Miniproposals

Miniproposal: Date Filed: Title: First Author: Session Leader: Joe Daigle (shots 1-15) Session Leader: Vinny Bertolino (shots 1-15)

Engineering Operator Run Comment

RF CONDITIONING

Session Leader Plans

Physics Operators Plans

Session Leader Summaries

Physics Operator Summaries

Entered: Jul 7 2004 04:36:55:177PM

Author: To Be Determined

Physics Operator Summary for 951103 EF4 Power supply fixing SL: Bertolino/Daigle PO: Wolfe EO: Bertolino/Daigle The goal was to get the EF4 supply working, at least in 2-quadrant mode, and provide enough tests so that the engineers would be convinced it was safe to tie in the reverse half of the supply, enabling four quadrant operation. As it turned out, the day went very well, and we ended up with successful four quadrant operation and even got a good plasma as the last shot of the day. The procedure was the same as yesterday. The hybrid was used to produce coil current waveforms based on an actual shot, to provide realistic and reproducible induced voltages and loads against which to test the EF4 supply. The initial PCS shot was the same as the end of yesterday's run. The first shot used additional filtering on the shunt signal; this did not seem to help very much. The next several shots were spent investigating the use of rectified signals from the primary CT's. The noise on this signal was high frequency spike (SCR type) noise, and was easily filtered to give a signal as clean as the bus rogowski. This was actually used to set the gate blocks, and worked. The threshold for starting the integrator was reduced from 750A to around 100A. However, this signal had the problem that it is unipolar, and extra electronics would need to be incorporated to make it change sign on double swings. It was then decided to try using a higher resistance shunt. A 200 uohm shunt from the EF3 bus (Irby shunt) was used to test this out. It worked very well. The noise level was reduced to around 15amps with a 100Hz filter. After examining the signals, this was tied in to the gate block circuitry on shot 7, and worked fine. The next few shots were attempts to torture the new configuration into failure (single phase or otherwise). On shot 8 I loaded a bipolar demand into the EF4 current programing, taken from 950526005. This worked fine, in the sense that the 2-quad supply went to zero and stayed there. Shot 9 and 10 were drawn to force the current to zero and then to come back up again in the same direction, to simulate a case in which the plasma feedback might cause multiple zero crossings. The supply also handled this test. Shot 11 was a test of the PLC clipping; the forward demand was increased to 1500A and the PLC limit lowered to 850A. The supply actually limited at about 1000A; aside from the usual sloppy PLC calibration, it worked fine. For shot 12, the reverse supply was tied back in (gates enabled). The first test was essentially a repeat of the unipolar current pulse from shot 7, with the final value carefully set to zero, and a positive voltage programmed into the V for the remainder of the shot. This seemed to work well. Shot 13 was a full test of four-quadrant operation, based on a reload of the bipolar waveform of shot 8. The double swing worked fine, with a deadtime of about 110msec. The supply was declared to be working. During next week, the engineers will clean up the patches and replace the Tektronix filter used today, and have a new 200 uohm shunt made, to implement the fix in a permanent way. We were ready to declare success and withdraw, but John Rice requested a plasma to check out a repair on one of his diagnostics. We therefore loaded from 951031020, modified to use steady-state fill, in view of the difficulties encountered on Wednesday with the pulsed startup. Shot 14, at a fill of 1.8e-5 fizzled. Shot 15, at a fill of 1.3e-5, ran fine, a good 600kA, nl04=5e19, full length shot. As noted on 951031, the x-point is still a bit high, probably because the gain on ZXL is too low, but it was well-diverted with good gaps on both sides. Additional observations: The OH2U coax resistance was between 1.1 and 1.7 uohms all day. There seems to be a correlation with machine temperature. The top magnet arms were inadvertently allowed to ratchet up in temperature during the afternoon. Apparently the cooling was switched to manual because we still have bad TC's included in the average. We allowed about 30 minutes between shots 13 and 14 to get the high temp down to -150. The EF4 supply still needs to be tied in to the fizzle detector. Scorecard: Tests 13 Fizzles 1 Plasmas 1 ======= == Total 15

Session Leader Comments

Physics Operator Comments
Nov 3 1995 08:39:43:950AM	951103001	Steve Wolfe	More power supply work today. Start from model from yesterday. Coldstart went ok. Supplies ran ok. EF4 better, but may glitch at .3 sec. Engineers are staring at traces. OH_resist: oh1: 0.2122 uohm oh2u: 1.2753 uohm oh2l: 0.8082 uohm
Nov 3 1995 09:15:40:330AM	951103002	Steve Wolfe	Shot #951103002 Repeat, with new filter on EF4 primary CT's, to see if they will give a more useable signal. The CT signal now looks very good. They need to adjust the amplitude, but noise does not seem to be a problem. oh1: 0.1905 uohm oh2u: 1.3563 uohm oh2l: 0.7957 uohm
Nov 3 1995 10:14:35:010AM	951103003	Steve Wolfe	Shot #951103003 Repeat with new filtered CT signal tied in to tristate logic in EF4. Looks good to me, the P.S. experts are huddling oh1: 0.2050 uohm oh2u: 1.3224 uohm oh2l: 0.8063 uohm
Nov 3 1995 11:10:55:750AM	951103004	Steve Wolfe	Shot #951103004 Repeat, with more board changes. Vinnie is interested in using the bus rog instead, tied in through an AFOL from the interface room. It seems possible, but I don't like depending on a rogowski integrator gate to run power supply internals, especially since we are already depending on it to run the feedback. Too much dilution of multi-point safety. However, we'll look into it. oh1: 0.2122 uohm oh2u: 1.2638 uohm oh2l: 0.8000 uohm
Nov 3 1995 01:44:11:590PM	951103005	Steve Wolfe	Shot #951103005 Going to a higher resistance shunt. They have installed the 200uohm shunt formerly used in the EF3 bus to be used only in the gate block logic of the EF4 supply. For this shot, only look at the signal. Noisy, but it looks like high frequency. Should be fixable. For next shot put in the Tektronix filter. oh1: 0.2815 uohm oh2u: 1.2573 uohm oh2l: 0.7766 uohm
Nov 3 1995 01:55:19:570PM	951103006	Steve Wolfe	Shot #951103006 Filter installed. Looks nice and clean, around 30A noise. Tie it in next shot. oh1: 0.2935 uohm oh2u: 1.2895 uohm oh2l: 0.7849 uohm
Nov 3 1995 02:19:02:030PM	951103007	Steve Wolfe	Shot #951103007 Tied in to tristate control, with filter at 100Hz (was 300Hz last time) Currents look just the same. Looks like it will work. oh1: 0.3189 uohm oh2u: 1.3114 uohm oh2l: 0.7681 uohm --
Nov 3 1995 02:31:53:550PM	951103008	Steve Wolfe	Shot #951103008 Put in a bipolar demand on EF4 by cutting in the trace from 950526005 Extend FD to 1.2, EF4 invert to 0.5 Worked well. oh1: 0.3366 uohm oh2u: 1.3453 uohm oh2l: 0.7511 uohm
Nov 3 1995 02:32:09:640PM	951103008	Steve Wolfe	Shot #951103008 Put in a bipolar demand on EF4 by cutting in the trace from 950526005 Extend FD to 1.2, EF4 invert to 0.5 Worked well. oh1: 0.3366 uohm oh2u: 1.3453 uohm oh2l: 0.7511 uohm
Nov 3 1995 02:43:40:550PM	951103009	Steve Wolfe	Shot #951103009 Ask it to come back on in the same direction after it goes to zero Extend forced invert time to 1.2 sec Never actually got to zero. Try again. Lots of magnetics read errors on the Bp loops. oh1: 0.2109 uohm oh2u: 1.5939 uohm *** It went up this time. oh2l: 0.7163 uohm
Nov 3 1995 02:52:56:800PM	951103010	Steve Wolfe	Shot #951103010 Last time I asked for it just to go a little past zero, and it didn't make it. This time, ask for +1000A before turning back negative. Worked well. went to zero, came back on, went back to zero. Errors on magnetics and hybrid store actions. oh1: 0.1899 uohm oh2u: 1.6882 uohm oh2l: 0.7174 uohm -
Nov 3 1995 03:30:46:960PM	951103011	Steve Wolfe	Shot #951103011 Ask for 1500 Amps in forward direction, to test PLC limit. Vinnie will set "750A", expecting 1000. Seems to limit at just under 1000A. Rest of pulse looks ok, as before. oh1: 0.2674 uohm oh2u: 1.5948 uohm oh2l: 0.7005 uohm -
Nov 3 1995 03:42:58:810PM	951103012	Steve Wolfe	Shot #951103012 Tie in 2nd half of supply. Reload 951103007, tweak to make sure demand is at zero current at end. Also put V positive at end. Seems to have worked. goes to zero at 0.4sec oh1: 0.1947 uohm oh2u: 1.6382 uohm oh2l: 0.7017 uohm
Nov 3 1995 03:56:47:110PM	951103013	Steve Wolfe	Shot #951103013 Reload 951103008 -- bipolar EF4 current (to +3kA) Double swing worked fine. About 110msec of dead time. oh1: 0.1946 uohm oh2u: 1.5651 uohm oh2l: 0.6960 uohm
Nov 3 1995 04:39:20:000PM	951103014	Steve Wolfe	Shot #951103014 Try for plasma : Load 951031020. Get rid of prepulse, use a-side fill & puff. P0=1.8e-5 Turn on ECE tree Cooling the magnet back down is going to take a while. The top arms have been ratcheting up all day because the top was set at 25% manual fill. This was done because there are still bad TC's included in the average, and if it isn't on manual it goes berserk. Fizzle. Got magnetics errors again. oh1: 0.2071 uohm oh2u: 1.3674 uohm oh2l: 0.6647 uohm
Nov 3 1995 04:54:40:830PM	951103015	Steve Wolfe	Shot #951103015 Try for plasma. Drop fill to 1.3e-5 Plasma, full length. 600kA. Looks good. x-point still too high, good gaps all around. EF4 deadtime at just under 100msec. oh1: 0.2729 uohm oh2u: 1.1811 uohm oh2l: 0.6501 uohm -

Engineering Operator Comments
Shot	Time	Type	Status	Comment
1	08:28:53:040AM	Test	Ok
2	09:08:15:920AM	Test	Ok
3	10:07:10:800AM	Test	Ok
4	11:02:38:310AM	Test	Ok
5	01:33:54:000PM	Test	Ok
6	01:50:41:070PM	Test	Ok
7	02:11:39:600PM	Test	Ok
8	02:26:29:050PM	Test	Ok
9	02:38:36:380PM	Test	Ok
10	02:46:25:180PM	Test	Ok
11	03:26:07:700PM	Test	Ok
12	03:38:47:020PM	Test	Ok
13	03:48:40:280PM	Test	Ok
14	04:31:12:230PM	Test	Ok
15	04:41:44:570PM	Plasma	Ok