HighField Magnet Workshop

1
Nb3Sn Magnet Instability Observations at
LBNLAlan Lietzke, Paul Bish, Bill Lau, Sara
Mattafirri, Mark Nyman Superconducting Magnet
TestingLawrence Berkeley National Laboratory
2
Introduction

• Magnet Instabilities
• Mechanical (wire moves in B-field)
• Diamagnetic (B-field moves thru wire)
  Flux-Jump
• Observations
• Summary of instability observations at LBNL
• Observational challenges
• Examples
• Summary Conclusions
• Future Plans

3
Experimental GoalsFast-Flux Investigations

• Initial (1996) Goals
• Identify quench trigger.
• Locate mechanical triggers.
• Differentiate between training and thermal
  quenches.
• Recent Focus
• Mechanical or flux-jump quench triggers

4
Summary Instability Observations at LBNL
5
Summary (boot-legged)Instability Observations
at LBNL
6
Observational Challenge PS Noise

• Voltage View
• PS ripple
• PS switching transients

• dV/dt view
• PS switching transients

PS transients and ripple
Flux-jump
7
Quench Trigger

• Mechanical
• or
• Diamagnetic?

8
Fast-Flux Differentiation

• Stick-slip
• Higher frequency.
• Oscillation.

• Flux-jump
• Lower freq.
• Little overshoot, no ringing.

300V/s, 30mV flux-jump
300V/s, 30 mV Stick-slip
9
Flux-Jump Variations

• Cascade (sometimes to quench)

• Single

10
D20 (1997) 1st Four-Layer Nb3Sn

• Geometry
• Dipole
• Cos-theta
• Four-layer Nb3Sn
• 50 mm clear bore
• L gt 1 meter.

11
D20 (1997) Training

• Quench
• 13.5T, but slow training.
• Stick-slip limited (??)

• Fast-Flux
• Stick-slip slow training
• Flux-jumps?? rare ( 8/ramp).

Stick-slip training
Flux-jumps
12
D20 (1997) Fast-Flux Signals

• Typical (100s)
• Stick-slip
• 40mV, 500V/s

• Rare (6-8/ramp)
• Flux-jump??
• 3mV (small), 25V/s

Flux-jump?? Low-B, no training, no decel., no
ring., but repeatable, and little net flux change.
Accel
Ring
Decel
13
D20 Fast Flux Change

• Summary
• Trained to its 4.3K Iss. (reluctantly!).
• Many fast-flux changes
• Mostly mechanical.
• Little memory.

• Conclusion
• Flux-jumps NOT a problem (at this Jc, Istrand,
  RRR).
• Slow stick-slip training slow quench
  training.

14
SM-02 (Feb02) 1st Magnet Failure

• 14 low Cu SC 7 hi-purity Cu
• RRR 250 (7-Cu strands)
• 42 Iss (Istrand 285A).
• Splice quenching (??, hi-RRR sig.).
• No ramp-rate hump.

• Fast-Flux data (finally) examined Aug03
• Flux-jump quench onsets.
• Cascade, mirror symmetric.

15
SM-02 (Feb02) Violent Quench-Start

• Large inductive voltages
• 10x D20
• Swamped PS transients.
• Swamped resistive growth
• 20m/s 4V/s
• Erratic signals
• No ringing!

16
RD3c 11T Racetrack Common-Coil Dipole

• Cost-effective field-quality demo
• Reuse 14.5 Tesla POP outer coils.
• Reuse 19 Tesla support structure
• Iron yoke
• Bladder Key
• Al shell

17
RD3c Magnetic Design

• Challenges
• Correct large positive sextupole (RD3b outer
  coils).
• Large (a2) quadrupole (reused yoke).
• Large SC-hysteresis (Nb3Sn)
• Smallest measurable bore (max. force on
  correction coil).

18
RD3c Training Performance

• Slow training
• Erratic plateau.
• 0.9 Iss.
• Stick-slip.
• Many (repeating) fast-flux events.
• Study platform for FFC events.

19
RD3c Stick-Slip Quench Triggers

• Accel/decel, ringing
• gt2000V/s
• Freq1/size
• lt 1 trigger quenches.

20
RD3c Fast Flux Changes

• Two Kinds
• Flux-Jumps
• Slow (10ms)
• I lt 50 of Iss
• Polarity dB/dt.
• Every ramp.
• Stick-Slip
• Fast (0.1ms)
• Training
• Yes lt 8KA.
• No gt 8KA.

21
RD3c Multi-Ramp FFCs

• Flux Jump
• Same every ramp reversed polarity during
  down-ramp.
• Stick-Slip up-ramp only forgets every ramp
  gt 8kA.

22
RD3c Fast-Flux Summary

• Quench-training stalled 10.7 kA (92 Iss).
• Huge (gt2KV/s) stick-slip triggers.
• Stick-slip training stalled
• No memory above 8 kA (20 lower).
• Flux-jumps
• Record number did not limit performance.
• All from higher Deff coil.
• independent of ramp repetition and direction.
• Fast-flux-change polarity reverses with dI/dt.

23
NMR (2003)

• Magnet Performance
• 7T (60 of Iss)
• No training effect.
• No ramp-rate hump.

• Conductor/splice performance
• RRR 5.6
• 0.2 lt Rsplice lt 0.3 nOhm.

24
NMR
Quench start
Splice quench.

• Quench origins
• Outside of splice.
• All but 1st in SC-12.
• Propagation
• 50 into nearby splice in 1 ms.
• Max. splice temperature after quench 40K.

• Conclusions
• Splices low heating, well cooled, but quenching
  nearby.

25
UND (2003)1st NbSn Undulator

• 6-0.7mm, high-I strand (1kA)
• RRR low (??)
• 40 Iss (Istrand 333 A).
• Splice quenching every quench
• After quench-start (no damage)

• Fast-Flux
• Many flux-jumps
• All quenches had F-J start.

26
UND (2003) Quench Augmentation
F-Js during quench augmentquench rate.

• Cascade during quenching
• Increases effective quench speed.
• Reduces peak temperature.

Small flux-jumps start quench.
27
HD1 Coil Mod-B Cross-Section
150MPa Pre-stress Target

28
HD1 Quench Training

• 1st Quench
• 8.7 kA
• 13.3 T
• 78 of Iss.
• Layer-1
• Return-end spacer
• Cond. Limits
• 11.3-11.5 kA
• 16.6-16.9 T

29
HD1 Typical Quench Trigger (15.8T)

• Fast motion
• 2 kV/s (huge)
• Delay
• 0.6 ms
• decaying tremors
• Resistive growth
• large surges (multi-turn, or F-J?)

Flux-jumps (?) during quench augment quenching
rate.
30
HD1 Typical Quench Trigger (15.8T)

• Fast motion
• 18 mV
• Magnetization decay
• 0.6 ms
• 23 mV
• Rapid resistance growth
• 400V/s

31
HD1 Fast-Flux Summary

• Quench training stalled 16 Tesla (93
  Iss).
• Stick-slip training stalled?? (not examined
  yet).
• Flux-jumps
• 100s, most ever seen in a large magnet.
• Did not limit magnet!

32
SM06 (Apr04)

• Set-up
• Oxford conductor
• Small common-coil
• SC13 SC14
• Reacted with NMR 11 12.
• Lightly loaded.

33
SM06 (Apr04)

• Training
• All in SC13 (0.68 Iss).
• All but 1st in outer layer.

• Observations
• SC13 flux-jump trigger
• 1/10 had splice quenching.

34
SM06 T03s Quench Initiation

• Massive flux-jump trigger
• 4kV/s.
• 3V, 750us.
• Massive parallel propagation
• 2300V/s.
• 10 turns.

35
SM06 T03s Flux-Jump Trigger

• Quench-trigger
• 4 V/ms F-J
• Reverse polarity.
• Precursor
• Oscillatory
• Reverse polarity.

36
SM06 T03s Flux-Jump Precursor

• Oscillatory
• Reverse polarity.

37
SM06 (Apr 04) Fast-Flux Comparison

• Experimental Set-up
• Quenched-start.
• Ramp
• Up-Down-Up.
• Trigger
• Vn-m (5mV)
• Observed
• Mostly flux-jumps
• 7/400 stick-slip
• 1st up-ramp.
• Max. freq. _at_
• 2.5 kA
• 25 of Iss.

38
SM06 (Apr 04) Fast-Flux Comparison

• Experimental Set-up
• Quenched-start.
• Ramp
• Up-Down-Up.
• Trigger
• Vn-m (5mV)
• Observed
• Mostly flux-jumps
• 7/400 stick-slip
• 1st up-ramp.
• Max. freq. _at_
• 2.5 kA
• 25 of Iss.

39
SM06 (Apr 04) Fast-Flux Comparison

• Experimental Set-up
• Quenched-start.
• Ramp
• Up-Down-Up.
• Trigger
• Vn-m (5mV)
• Observed
• Mostly flux-jumps
• 7/400 stick-slip
• 1st up-ramp.
• Max. freq. _at_
• 2.5 kA
• 25 of Iss.

40
SM06 (Apr 04) Fast-Flux Comparison

• Set-up
• Trigger 5mV
• Counter reset
• dI/dt reversal.
• Faster 2nd ramp
• 2x faster
• Statistics
• UpRamp-1 440 cnts
• Coil-N
• DnRamp-1 215 cnts
• Coil-M (49)
• UpRamp-2 360 cnts
• dI/dt 2x
• Coil-N (82)

41
SM06 Summary

• Higher bandwidth amplifiers
• dV/dt extended from 10 kHz to 80 kHz.
• Still have trigger problems.
• Coil M limited performance
• Zero training 68 Iss.
• Lower RRR (5.2 vs. 7.1)
• Ramp-rate variations
• M flux-jumped less than N
• Fewer flux-jumps at higher speed.
• Larger flux-jumps at higher speed??

42
Summary of Observations

• Incremental DAQ Progress
• Noise suppression.
• Higher bandwidth viewing.
• Triggering improvements.
• Differentiating mechanical from flux-jumps.
• Flux jumps have become a problem recently
• 2002 SM02, mixed (Cu/Sc) strand (high Istrand).
• 2003 NMR, SM06 Low RRR.
• 2003 UND, UNDA High Istrand, Low RRR.

43
Conclusions

• Flux-jumps are only a problem if they avalanche
  adjacent elements
• Observation and localization continues to be a
  challenge
• Need a way to manage them (if they cant be
  eliminated.

44
Future Plans

• Better FFC-DAQ trigger
• Quieter PS environment.
• Better localization
• More channels,
• FFC antennas (coil arrays).