Alcator C-Mod Run 930914 Information

Back to runs

Miniproposals
Miniproposal:013
Date Filed: 7/16/1993
Title:Discharge development
First Author:Steve Wolfe
Session Leader:Steve Wolfe (shots 1-21)
Miniproposal:019
Date Filed: 7/22/1993
Title:Plasma current feedback control
First Author:John Goetz
Session Leader:Steve Wolfe (shots 1-21)

Operators
Session leader(s):Steve Wolfe,Steve Wolfe
Physics operator(s):Steve Fairfax
Engineering operator(s):Bill Cochran,Frank Silva

Engineering Operator Run Comment
first run after 3 week shutdown. Hope to

Session Leader Plans

Physics Operators Plans

Session Leader Summaries
Entered: Jul 7 2004 03:49:08:587PM
Author: To Be Determined
Run summary for 930914
Session Leader: Wolfe
Physics Operator: Fairfax
Eng. Operator: Silva

Primary MP 013 - Discharge development
Secondary MP 019 - Current feedback

The goals of the run were:
1) Get breakdown and current rise for the first time this month
2) Use double swing capability to extend flattop and perhaps increase current
3) Try Ip feedback

The first was accomplished, and the second partly accomplished. We did not get
to try Ip feedback.

The first attempt at plasma was #13. We got fizzles on 15 and 16, and a good
long plasma on #17. The change between 16 and 17 was to increase the gain on
Br to try to correct a small positive radial field. We ended up with a small
negative Br, but it ran. Current lasted until a hard disruption at 600msec.
OH1 double swung successfully, and plasma current was held up (430:360kA). The
disruption seemed to be a vertical instability, with kappa_FFIT about 1.6.
Shot #19 was somewhat longer, with current held nearly flat (460:430kA for over
450msec).The shot again ended with a vertical disruption at 660msec, in this
case the current going to zero in 1.5msec. Both were good lower nulls. While
kappa was increasing only slowly (according to EFIT), the size was shrinking
fairly rapidly from 300msec on, with both inner and outer gaps increasing. The
minor radius got below 20cm; the major radius was somewhat outside, around 69
cm. The strike points were swept over both inner and outer divertor
structures, and in the slot. The disruption occurred just after the Lazarus
N/Ncrit exceeded 1.0. Despite the fact that the Z-meter was getting very large
Z-eff values, the li's were reasonable: dropping from 1.75 to 1.5; also, the
surface voltage was only 1.25 V (EFIT).

There were only the two "good shots". The fizzles on 18,20, and 21 were
probably due to the disruptions on 17 and 19, although the radial field could
probably still use some tuning.

Suggestions:
-----------
The last two attempts fizzled, but the programming changes should be tried:
they consisted of changes to OH1, OH2 voltages and EF1 current to try to
restrict the shape changes, and gas puffing to increase the density (which was
VERY LOW). Further, it would be reasonable to move the RCUR in a couple of
cm, and allow the limiter to be moved in. If fizzles persist, some tuning to
the radial field may help; also the index at commutation is a bit negative.

Problems:
---------
Power supply tests/problems occupied the first 12 shots. The main difficulty
was that the OH2 stopped listening to the programming when the alternator
voltage dropped. This was traced to the limit on inversion end stop
proportional to alternator voltage, and got fixed by shot 12. There was also a
problem with the gas when A-side valve had hybrid enable. The symptom was a
fault appearing at the transition to PULSE. This was "solved" by
not enabling A-side, and only using B-side upper. There seems to be a drain
on the +15V supply when the A-side enable is turned on. This will be looked at
after the run. We also had incorrectly set the switch close times on OH1,OH2,
and EF1, leading to a failure of some of the switches to close properly. This
was corrected at shot 14.

There were several TF crowbars.

After these problems were cleared, we found a very worrisome
Hybird problem: Despite the fact that the PID on the Ip wire was zeroed, it
put a 1V signal on the V_EF3 output beginning at -0.046 seconds, when the
controller was switched. This was discovered on shot 14. This may be a
hardware problem. We were unable to zero the controller using the zero button,
apparently due to a software problem: it just sat there after typing "yes" in
response to the "Are you sure?" prompt. We then got rid of the controller by
setting it to "TF" and manually zeroing the points. After the run, Steve Horne
determined that the PID on Wire 15 (Ip) is indeed cooked; it is putting out
10V. He is attempting to fix it.

Score card:
-----------
Tests: 12
Duds: 2 (1 bad switch times, 1 EF3 error due to Hybrid malfunction)
Fizzles: 5
Good shots: 2
--
Total 21

Physics Operator Summaries
Entered: Jul 7 2004 04:36:50:333PM
Author: To Be Determined
Physics operator summary
September 14, 1993

Session Leader: Steve Wolfe
Physics Oper: Steve Fairfax
Eng. Oper: Frank Silva
Power Sys: Bill Cochrane
State; Pierson
Aux. Sys. Joe Sorci

This run was the first with double-swing available on OH1, OH2U&L, and EF1U&L.
The first 7 shots were devoted to patching a problem that was causing the
OH2U to trip off at 20 kA and to fixing the gas fill/puff system. The OH2
problems were fixed on shot 6. The gas puff logic was changed before shot 7;
the system still faulted when it went into PULSE state.

Shot 8 11:21 AM A-side gas pulsing disabled as it seems to be the cause
of the fault going into PULSE state. All power supplies
except EF4 are on so as to test the system without
any chance of getting plasma (until the pulse gas is
shown to work.)

Pulse gas did work on Shot 8.

The OH2U power supply did not trip its breaker after
a type 2 fault. It took about 30 minutes to find and
correct the problem.

Both OH2 supplies reduced their output voltage to 50
volts when the TF supply went into regulation and the
alternator output voltage dropped. The regulators are
designed to limit inversion voltage as a function of
power supply current and AC input voltage, but the
gain on the input voltage seems too high. The voltage
to the multiplier was raised from 5 to 7.5 volts. It
took me and Steve W. over 40 minutes to figure out this
problem.

Shot 9 1:06 PM Test of OH2 and TF supplies to see if fix worked.

Still no good. Changed margin angle setting
from 60 to 40 in both OH2 power supplies.

Shot 10 1:27 PM Test of OH2 and TF with commutation on OH2.

Shot 11 1:39 PM Same as shot 10 with no TF

Shot 12 1:50 PM Decreased OH2 inversion end-stop from 40 to 30
Increased PID gain in OH2 wires from 1.84 to 5.

Found that OH2U switch was not closing due to my
incorrect setting on switch close timing.

Increased EF1 PID gains from 2.5 to 5.
Set -I limits to -10 kA OH1 and OH2

Shot 13 2:17 PM Plasma Attempt. OH1 switch did not close. Found
current demand for OH1 was 2 amps when swtich was
closing. Delayed TF start 300 msec to -1.35 sec.

Shot 14 2:47 PM Good pulse but no plasma. EF3 was being turned on at
-46 msec by the Ip controller even though the PID gain
was set to 0.00000 for all time. This is either a
failure in the multiplier hardware or a software bug.
Tried a work-around of setting the Ip controller to run
TF current (not connected).

Shot 15 3:35 PM Fizzle

Shot 16 3:48 PM Fizzle. Increased Br gain from 1 to pi.

Shot 17 3:58 PM 440 kA. Changed OH1 and OH2 volt program to hold current
at 440 ka. Changed EF1 current program to try and hold
elongation constant at 1.5 or so.

Shot 18 4:15 PM Fizzle. Tried to let Moly monitor and NPA open for
next shot. Changed wall setpoint to -10 degrees.

Shot 19 4:25 PM Plasma. Disrupted at 600 msec. Changed EF1 current
for more rapid ramp-down to control elongation. Wall
temperature at -10 degrees.

Shot 20 4:48 PM Fizzle. Changed nothing.

Shot 21 5:03 PM Fizzle.

Summary: Not bad for a Tuesday after a 3-week shutdown. Double swing seems to
work with only minor new constraints on the operator. We probably
have a few days of wall conditioning ahead of us. The PCS bug that
caused the EF3 problem on shot 14 needs immediate attention.

My method for closing the SCR switches prior to the shot is to demand
full current rather than a ramp. This results in some odd-looking
currents before switch closure but is otherwise harmless.



Session Leader Comments
Sep 14 1993 07:41:05:380AMSteve WolfeRun plan for 930914
--------------------
The goals are:
1) Get breakdown and current rise for the first time this month
2) Use double swing capability to extend flattop and perhaps increase current
3) Try Ip feedback

Approach:

Start from an old shot, eg 930817022. Extend inversion times on all supplies to
1second. This shot was from the pellet run, but we won't be using pellets. It
had moderate elongation, peak current 440kA at t~0.2sec, and lasted, with
decaying current,until 600msec. Only OH1 ran out of current (at 380msec). If we
pre-program OH1 voltage to remain constant at ~-150V until 750msec (or longer?)
we ought to be able to hold Ip. A possible target: hold Ip=450ka +/-50ka from
200ms to 750ms, ramp down from 750ms to 1sec.

Try turning on Ip feedback. Two and a half possible controllers:
OH1 only (maybe OH2's in turns ratio)
Balanced all coils controller (essentially turns ratio)
Test by tring to change slope of current, and/or gas puffing.

Adjust OH2, EF1, EF2, EF4 to hold more or less constant equilibrium (ie
kappa,triangularities, as well as Ip, within 10%

Raise current (limit 1MA) by holding higher voltage on OH1 and OH2 between 100ms
and 200msec to maintain current ramp rate around 2MA/sec.

Sep 14 1993 09:40:09:800AM930914001Steve WolfeOh2-only test shot at 20kA. OH2U trips out on forward current rise,
as on last tests yesterday. OH2L looks ok.

Sep 14 1993 10:06:30:920AM930914002Steve WolfeChanged the inversion end stop on OH2U and try again.
Still no good

Sep 14 1993 10:10:36:990AM930914003Steve WolfeLoad setup as for actual run. Changed the OH2U&L to be what they are,
instead of CUSTOM predictors with sum and difference. The idea is the
supplies may behave better with voltage program from hybrid.

Result: Same problem.

The RF Limiter has now been moved to .95 in preparation for the plasma
part of the run.

Sep 14 1993 10:22:04:340AM930914004Steve WolfeFound (and fixed?) a problem in the OH2U regulator.

Result: No change. That wasn't it.

Sep 14 1993 10:49:22:370AM930914005Steve WolfeChange conduction monitor threshold from 6kA to 10kA on both OH2s.

We are aslo having a problem with the gas logic. It gives a fill fault
on entering PULSE, and the hybrid signal doesn't gt through to the
valve. M.Graf is working on it.

Result: OH2s both ran ok

Sep 14 1993 11:01:34:610AM930914006Steve WolfeAnother OH2 test -- load from 930913017 (test from yesterday with
double swing).

Both ran, but not symmetric. OH2L saturates at -.85sec at 24kA, while
OH2U doesn't reach this current until -.51sec.
Apparently the inversion end stops are not set the same. Change it back
for next shot.

Sep 14 1993 11:11:30:800AM930914007Steve WolfeOH2 and gas test:

Gas still faults when it goes into pulse. OH2s look good.

Sep 14 1993 12:10:59:360PM930914008Steve WolfeTurn everything on except EF4, turn off Hybrid enable on A-side.
This should prevent us from getting plasma, but test everything else,
including whether turning off A-side fixes the gas problem.

Vessel set point has been lowered to 5C. POrts are 30C, coils are 5C.
Gradient on B-port is now 87C. Others have improved.

Gas went ok.

OH2s didn't get to proper current. It looks like neither
OH2 is following its demand after -.28sec. They go down to 50V while the
demand is going up. Also, the OH2s did not commutate.

Sep 14 1993 01:13:09:310PM930914009Steve WolfeThe problem on OH2 seems to be related to the alternator voltage drop
when the TF starts to regulate. The solution proposed is to change the
gain applied to the inversion endstop. This shot will be OH2 and TF
only.

Also fixed before this shot: the breaker was not pulling out on a
fault on one of the OH2 cabinets. This was fixed and tested.

It is confimed that the hybrid enable to A-side valve is responsible
for the gas fault. Apparently the 15V supply is pulled down when this
switch is on. We will run without A-side valve.

OH2s reached 21ka at -.34sec, but comes down to 15kA by 0sec. Looks
like the same problem.

Sep 14 1993 01:38:50:300PM930914010Steve WolfeBetter, but still droops prior to commutation. It did commutate this
time.

Sep 14 1993 01:44:22:920PM930914011Steve WolfeOH2 only. No TF this time.

Result: Looks better. No droop, but it never reaches target point.
May need more gain?

Sep 14 1993 02:11:33:720PM930914012Steve WolfeIncrease gain from 1.85 to 5 on PID (P) on IC_OH2s.
INcrease margin angle.

Turn everything on, and try for plasma.

Result: TF crowbarred. OH2U only got to 10kA and then dropped.
OH2L got to 18kA and flattopped. Both OH2s were driven up to greater
than 25kA after commutation. OH1 double swung to -20ka, probably as
it should have.

The switch times were wrong. The OH2U switch didn't close because
the close time was at -1.0 and the current was zero.

Sep 14 1993 02:29:21:620PM930914013Steve WolfeINcrease gain on EF1s to 5.0
Change switch times to same as on 8/17.
Set limits on reverse current to -10kA on OH1&2s.

Result:
Crowbar.
OH1 switch close was still early (-.94 instead of -.85)
OH2s look ok now.

Sep 14 1993 03:32:37:260PM930914014Steve WolfeChange OH1 switch close to -.85 and changed the program to insure
there is demand at -1.0 sec.
Delay TF 300ms.
The channel 34 TV camera is off; possibly a breaker blown in the cell.

No crowbar.
EF3 had an Mout on at -0.046 sec that seems to be a PCS program error?
The Ip PID was zero, but nevertheless we were getting an output from it
on the Mout. It could be that the PID DAC isn't being loaded with the zero
properly.

Sep 14 1993 03:40:53:220PM930914015Steve WolfeTurn off the M's on the Ip wire (zero didn't work so we set it to TF current
with zero gain.

Result:
Fizzle (hey, it's better than nothing!)

Br is slightly positive, Bz' is slightly negative.

Sep 14 1993 03:53:22:180PM930914016Steve WolfeNo change.
Another fizzle.

Sep 14 1993 04:08:57:170PM930914017Steve WolfeIncrease Br gain from 1 to 3.14

Long pulse! Looks good. > 400kA, lasts till 600msec.
Lower null from 280 msec to 600
Kappa grows to 1.6 - note that EF1s were held constant current.
Vertical instability grows slowly from 400msec, finally explodes at down at 600+

Sep 14 1993 04:22:45:850PM930914018Steve WolfeINcrease V_OH1 voltage to -200V in flattop phase
Increse V_OH2 from -20 to -40V
IC_EF1 2700a 350msec and 1500a at 800msec

Result:
Fizzle

Sep 14 1993 04:39:48:560PM930914019Steve WolfeLower wall set point to -10C
Enable Mo monitor gate valve.

Result:
Good long shot. Almost 700msec.
Density is only 2e19 later in shot. Hards growing to 8 units.
kappa around 1.5 -> 1.6. Final disruption is vertical.

Sep 14 1993 04:53:40:420PM930914020Steve WolfePulse gas2 at 450-550, 150-300ms
Increase V_OH1 from -200 to -300V
V_Oh2 from -40 to -50V
I_EF1 goes to 500A at 800msec.

result:
Fizzle - TF crowbar at 1sec

Sep 14 1993 05:09:01:160PM930914021Steve WolfeNo changes. see if it comes back.

Result:
Fizzle.

Physics Operator Comments

Engineering Operator Comments
ShotTimeTypeStatusComment
109:35:21:830AMTestBadoh2 type 2 fault more than one scr conducting
209:50:34:940AMTestBadsame as one
310:07:47:790AMTestBadsame as 2 and 1 except no double swing
410:19:41:260AMTestBadsame
510:45:24:190AMTestOkOH2s forward
610:55:37:020AMTestOkdouble swing ok
711:05:36:590AMTestOkpulse gas problems
811:20:52:760AMTestOkall supplies except ef4
901:06:31:940PMTestOkTF & OH2
1001:27:33:610PMTestOkoh2 test
1101:39:33:390PMTestOkOH2 only
1201:51:24:760PMPlasmaOkno bd plasma shot #1
1302:16:49:430PMPlasmaBadFaults
1402:47:10:130PMPlasmaOkno breakdown
1503:32:07:940PMPlasmaOksuspected flash
1603:43:42:880PMPlasmaOkgood
1703:55:35:220PMPlasmaOk440ka .6sec
1804:14:47:510PMPlasmaOkno faults
1904:25:50:550PMPlasmaOkgood
2004:44:45:870PMPlasmaOkps faults, fizzle
2105:04:09:580PMPlasmaOkbye !