Alcator C-Mod Run 1060208 Information

Back to runs

Miniproposals
Miniproposal:298a
Date Filed: 5/20/2002
Title:C-Mod Plasma Startup
First Author:Jim Irby
Session Leader:Steve Wolfe (shots 1-24)
Session Leader:Marco Ferrara (shots 5-24)

Operators
Session leader(s):Steve Wolfe,Marco Ferrara
Physics operator(s):Joseph Snipes
Engineering operator(s):Andy Pfeiffer,Ed Fitzgerald,Bill Parkin

Engineering Operator Run Comment
Complete MP 298a. Power Tests, MP298A Plasma startup Session leaders: Wo

Session Leader Plans

Physics Operators Plans
Entered: Feb 7 2006 02:49:40:790PM
Author: Joseph Snipes
Engineering setup for Wednesday 8 February 2006

MP298A Plasma startup

Session leaders: Wolfe, Irby
Physics operator: Snipes

Overnight ECDC in D2, and bake at 60 C.

Run begins at 09:00 and ends at 17:00

Power systems as on 1051101001 (0.8 MA, 5.4 T, lower single null)

A-coil configuration: normal [+Dtop -Dbot -Jtop +Jbot] and 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 D2 Hybrid enabled (PG3)
leave NINJA as is Leave DISABLED

Enable the following gate valves, assuming no vacuum problems:
ECE, VUV

Torvac setup: Gate valves open

Run plan: Start from 1051101001 and try to get reproducible full-length
plasmas. If successful, use the plasmas to clean up the machine and
slowly reduce the hydrogen content.

Session Leader Summaries
Entered: Feb 8 2006 05:42:38:293PM
Author: Steve Wolfe
Power system tests and
MP#298a C-Mod Startup

SL: Irby, Wolfe, Ferrara
PO: Snipes
EO: Pfeiffer, Cochran, Fitzgerald, Parkin

This was quite a successful first plasma run day of the year. 

Yesterday's problems were dealt with early. 
Gary teed a meter off the PG3 (B-main) driver cable and we ran a test cycle
(shot#1) before we closed up. On this test the valve opened and we got
lots of gas, although the response was slow. It is likely that the piezo just
needed to be exercised harder than we were hitting it yesterday, 

Bill Cochran "fixed" the EF4 reverse bridge power supply problem, or at least
by the time he finished inspecting it and not finding any problems it started
working again. Shots#2-4 were successful EF4 tests. 

Marco got his Chopper characterization tests on Shots#5-7. Data look good, and
he should be able to both construct a good power supply transfer function
model for simulation and analyze the power supply noise in the vertical
control loop, and thereby also characterize the plasma contribution. 

The dpcs_fizzle asynchronous control routine was tested with all PF supplies
(and gas valves) on shots #8-12. The initial test, using an external program
(SOURCE_SHOT=1060110999) worked quite well, successfully ramping the OH2's and
OH1 positive, though not reaching the program value of 15kA. All of these
tests were done with the OH fizzle delays set to 300msec (and the TF delay to
600msec). These are the recommended settings. On shots 9,10 we imported seg4
(async) from 1050110999 into the model, changed Source_Shot to $shot, and made
some minor adjustments, increasing the gain on the OH2's and changing the
V_out for the chopper to 0 from 1000V (which had been intended to turn the EFC
current off). The behavior on shot 1060208010 seemed good, so I changed the 
SOURCE_SHOT to point to that. Shot 11 tested this setup. The dpcs_fizzle 
settings have been left with that setup, and I recommend that this be
maintained. The settings are in
DPCS::TOP.LOADABLES.PROCEDURES:ASYNC_PROS:PROCEDURE_01 
Turning this node OFF disable the digital fizzle detector and turning it ON 
enables it. 

Shot#13 was the first plasma attempt, taken around 12:30 when the TF finally
got down to temperature (see below). It was successful, getting about
150msec and 500kA. Shot#14 was slightly longer and higher current. It was
discovered that the TCI signal was not being sent to the control system, 
so we disabled nl04 feedback and let it run off the walls (no pg3 puff
in seg2). The run continued with duds, fizzles, and a couple short
plasmas. Then we got full length 800kA plasmas (loop voltage ~1V) on shots #22
and 24. The argon/neon puff was turned on for the last shots, and John and
Alex report that HIREX and NeSoX are working.  

Despite the duds and fizzles, the oh coax resistance readings remained low
today. It's probably too early to declare the dpcs_fizzle sequence to be
the cure for this problem, but so far so good!

The major problem today, as yesterday, is the slow cooling of the C-right
vertical TF arm, which seriously limited the shot cycle today. Also, the TCI
signal needs to be restored to the control system for density feeback,
although we can use NL_ZMETER in the short term; Earl revived that signal
during a cell access. The bolos and Neuts/Hards signal suffered a loss of
trigger on their CAMAC digitizers after the cell access around shot 15 or 16.
The neuts/hards signal digitized at DPCS indicates we had no runaways at all
today. 

Minor problems today included a few data system hangs, and my Web-based logbook
sessions kept blowing away Firefox (or the other way around).  


Physics Operator Summaries
Entered: Feb 8 2006 05:00:22:997PM
Author: Joseph Snipes
Overall it was a very successful startup run. The run started out with PLC tests of EF4 that worked fine.
Then, there were EFC tests designed by Marco Ferrara that also worked fine. Then, it came time to try
plasmas and we were slowed down significantly by cooling problems, particularly associated with the TF.
Nonetheless, shot 1060208013 was the first plasma attempt and we got a short plasma out to 0.5 MA. There
were a couple of other plasma shots (14 and 17) interleaved with fizzles or duds before we finally got a
full current (0.8 MA) full length (2.0 s) plasma on shot 1060208022. I started with Hutch's magic startup
from 1051101023 and it worked initially but then seemed to drift off during the day. Attempting to
anticipate gas evolving off the walls on shot 23 was ineffective and resulted in a dud. Returning to the
startup from 22 for shot 1060208024 resulted in another full length shot, so the startup from 22 and 24
seems to be the best at the moment and would probably be a good place for Ron to start tomorrow.

Shot,Duration(s),Ipmax(MA),Result Shot,Duration(s),Ipmax,Result
| 01 0.000 0.003 Dud | 02 0.000 0.009 Dud |
| 03 0.000 0.005 Dud | 04 0.000 0.005 Dud |
| 05 0.000 0.002 Dud | 06 0.000 0.003 Dud |
| 07 0.000 0.002 Dud | 08 0.000 0.026 Dud |
| 09 0.000 0.025 Dud | 10 0.000 0.025 Dud |
| 11 0.000 0.026 Dud | 12 0.000 0.026 Dud |
| 13 0.148 0.517 Plasma | 14 0.250 0.705 Plasma |
| 15 0.000 0.024 Dud | 16 0.000 0.025 Fizzle |
| 17 0.381 0.791 Plasma | 18 0.000 0.026 Dud |
| 19 0.000 0.025 Fizzle | 20 0.000 0.024 Dud |
| 21 0.000 0.025 Fizzle | 22 1.971 0.823 Plasma |
| 23 0.000 0.024 Dud | 24 1.941 0.825 Plasma |

Date 1060208: 24 shots. 5 plasmas. 3 fizzles. 4 duds. 12 tests. 0 System Error.
(Plasma = Ip gt 50.0 kA Dud = H_alpha lt 0.40 V)
Total plasma duration: 4.69 s (exceeding -100 kA)

Session Leader Comments
Feb 8 2006 10:00:19:097AM1060208002Steve Wolfe
Shot#2: successful reverse bridge ef4 test.
mflux field looks like good vertical field, which
implies all flux loops are ok. 
Got gas ok from pg3. 
Next shot: another ef4 test with lower plc current setting
(2kA instead of 4kA as on shot 2) 

Feb 8 2006 10:05:54:407AM1060208003Steve Wolfe
Shot#3: successful reverse bridge ef4 test (1kA each).
Got gas ok from pg3. 

Next shot: double swing ef4 

Feb 8 2006 10:42:06:800AM1060208006Steve Wolfe
Shot#5: Marco's test on EFC with zero volt demand
Current average is ~1750A, no obvious increase in 
noise envelope on Aout ZCUR. 

Feb 8 2006 10:45:28:157AM1060208006Steve Wolfe
My logbook has blown away twice now. 
Shot #6: Looks good with envelope decreasing 
as frequency increases. Looks good. 

Feb 8 2006 10:59:05:870AM1060208007Steve Wolfe
Shot#7: Try the last case, slower frequencies. 
Looks good . 
Next: PF only tests with digital fizzle detector

Feb 8 2006 11:01:54:353AM1060208008Steve Wolfe
Shot#8: Load as for plasma: seg1 from 1061101023,seg2 from 1061101007
TF fizz delay to 600ms, OH's at 300ms
\HYBRID::TOP.DPCS.LOADABLES.PROCEDURES:ASYNC_PROS:PROCEDURE_01
turned on in hybrid tree. 

Enable, gas, PF's, EF3, not TF


Feb 8 2006 11:12:12:223AM1060208008Steve Wolfe
Shot#8: Load as for plasma: seg1 from 1061101023,seg2 from 1061101007
TF fizz delay to 600ms, OH's at 300ms
\HYBRID::TOP.DPCS.LOADABLES.PROCEDURES:ASYNC_PROS:PROCEDURE_01
turned on in hybrid tree. 

Enable, gas, PF's, EF3, not TF
Reverse swing works, but doesn't get to the targets (15kA)
OH1 gets to +7475A, OH2U,L to ~ +8500A by the time the
real fizzle (with delay) cuts in. 
The OH1 voltage is railed, so I can't do any more on that.
OH2's have a spike at .174sec, which is probably the gain switch.


Feb 8 2006 11:28:00:780AM1060208009Steve Wolfe
Shot#9: Import seg4 from 1060110999
Change initial gains on oh2u and oh2l from 1 to 2.
In traverser, change HYBRID::TOP.DPCS.LOADABLES.PROCEDURES:ASYNC_PROS:PROCEDURE_01:SOURCE_SHOT
from 1060110999 to $shot

Still only gets to +9500A on oh2's, probably just fine.
Voltage jumps to 200V, then drops to ~80V or so. 
We could tune it, but that will probably do for now.

The chopper is being slammed off, then comes back up 
when the hw fizzle detector turns on. I think I'll
get rid of that, while we're waiting. I actually had a
standard V waveform, which means it was at +1000V for 
about 100ms. Set it to zero across the board.

Feb 8 2006 11:32:37:513AM1060208010Steve Wolfe
Shot#10: Modified Chopper vout signal to be zero all the
way across. I think this will be a good place to leave it. 

Next shot: Change tree node source shot to point to seg4 
of 1060208010


Feb 8 2006 11:43:46:493AM1060208011Steve Wolfe
Shot#11: Change tree node source shot to point to seg4 of 1060208010
Note: seg4 in the model is still turned on async. It should deal with 
that and then use the other one. I've never tested this combination.

Looks ok. Nothing obvious in the log file but it didn't seem to have
a problem. 

next shot: Reload seg4 from the original shot, and be sure it's 
turned off. 

From now on (until we change it), the source for the fizzle detector
will be 1060208010 segment 4. 

Feb 8 2006 12:38:53:097PM1060208013Steve Wolfe
Shot#13: Try for plasma.
Got one! 150msec, 500kA. 
Dirty, but bright. 
Good startup. 
Z-meter signal is missing. 
Looks like plasma is running low? 

Feb 8 2006 01:33:31:983PM1060208014Steve Wolfe
Shot#14: Repeat. Got 250msec, 680kA

Irby reports that the tci feedback signal is inverted!
Doesn't matter, because there is no signal on channel 80
at the input of the dpcs.

next shot: turn off nl04 feedback

Feb 8 2006 02:37:25:003PM1060208017Steve Wolfe
Shot#17: Plasma to almost 400ms
Still no input on dpcs channel 80 (nl04)


Feb 8 2006 03:03:29:760PM1060208017Steve Wolfe
Shot#17: Plasma to almost 400ms
Still no input on dpcs channel 80 (nl04)
Had an oh1 self-power circuit fired after the 
disruption. The dpcs_fizzle demand on the oh1 is
higher than it can actually handle, by at least a 
factor of two. Maybe I should back this down, but
I don't think it is responsible. 

The gains are initially 1 and jumping to 2 after 100ms,
and the Mout value is 12V and switches to 23V after 100msec
(clipped to 10V at the output) and the supply stops responding 
around 5V or so.  THe controller gain is .05775, same as in 
segment 1.  I could reduce the gain to 0.5 across the board,
and it would probably not change the behavior. 


Feb 8 2006 04:14:07:070PM1060208022Steve Wolfe
Shot#22: Plasma full length
TCI is otl. zmeter is falling throughout
the shot. No hards as far as I can tell from 
channel 84 of dpcs; the camac is not triggering 
on neuts/hards or the bolos

EFIT didn't run

Feb 8 2006 04:47:25:283PM1060208024Steve Wolfe
web-based logbook (firefox) just blew itself away for the third or
fourth time today. 

Feb 8 2006 04:50:33:313PM1060208024Steve Wolfe
web-based logbook (firefox) just blew itself away for the third or
fourth time today. 

Shot#24: full length plasma. 

Physics Operator Comments
Feb 8 2006 07:38:45:773AM1060208001Steve Wolfe
Shot1: Set up a pulse gas 3 test 
Seg1 only, pg3 programmed for 100V from [-.768,+.286]
Fizz det at .08
There is a meter on the tee before the piezo; someone
will watch for signal downstairs during no power cycle.


Feb 8 2006 07:51:31:800AM1060208001Steve Wolfe
Got gas, but somewhat delayed. Ended up going to the mTorr 
range. Looks like about a 200msec delay between start of the
pulse and the start of the strong rise (4mTorr/sec gauge).

Back off on pulse length, but leave it on for later tests to 
keep exercising the valve. 


Feb 8 2006 07:54:40:830AM1060208001Steve Wolfe
Shot 2 (maybe): loaded pg3 pulse starting at -.2sec
We'll wait until later to take more pulses. 

Wolfe exiting PCS. 


Feb 8 2006 09:43:00:290AM1060208001Joseph SnipesShot 1: Gas test run by Wolfe before I arrived.
Feb 8 2006 10:05:15:827AM1060208002Joseph SnipesShot 2: PLC test of EF4 supply following extensive EF4 power supply tests in the power room.
Feb 8 2006 10:05:45:657AM1060208003Joseph SnipesShot 3: Another PLC test of EF4 supply and a gas test as well.
Feb 8 2006 10:28:25:367AM1060208004Joseph SnipesShot 4: Loaded 1051031011 to do a double swing of EF4 test shot.
Feb 8 2006 10:34:32:583AM1060208005Joseph SnipesShot 5: Loaded /home/mferrara/1060207500 to do an EFC test with no fast feedback on ZCUR.
Feb 8 2006 10:46:49:187AM1060208006Joseph SnipesShot 6: Loaded /home/mferrara/1060207501 to do an EFC test with fast sine waves.
Feb 8 2006 10:51:35:123AM1060208007Joseph SnipesShot 7: Loaded /home/mferrara/1060207502 to do an EFC test with slow sine waves.
Feb 8 2006 11:04:52:617AM1060208008Joseph SnipesShot 8: Loaded 1051101023 and then loaded segment 2 from 1051101007, which is a standard lower single null
0.8 MA, 5.4 T plasma segment. Segment 1 from 1051101023 should be a reasonable startup to try dubbed Hutch's
magic startup by Wolfe. For this shot, only the PF and EF3 will be on for a fizzle detector asynchronous
switch in the absence of plasma current.
Feb 8 2006 11:19:44:563AM1060208009Joseph SnipesShot 9: Imported 1060110999 segment 4 into segment 4 and turned it on and clicked the sync button to make it
asynchronous. Then, changed the gain settings on OH2U and OH2L to 2.0 at the early time at -0.2 s. This is
another test pulse with PF and EF3 only.
Feb 8 2006 11:33:05:327AM1060208010Joseph SnipesShot 10: Repeat setting the EFC voltage to zero throughout the pulse.
Feb 8 2006 11:38:41:917AM1060208011Joseph SnipesShot 11: One more test shot to ensure the changes were made correctly.
Feb 8 2006 12:18:09:260PM1060208012Joseph SnipesShot 12: Imported segment 4 from 1051101023 and turned the segment to synchronous and off. Try one more test
with PF and EF3 only.
Feb 8 2006 12:20:38:993PM1060208012Joseph SnipesWaiting for TF cooling . . .
Feb 8 2006 12:40:09:970PM1060208013Joseph SnipesShot 13: Attempt plasma with all breakers on the coils closed. No PCS changes. Got plasma up to 0.5 MA.
Disrupted at 0.13 s.
Feb 8 2006 01:08:27:477PM1060208013Joseph SnipesCooling problems on EF3 . . .
Feb 8 2006 01:32:06:827PM1060208014Joseph SnipesShot 14: Repeat. Almost 250 ms plasma up to almost 0.7 MA. Density feedback
is inverted. Try turning off the gain on nl04 for shot 15.
Feb 8 2006 01:39:36:997PM1060208015Joseph SnipesShot 15: Turned off nl04 gain. Dud!
Feb 8 2006 02:14:19:420PM1060208016Joseph SnipesShot 16: Increased prepuff by 3 ms to 24 ms duration. Fizzle!
Feb 8 2006 02:32:01:273PM1060208017Joseph SnipesShot 17: Reduced EF4 current by 10 A to -1335 A. Better shot out to 0.38 s and up to 0.75 MA.
Feb 8 2006 02:54:39:280PM1060208018Joseph SnipesShot 18: Repeat. Dud! Br0 jumped down by -1 mT for some strange reason.
Feb 8 2006 03:11:47:383PM1060208019Joseph SnipesShot 19: Added 0.75 mT offset to Br0 to compensate for the jump in Br0 from shot 18. Fizzle! Br0 increased
more than requested.
Feb 8 2006 03:41:11:403PM1060208020Joseph SnipesShot 20: Reduced Br0 offset to 0.25 mT. Try again. Dud!
Feb 8 2006 03:53:37:070PM1060208021Joseph SnipesShot 21: Reduced Br0 offset back to 0. Decreased EF4 current by 5 A to -1340 A. Increased prepuff by 5 ms
to 29 ms duration. Fizzle!
Feb 8 2006 04:13:27:710PM1060208022Joseph SnipesShot 22: Decreased EF4 current by another 5 A to -1345 A. Decreased prepuff by 4 ms to 25 ms duration.
Try again. Good shot. Full length 0.8 MA all the way down out to 2 s duration.
Feb 8 2006 04:30:43:813PM1060208023Joseph SnipesShot 23: Added 2 psi of neon to the 1.2 psi of argon in B-side lower PG1 and enabled it from 0.3 to 0.35 s.
Reduced prepuff by 6 ms to 19 ms duration to attempt to anticipate the excess gas evolved from the walls.
Dud!
Feb 8 2006 04:51:35:733PM1060208024Joseph SnipesShot 24: Returned prepuff back to 25 ms duration. Good full length shot near repeat of shot 22 though it
disrupted at 1.94 s. Startup looks a bit worse than on shot 22 though.

Engineering Operator Comments
ShotTimeTypeStatusComment
107:41:05:680AMTestOk
209:52:53:990AMTestOk
310:02:38:890AMTestOkEF4 2Ka test
410:19:35:103AMTestOkEF4 doubleswing
510:31:01:693AMTestOkEFC only
610:41:51:080AMTestOk
710:50:10:890AMTestOkefc
811:02:56:947AMTestOkTest no TF
911:18:09:457AMTestOkTest, No TF
1011:28:57:437AMTestOkTest, No TF
1111:38:04:217AMTestOkTest No TF
1211:58:32:567AMTestOkTest, No TF
1312:31:25:757PMPlasmaOk
1401:22:50:813PMPlasmaOk
1501:37:24:293PMPlasmaOk
1602:11:17:123PMPlasmaOk
1702:29:29:680PMPlasmaOkPF breaker pulled after disruption
1802:48:14:330PMPlasmaOk
1903:05:44:150PMPlasmaOk
2003:39:20:370PMPlasmaOk
2103:52:05:523PMPlasmaOk
2204:06:34:690PMPlasmaOk
2304:29:09:987PMPlasmaOk
2404:46:04:097PMPlasmaOk