, .445sec on the injection
This was different, looks like the chopper railed?
That may be something I did? Could be a crossover problem?
|
| Apr 27 2006 12:25:28:680PM | 1060427006 | Steve Wolfe |
Shot#6: I think I did that one. Undo the clearin, zxl changes
back ZXU off to .055
Plasma ran through, breakers pulled after the shot - At least that's
not my problem!
Several injections, but it survives.
L-mode, got some RF.
Nailed the shape this time. q95=3.9, could drop Ip program to 1.28MA
|
| Apr 27 2006 01:41:38:837PM | 1060427006 | Steve Wolfe |
Shot#7: drop Ip program to 1.28MA
We are having a long delay (1:30 and counting) to add some
instrumentation to the fault chain in the TF power supply.
There was a type II fault on shot 6 at 1.595sec that caused
the breakers to open at 4.6sec, followed by a rather larger
voltage spike in the TF supply. We would rather avoid this
happening again.
|
| Apr 27 2006 02:25:45:803PM | 1060427007 | Steve Wolfe |
Shot#7: drop Ip program to 1.28MA
plasma full length, no breakers pulled this time.
H-modes (2), lots of injections, but it hung in there.
First h-mode gets to 163kJ. J still faulting, max Prf 4MW
q95 between 4.1 and 4.2.
Next shot: nl04=0.8e20
|
| Apr 27 2006 02:49:31:210PM | 1060427008 | Steve Wolfe |
Shot#8: drop nl04=0.8e20, Ip program to 1.30MA
Give JR 30msec of Argon (remote request from PPPL)
Another early disruption. Injection at 0.32sec finally
got us. There was one at essentially the same time on #7
but we got past it. Prad started up at .30sec
Still happening just after it diverts.
|
| Apr 27 2006 03:14:45:070PM | 1060427009 | Steve Wolfe |
Shot#9: Add a small pg3 puff (20V from .25 to .5sec, start up at .1)
Move RXL out after .31sec
Same disruption, around .42sec,
Didn't get the extra pg3 puff, density too much higher than the program.
JANT video frame 12 shows something exploding in the divertor around ? C-port ?
|
| Apr 27 2006 03:33:43:913PM | 1060427010 | Steve Wolfe |
Shot#10: Drop nl04 feedback until .4sec to 2.0 from 9
plasma full length, no disruptions, but tf breakers pulled after
shot. 170kJ, Amanda claims we've got ELMs
Still didn't get early seg2 gas puff. No big injections until
0.9sec. Small ones at .18 and .54sec
J-port ran clean at 1MW, Prf=3.4MW
Next: Nl04 to 1.1e20
|
| Apr 27 2006 04:10:57:513PM | 1060427011 | Steve Wolfe |
Shot#11: Nl04 to 1.1e20, per SL.
Increase pg3 in seg2 from 20 to 35 early, still hoping to fill in the dip in
ne(t), which probably isn't hurting anything anyway.
Pellet set for 1.75sec
Plasma lasted until pellet. Multiple H-modes.
density only got to 1.0 instead of 1.1e20.
Lots of faulting on J.
Next: 1MA, 1.e20 (program)
|
| Apr 27 2006 04:38:57:403PM | 1060427012 | Steve Wolfe |
Shot#12: Nl04 to 1.0e20, Ip to 1.0MA per SL (requested by AUG)
Fizzle. Got to 64kA and died. Should have gone.
One spike on Dalpha, but it doesn't look like a bounce.
Drop pg4 by 1msec and go again.
|
| Apr 27 2006 05:02:15:917PM | 1060427013 | Steve Wolfe |
Shot#13: Pg4 from 23 to 22ms (purely for superstition)
Nl04 to 1.0e20, Ip to 1.0MA per SL (requested by AUG)
Plasma through pellet, ramped on down
breakers pulled after shot; engineers can check newly instrumented point.
It looks like it is the junction temp interlock.
2.6 H-modes, not very good. Best is 128kJ
q95=5.25.
J had multiple trips, D and E a few.
One early injection at .48sec didn't cause a problem.
Also had some injections during the h-modes, which didn't help at all.
END OF RUN. More next week (probably)
|