Alcator C-Mod Run 930618 Information

Back to runs

Miniproposals
Miniproposal:
Date Filed:
Title:
First Author:
Session Leader:Steve Golovato (shots 1-27)

Operators
Session leader(s):Steve Golovato
Physics operator(s):Steve Wolfe
Engineering operator(s):Frank Silva,Vinny Bertolino

Engineering Operator Run Comment
Improved plasma control

Session Leader Plans

Physics Operators Plans

Session Leader Summaries
Entered: Jul 7 2004 03:49:07:930PM
Author: To Be Determined
Summary of run 930618

Physics Operator: Wolfe
Asst. Operator: Golovato
Eng. Operator: Silva/Bertolino

The goal of this run was to continue development of diverted plasmas.
A secondary goal was to study the stability properties of the elongated
plasmas and control schemes.

The run was a qualified success. Divertor discharges were produced, and
some progress was made toward putting the outboard separatrix into the
lower divertor chamber. Considerable data was collected on the behavior of
the system with different settings of the proportional and derivative gain
on the ZCUR wire. However, we were unable to prevent losing the diverted,
or even the limited higher elongation plasmas, to vertical instability
disruptions; we have some nice examples of these for the pathology
specialists. Maximum elongations achieved, according to FFIT, were about
1.6. It is not really surprising that these would be unstable with the
present feedback capabilities, namely using OH2 and EF1 as the control
coils; in fact, I'm surprised we can control these equilibria as well as
we did.

Wall Conditioning/Vacuum:
-------------------------
ECDC was run for about an hour prior to the run, ending at 07:30. The run
began at 09:45, with the delayed start being attributable to cooldown. A
leak opened up (N2 on the RGA) just prior to start of the run; the N2
signal was just below 1.e-9 on the RGA, and no He signal was observed,
implying that the leak was not at one of the inner horizontal ports, all
of which are being purged. The leak was probably present for the entire run.

Detailed Summary:
-----------------
We began from programming similar to 930616028, which had shown evidence of
being diverted. Shots 2 (no magnetics data) and shot 3 are examples. Shot
930618003 reaches a maximum elongation of 1.5 (according to FFIT) at 0.2
sec and seems to be a nice symmetric, limited but close to diverted,
discharge. It loses vertical position control after 260msec around the time
the OH2's run out of current.

Shots 4-9 were attempts to divert by holding the OH2's more positive,
effectively pushing the inner midplane off the inner wall. These got
higher elongation (1.6 - 1.7 according to FFIT) and actually diverted, but
lost vertical position control earlier and disrupted into the divertor. A
search of PID gain space is included in this sequence.

Shots 10-12 were a continuation of the symmetric programming, aiming for
lower elongation by backing off on the OH2 adjustment made earlier, ie
allowing them to run down a little faster. The elongation in fact did not
change much, remaining around 1.6, but there was some reduction in the inner
gap. Shot 12 is a good shot, although it only stays diverted for about 60
msec (.176 .. .236). Shot 13 was an attempt to put the separatrix into the
lower chamber by reducing the EF2 current; this moved in the right
direction, but also reduced the divertedness, not unexpectedly.

The remainder of the run was devoted to lower null operation, with the
plasma moved down below the midplane. There is a good scan of proportional
and derivative gains, and some very impressive oscillations on the
vertical position (see the logbook). Longest time on lower divertor was
about 90msec on shot 20, which exhibited a growing oscillation and finally
dived into the divertor. The strike point position was apparently on the
outer strike plate around 175msec and then moved up to the nose piece by
208msec, according to the FFIT reconstruction. Elongation was 1.5:1.6, and
triangularity in the range of 0.4 to 0.5.

Problems:
---------
On shot 2, one of the magnetics CAMAC modules failed to INIT, causing the
data from the previous shot to be re-stored. This was immediately
noticebale, because the previous shot had been a dud, while shot 2 had a
good plasma. However, the possibility for confusion exists if this occurs
in a series of normal shots. There should be some thought given to flagging
CAMAC errors in some way more apparent than having a message appear on the
ACTION MONITOR, which is typically not watched unless someone notices a
problem.

PF commutation failures occurred on shots 25 and 26 (OH2 and EF1). At least
the first of these was probably caused by having too high a derivative gain
in the hybrid ZCUR controller.

After the run, it was discovered that the RCUR waveform programming
inherited from the previous week and a half was invalid. The values drawn
correspond to an over-range of the WAVEGEN DAC by about a factor of two (ie
it called for a gain between -18 and -25, while the maximum range allowed
is +/- 10.) This was the reason that attempts to move the radial position
by incrementally changing the existing programming on shot 20 and again on
shot 26 failed to have any effect. Two actions should be taken to correct
this:
1) The present programming should be adjusted for the next run. The
minimum programmed radius with present value of FACTOR is about
0.64, I think; the inherited value was 0.61, and the value after
shot 26 is 0.63. Also, as I understand Ian's predictor, there should
be a one cm offset, such that asking for .64 should give a position
of 0.65 if everything is working. This is because the center of his
flux points is at 0.675, while the nominal center used in the tree
is 0.665.
2) The PCS software should be fixed to account properly for drawn
limits on the IP-referenced channels, making it impossible to draw
waveforms that cannot be represented in the WAVEGEN DACs.

The EF3 turn-on timing drifted by about 1.5 msec during the run. ALL of the
fizzle shots (1,7,15, and 27) were associated with EF3 turn-ons at the
extreme early end of the range. Shot #001 had this property because of
inherited programming from a previous run; #0027 was moved earlier to compensate
for the observed drift to later times, which has been anecdotally correlated
with increased runaway production, and apparently overdid it (a 1 msec move
is the smallest the wavegen can accomodate.) The others were random
fluctuations in the timing.

The presence of a cooling-associated leak continues to be a problem.

Score-card:
-----------
23 "good" shots
4 fizzles
0 no-breakdowns
--
27 total shots

Physics Operator Summaries

Session Leader Comments
Jun 18 1993 08:53:30:700AMSteve Golovatoone hour of discharge cleaning before run. shut off at 7:30am.
base pressure before cooling and cleaning started 1e-7.

goal of run will be to continue learning to make a diverted plasma.
first will try up-down symmetric, then will try moving it down toward lower
divertor. also may try turning off feedback during shot to look at
instabilities (if any) in the elongated plasma.

first shot will be repeat of shot 930616028 (shot 28 of wednesday) but
without the argon puff. this is one of the shots that may have been
diverted.

oh2's still have unbalanced demands.

session leader - wolfe
assistant - golovato
engineering operator - silva




Jun 18 1993 09:45:17:000AM930618001Steve Golovatoshot 930618001

A leak has opened up: N2 is just less than 1.e-9. Base pressure is 2e-6.
We will go for it anyway.

Prefill set to 4.5e-5. All else just like 930616028, except no Argon puff.
Prefill for 930616028 was set to 4.e-5.

problem with ef3's yesterday was not fixed in the pcs setup for
930616028. ef3 pre-programming from yesterday (930617013) was
inserted. will retry shot with this setup. also, argon pulse
was not disabled, it is now.





Jun 18 1993 10:03:00:370AM930618002Steve Golovatoshot 930618002

good shot. problems with magnetics diagnostics. looks like it didn't
disrupt until 300 ms. daq2 camac magnetics traq3 did not init.
-------------------------------------------------------------------------
WARNING : MAGNETICS DATA IS FROM PREVIOUS SHOT.
-------------------------------------------------------------------------
current was actually about 300ka (from hybrid).


Jun 18 1993 10:18:03:890AM930618003Steve Golovatoshot 930618003

lowered pre-fill to 4.e-6, otherwise a repeat to get magnetics data.
nice shot, 320ka, first disruption at 330ms.
didn't divert according to ffit.

Jun 18 1993 10:22:14:640AM930618003Steve Golovatoshot 930618003, further comments for shot 003.
max elongation 1.5 at at 0.2 sec., lose vertical position control at
0.26 sec when oh2's run out.

Jun 18 1993 10:20:21:660AM930618004Steve Golovatoshot 930618004

raising oh2 voltages (making thehem less negative) by 20 volts each after
120ms (maintaining imbalance).

Jun 19 1993 01:39:53:190PM930618004Steve WolfeTEXT: shot 930618004.

diverted after about 175 ms.
went vertically unstable downward at 180 ms where ffit kappa=1.7 .
goes in 2-3ms according to ian's reconstruction.

Jun 18 1993 10:52:25:360AM930618005Steve Golovatoshot 930618005.

turned off ef1 controllers on vertical postion, letting only oh2 control
vertical position. lowered pressure to 3.8e-5.
disrupted slightly earlier, otherwise similar to previous shot.
hard x-rays. electrons hit outboard limiter lower than before.
evidence off upper x-point on tv.

Jun 18 1993 11:03:17:230AM930618006Steve Golovatoshot 930618006.

raised derivative gain from .4 to 2. on zcur.
lost plasma up. oscillated. perhaps too much derivative gain.

Jun 18 1993 11:29:29:100AM930618007Steve Golovatoshot 930618007.

lowered derivative gain to 1.
fizzled. Plasma went vertically unstable at 30 msec.
We have a -1mT radial field, apparently because the hybrid thinks the
field is 9mT and we are asking for 7mT (damned offsets!!)
Raise the S on BR to 8mT for next shot.
Also raise fill pressure to 4.3e-5.


Jun 18 1993 11:47:13:500AM930618008Steve Golovatoshot 930618008.

similar to 005. disrupts at 180ms, no oscillations in vertical position.




Jun 18 1993 12:08:20:240PM930618009Steve Golovatoshot 930618009.

set zcur proportional gain to .2 from .4 and derivative
gain back to 2. from 1. (as originally suggested by ihh).

raised gas pulse from 40-80ms to try to reduce moly influx.
lowered pre-fill to 3.8e-5 form 4.3e-5.

better shot, disrupted at 210 ms. diverted up first, then double null, then
down, then lost down.



Jun 18 1993 12:31:12:480PM930618010Steve Golovatoshot 930618010.

Drop each OH2 by 10 V after 120m0 msec (half-way back to original settings.)
Now at -30 and -40 V.

runaway shot, but didn't appear to hit outboard limiter this time.


Jun 18 1993 12:51:59:860PM930618011Steve Golovatoshot 930618011.

Reload gas programming from shot 8. No other changes.
better shot., appeared to start moving down at 200ms but
recovered (feedback control worked?)
first disruption at 225ms (not a vertical instability).
shot lasted to 320ms.



Jun 18 1993 01:16:15:480PM930618012Steve Golovatoshot 930618012.

raised proportional gain on zcur from .2 to .5 (more than original .4)
raised pulsed gas between 70-170ms.

good shot.
doesn't go vertically unstable untitil 300 (when coil currents run out).
z-position has some oscillations ( (where kappa peaks at 1.7but
stays well-positioned until 300 ms.


Jun 18 1993 01:43:16:060PM930618013Steve Golovatoshot 930618013.

Lowered EF2 current demands after 100 msec.
Did not make any changes to deal with a hesitation on previous shot current
rise; we'll see.

not a bad shot, disrupted at 220 ms. marginally diverted.
Separatrix is pretty much in the right place.
Still has a less negative than usual Bz at breakdown.

Jun 18 1993 01:59:00:950PM930618014Steve Golovatoshot 930618014

Lower Zcur demand to -1.2 cm from 162 msec. Note: did not get rid of Br
feedback, so this will fight the change.

Lowered EF4 current by 50 amps at 0. to correct Bz at breakdown.

Worked ok. lowered by about 1cm. Still lots of hards.


Jun 19 1993 01:30:35:430PM930618015Steve WolfeTEXT: shot 930618015

extended ef1 currents in time, used to ramp down at 180ms.
ramp rate reduced.
fizzled. Ian thinks there is a radial field problem.
Wolfe doesn't. They agree the Bz is too negative.
Also the density looks a little high

Jun 18 1993 02:26:52:700PM930618016Steve Golovatoshot 930618016
Raise EF4 by 25 amps (undo half of previous change).
Lower fill to 3.5e-5.

Shot went. lose it at about 270 ms when oh2's run out.
Diverted down from 180 to 240 ms but close to double null.


Jun 18 1993 02:53:55:590PM930618017Steve Golovatoshot 930618017

lowering zcur to -2 cm.
turned back on ef1 controllers on zcur at +/-3.e-3 (same as oh2s)
lower fill pressure to 3.2e-5

lots of hard x-rays but didn't appear to hit outboard limiter.
moved down by 1.6-1.8cm, disrupted at 220ms.







Jun 18 1993 03:14:26:490PM930618018Steve Golovatoshot 930618018

set zcur to -2cm from 130 msec on (was 162)
Raise OH2's by 5 V from 120msec (to -35,-25)
Raise pressure back to 3.5e-5

nice shot. z-position oscillates (average about -1.4) but lasts to 300 ms.
hard x-rays lower but still there.
diveted down from about 170-270ms.





Jun 18 1993 03:36:09:140PM930618019Steve Golovatoshot 930618019

Try the double controller trick on EF1's:
From 30msec turn off controller on IC_EF1U
and drive V_EF1U and V_EF1L from IC_EF1L

It worked. Shot is very similar to previous,
except the noise on the xray centroid is back.
Still getting oscillations on the vertical position
that get bigger when OH2 goes away.


Jun 18 1993 03:52:02:670PM930618020Steve Golovatoshot 930618020

moved rcur waveform out by 1cm strting at 116 ms.
z-centroid oscillation grows with time, lose plasma down at 250ms.


Jun 18 1993 07:25:10:950PM930618020Steve WolfeThis shot was an attempt to move the plasma out. However,
the xray centroid indicates that the position is actually inside of that on 18,
while the FFIT R_CUR adn R_GEOM values are essentially identical.

I believe that the R_CUR control is not actually working very well.

Jun 18 1993 04:10:20:410PM930618021Steve Golovatoshot 930618021

divert plasma earlier by bringing up ef1 earlier, at 100ms rather than 160ms.
raised ic_ef1_l to 3000a at 100ms.
lowered ef2's to 600a from 120ms on.

worked, plasma lasts slightly shorter time.
z oscillates more slowly.


Jun 18 1993 04:28:48:790PM930618022Steve Golovatoshot 930618022

move ef1 crowbar to 40 ms from 70 ms.
raise Zcur proportional gain to 1.0 from 0.5
Move ZCUR pids to 30 msec (were at 70msec).

shorter shot, lasts to 180 ms.
z oscillates throughout.




Jun 18 1993 04:44:56:680PM930618023Steve Golovatoshot 930618023

Lowered proportional gain on zcur back to .5
Lasted 220ms. Only oscillates when it moves down after 140ms, sl
Slower oscillation, a lot like shot 21.





Jun 18 1993 04:58:38:620PM930618024Steve Golovatoshot 930618024

Raised derivative gain on zcur to 4 from 2.
Much like last shot. oscillations faster, lower amplitude, throughout shot.



Jun 18 1993 05:09:55:270PM930618025Steve Golovatoshot 930618025

Lowered ef2s to 250 from 600 from 100ms.
Increased zcur derivative gain to 8 from 4.
ef1 and oh2 commutation failure. too much derivative gain?
Growing oscillation from the start.


Jun 18 1993 05:50:42:930PM930618026Steve Golovatoshot 930618026

Lower zcur derivative gain back to 4.
Move RCUR demand 1cm out.
EF1L had fault.
Oscillation looks a lot like shot 21 again.
The RCUR adjustment didn't do anything.
xrays are getting bad, it looks like EF3 is coming in late again.



Jun 18 1993 07:27:40:360PM930618026Steve WolfeThis shot was another attempt to move the plasma out.
It didn't seem to work.

I believe that the R_CUR control is not actually working very well.
See comment on #20

Jun 19 1993 01:37:27:810PM930618027Steve WolfeTEXT: shot 930618027

Bring EF3 on 1msec earlier.
Hold EF4 up at 400 amps through the end of the shot to try to move th
the plasma out a bit.
fizzled. for unknown reasons.

Physics Operator Comments

Engineering Operator Comments
ShotTimeTypeStatusComment
109:34:28:810AMPlasmaOksmall flash -- no faults
209:47:06:240AMPlasmaOksmall flash -- no faults
310:05:04:300AMPlasmaOk320ka plasma -- no faults
410:21:19:000AMPlasmaOk320ka plasma -- no faults
510:39:33:090AMPlasmaOk310ka plasma -- no faults
610:54:10:370AMPlasmaOk330ka plasma no faults
711:08:04:890AMPlasmaOkfizzle -- no faults
811:30:15:230AMPlasmaOk310ka plasma -- no faults
911:48:30:180AMPlasmaOk300ka plasma -- no faults
1012:09:58:560PMPlasmaOkdud, fizzel 240
1112:32:42:060PMPlasmaOksoso
1212:53:51:410PMPlasmaOk300ka plasma no faults
1301:18:20:220PMPlasmaOkstill going
1401:44:20:770PMPlasmaOkgoody
1502:00:10:330PMPlasmaOkfizzle -- no faults
1602:14:55:200PMPlasmaOk310ka plasma -- no faults
1702:35:35:650PMPlasmaOk300ka plasma -- no faults
1802:59:29:930PMPlasmaOk300ka plasma -- no faults
1903:19:37:250PMPlasmaOk300ka plasma -- no faults
2003:38:27:310PMPlasmaOk300ka plasma -- no faults
2103:57:56:300PMPlasmaOk300ka plasma -- no faults
2204:19:20:520PMPlasmaOk310ka plasma no faults
2304:33:48:560PMPlasmaOkok joe
2404:47:43:450PMPlasmaOk330ka plasma -- no faults
2505:02:43:780PMPlasmaOkphysics goof pulled breakers. syst ok
2605:20:57:130PMPlasmaOk300ka plasma -- ef1l comm fault -- fuses ok
2705:52:55:070PMPlasmaOkfizzle -- no faults