Miniproposals

Miniproposal: 042 Date Filed: 4/6/1994 Title: Power Supply and Hybrid Dynamics Characterization. First Author: Gerry Tinios Session Leader: Gerry Tinios (shots 1-16)

Engineering Operator Run Comment

dynamics continued

Session Leader Plans

Physics Operators Plans

Session Leader Summaries

Entered: Jul 7 2004 03:49:09:710PM

Author: To Be Determined

Session leader summary for runs 940426,27 Tinios (Session Leader) , Horne(Physics Ops) Run 940426 was mainly power supply tests. The first six shots were individual power supply tests under PLC control. Then we had to wait for TF to cool down, so we could use it for 5T shots. It turned out that the PLC reading for TF temperature as higher than actual. The first shot (007) where all power supplies were used under hybrid control was a complete failure. Only the TF and the chopper worked properly. Then we went back to running under PLC control with no TF. By shot 11, we completely gave up on EF2's (they would not fire under hybrid control) and EF3 (it showed faults under any kind of control) for the rest of the run. From shot 11 on, we ran under hybrid control with TF, but without EF2's and EF3. We tried to ramp up the OH/EF coils slowly using current feedback, then keep them on a flattop using voltage programming (no feedback). Then ramp them down using current feedback. This would be a test of the starting point for the shots for MP# 42. The voltages programmed for the second part had to be adjusted. Then it turned out that the third part (ramp down) had been programmed incorrectly. In shot 13, we forgot to load the voltage programming for the second part. The last two shots were completely current-controlled ramp up-flattop-ramp down. Some glitches in the rogowski signals were traced to the fact that the commutation resistors in the OH coils and EF1coils had been left in by mistake and then switched off shortly after t=0. It was discovered that our inability to fire EF2's under hybrid control was due to the fact that the demand was zero at the time when it was switched on. By switching them on a little later, when the hybrid was sending some demand to them, the problem was fixed on the following day. Run 940427 was the run for MP#42. The plan was to ramp up OH/EF currents with current feedback to values taken from the flattop of shot 931020003 which will be used for startup. Then turn off current feedback and feed the coils with enough voltage to fight resistive losses and modulate the voltage demand to one of the power supplies by various frequencies and observe the output of the power supply to determine its tranfer function. We wanted to do this with the TF on, to take into account the effect of the alternator slowing down during the shot. Frequencies up to 200 Hz were drawn by the wavegen. Higher frequencies were created by a function generator plugged into the Ip reference of the hybrid. On the technical side, the EF2's worked fine; EF3 kept giving faults, so we stopped using it after shot 4. It had to be left in the fault state, so that it can be inspected later. The TF kept crowbarring and it was not used after shot 8. Luckily we were able to get the OH2 shots done with the TF present. This is good since OH2's are important for vertical control and they are fed by the alternator. In shots 2-6 the OH2's were modulated at frequencies of 20-500 Hz. In shots 7-10 the chopper demand was modulated at frequencies of 20-1000 Hz. In shots 11-12 the EF2's were modulated at frequencies of 20-300 Hz. In shots 13-16 the EF1's were modulated at frequencies of 20-300 Hz. The obvious first result was that one does not see anything above 100 Hz on the current measuremnent for the coils on Robicon supplies (50Hz for TMX supplies). With the chopper we were able to see current wiggles at 1000 Hz (with some imagination). But, more close analysis is needed.

Physics Operator Summaries

Entered: Jul 7 2004 04:36:50:613PM

Author: To Be Determined

Physics Operator Summary of runs 940426,27 Tinios (Session Leader) , Horne(Physics Ops) Contine checkout of power supplies Take 3 plc current limit shots -- OH1, EF3, EF4 to investigate whether bad flux loop F17 is still useable. (It's not.) Granetz plans to fudge processed dat for this loop. Hybrid algorithms need rethinking, but we believe we can begin to try to get a breakdown. We tested gas; need to fix wire 14 in pcs, hybrid trees to do gas right. Gerry's MP was run, mostly. We think he has enough data to begin with, and given how many shots he lost due to power supply problems, if it turns out that he needs a few more shots I believe he should get them. These two runs were a fairly thorough test of the new PCS including multiple segments; a couple of minor bugs were discovered and fixed but basically everything worked. By the end of 940427 all power supplies were working except TF and EF3. TF crowbarred often enough that we turned it off. Cooling time to recover from a TF crowbar is 30 min or so. EF3 has some wierd problem that the engineers are working on. Run 940426 -- For the plc limit shots to test f17 -- Currents -- ef3 -- 6000 A ef4 4000 A OH1 -- 10000 A 2 -- tf only, 100 ka -- PLC worked 3 -- OH1 -- 10000 A -- worked 4 ef3 6000 -- power supply problem -- crowbar made it to about 5000. 5 Try again -- worked 6 ef4 test Long huddle waiting for TF to cool. Began MP 42 7 Gerry's shot 1 -- hybrid program. -- many problems beginning with EF3. TF will requite 1 hour to cool down from this shot! 8 disable all supplies except ef3; TF off. Good shot. This had segment 1 only, which has pid=0 starting at t=0, and zero drawn V. Hence ic_ef3 droops from t=0 to t=1. 9 Turn on segment 1; enable ef2 supply. TF still off. ef3, ef2 still misbehave. 10 ef3 hybrid, ef2 plc. EF3 bad, ef2 worked. ef2 won't run under hybrid program -- bad relay? 11 turn off ef2, ef3; try the rest under hybrid control. TF on, flattop from -1. sec to t=0. for Artur. tried to work. Oscillation on oh2_U visible, but current drooped, possibly due to oh1 voltage not being correct. 12 load another shot with corrected oh1 voltage, ramps starting at -1 instead of -2, and double pid gains. no tf this shot -- still too warm. 13 Discovered that all controllers in segment 2 are zero, so the rampdown never happens. Otherwise, more or less as expected. 14 segment 1 only. PID is on for full segment. 15 repeat 16 data acq test. By the end of the run -- all problems except intermittent behavior of ef3 are understood. Plan for tomorrow -- Gerry will prepare .pros to do the following: load segment 1 with complete current controlled shot, including pre-programmed voltages from V = M dI/dt. Load segment 2 with his wiggles, plus resistive term taken from data today. 940427 -- Gerry tried to calculate and preload V with V = M di/dt. shot 1 repeat of yesterday 15 -- all ok from here, but some problem with EF3 blew a fuse at end of shot. Compared m_out to Gerry's calculated V; they look ok -- typically within 2 volts!. 2 (12:33!) load Gerry's oh2_u shot. (51) This will oscillate oh2_u at 20,50,500 hz. 500 hz is from a signal generator gated by the "utility" input, wire 3. During t=0 to t=1 sec, no feedback -- coils are on pure voltage control only. Result -- Good Shot -- flattops look nice! oscillations are fine. faults in oh1 and ef3 during rampdown. 3 repeat 2 with faster digitizing rate. Result -- Looks the same as 2. -- including faults. cycle time dominated by tf cooling, about 20 min. 4 load Gerry 53, oh2_l wiggles. change tf flat-top to run t=0. to 0.8 sec. from t=-.5 to 1.0. This should significantly reduce the tf heating, hopefully giving more shots. Result -- seemed to work, but ef3 tripped again. EF3 will be turned off (breakers open) but control power still on, hopefully preserving the fault state so Joe D can trace it down tonight. Gerry believes that EF3 isn't critical to his analysis, since it doesn't produce radial field. 5 Load Gerry shot 52 -- oh2U, frequencies 100,200,300(300 from signal generator). Result -- Basically ok. TF crowbar again. Offset on signal generator looks too high -- adjust for next shot. Gerry brought the oh2_u v in segment 2 to zero at the wrong time (.55 instead of 1.0) which made the shot ugly, but his data is ok. 6 Gerry is preparing his next shot, (54) with ugliness hopefully dialed down to an acceptable level. Sim to 5, oh2L wiggles. Result -- very nice -- flattop ok ==> voltage tweak to signal generator was right. 7 EFC next, Gerry shot 55 100,200,300(generator) Worked ok. 8 Gerry 56 20,50, generator at 500 hz tf crowbar again. Bad news. 500 hz actually visible on chopper current. 9 Gerry shot 155 (like 55) generator at 750 (Using scope timebase) Increase amplitude of wiggles by factor of 3. 10 same as above, generator at 1khz. 11 ef2u Gerry 57 20,50,100,200,333(generator) 12 ef2l, Gerry 58, wiggle amplitudes increase by factor of 2. 13 ef1 u , Gerry shot 59, 20,50,333 Peculiar problem -- no oscillator output from m_out. Tried dropping controller amplitude, though that shouldn't have done anything. 14 60 ef1_u gas spikes on also. Oscillator gets through. Need to investigate this. 100,200,333 15 ef1 L Gerry 61 ok 100,200,333 16 ef1 L gerry 62 20,50,333

Session Leader Comments
Apr 27 1994 11:44:01:560AM	940427001	Gerry Tinios	Repeat last shot from yesterday. EF2's worked well after we switched their start time to shortly after the time when the shot starts, so they don't start with 0 demand. EF3 worked well during the shot but gave a fault after the shot. EF4 switches on late.
Apr 27 1994 12:58:10:410PM	940427002	Gerry Tinios	OH2U is supposed to have oscillations (20,50,500 Hz) on this shot. From t=0 to t=1 no current feedback. TF flattop from -0.5 to 1.0 This is to be the first shot pertaining to MP#42 Not a bad shot
Apr 27 1994 01:07:56:510PM	940427003	Gerry Tinios	Repeat of last shot with adjusted hybrid and bus rogowski digitisation rates/start times.
Apr 27 1994 01:53:57:900PM	940427004	Gerry Tinios	Now OH2L has 20,50,500 Hz oscillations. EF3 had a fault.
Apr 27 1994 02:21:51:500PM	940427005	Gerry Tinios	Now OH2U has 100,200,300 Hz oscillations. EF3 is not being used from now on. It had to be left in the fault state so it can be inspected later.
Apr 27 1994 02:28:13:740PM	940427006	Gerry Tinios	Now OH2L has 100,200,300 Hz oscillations.
Apr 27 1994 02:42:14:960PM	940427007	Gerry Tinios	Now EFC has 100,200,300 Hz oscillations.
Apr 27 1994 02:46:29:730PM	940427008	Gerry Tinios	Now EFC has 20,50,500 Hz oscillations.
Apr 27 1994 03:19:02:280PM	940427009	Gerry Tinios	Now EFC has 100,200,750 Hz oscillations. Amplitude of oscillation was raised by a factor of 3. No TF. Frequent TF crowbarring causes heating and concern, so we decided to stop using TF for the day.
Apr 27 1994 03:25:10:820PM	940427010	Gerry Tinios	Now EFC has 100,200,1000 Hz oscillations.
Apr 27 1994 03:34:28:140PM	940427011	Gerry Tinios	Now EF2U has 20,50,100,200,333 Hz oscillations.
Apr 27 1994 03:51:35:490PM	940427012	Gerry Tinios	Now EF2L has 20,50,100,200,333 Hz oscillations. Oscillation amplitude is the double of that of last shot (240V).
Apr 27 1994 04:16:20:000PM	940427013	Gerry Tinios	Now EF1U has 20,50,333 Hz oscillations. Demand oscillation amplitude is 240V. The function generator part (333 Hz) did not work. Some hybrid problem with controller limits.
Apr 27 1994 04:36:09:760PM	940427014	Gerry Tinios	Now EF1U has 100,200,333 Hz oscillations. Demand oscillation amplitude is 240V (105V for 333 Hz, it's still a mystery why the controller won't let us get 240) . Gas valves hybrid controlled for this shot.
Apr 27 1994 04:44:37:930PM	940427015	Gerry Tinios	Now EF1L has 100,200,333 Hz oscillations. Demand oscillation amplitude is 240V (tried to get that for the third frequency as well by playing with the PID gain instead of the controller. It worked.).
Apr 27 1994 04:49:07:820PM	940427016	Gerry Tinios	Now EF1L has 20,50,333 Hz oscillations. Demand oscillation amplitude is 240V (For the 333Hz we expect to get demand voltage oscillation amplitude of 500V! Got it, too, but it still does not show up on the bus rogowski).

Physics Operator Comments
Apr 27 1994 10:42:33:950AM	940427001	Josh Stillerman	Everything ran. EF3 failed in invert. Blown fuse.

Engineering Operator Comments
Shot	Time	Type	Status	Comment
1	10:19:31:750AM	Test	Ok	ef3 forced shutdown at end of shot
2	12:38:32:690PM	Test	Ok	ef3, oh1 faults
3	12:58:41:380PM	Test	Ok	ef3 forced invert, oh1 comm failure in inversion - margin angle pot move
4	01:26:32:170PM	Test	Ok	faults ef3 and tf
5	01:56:26:170PM	Test	Ok	tf comm fault
6	02:22:25:360PM	Test	Ok	no faults
7	02:32:11:370PM	Test	Ok	no faults
8	02:45:23:430PM	Test	Ok	tf comm faults
9	03:12:12:660PM	Test	Ok	no faults
10	03:22:31:280PM	Test	Ok	no faults
11	03:33:56:070PM	Test	Ok	no faults
12	03:48:56:380PM	Test	Ok	no faults
13	04:01:54:210PM	Test	Ok	no faults
14	04:26:51:920PM	Test	Ok	no faults
15	04:36:46:010PM	Test	Ok	no faults
16	04:45:05:490PM	Test	Ok	no faults