The ISIS Target Failure

1
The ISIS Target Failure

• Paul Hodgson
• The University of Sheffield

2
The Failure

• Target failed 21st December during a normal MICE
  shift
• This initially showed as an inability to lower
  the target beyond a certain point
• After repeated attempts to lower it the target
  eventually ceased to operate completely
• Raising and lowering the target frame had no
  effect

3
Target Schematics
Collar added here
4
Target removal and inspection

• Target removed from synchrotron 10th Jan
• Hard to see shaft tip because inside bellows but
  some evidence of melting
• Target transported to R52 for detailed inspection

5
The tip of the target was melted.
Estimate from photograph that target is 8.0 0.5
mm shorter
Target tip shorter but material concentrated in a
blob at tip no missing material
The Shaft was bent some 10º off axis
6
The Top Bearing

• No dust on the upper surface of the top bearing.
• No wear on the work surfaces of the top bearing
• No wear on the upper shaft area

7
The Bottom Bearing

• A small amount of wear on bottom part of shaft
• Matching wear pattern on the bottom bearings
• Collar fell out when bottom bearing removed

8
Summary of Observations

• No dust on upper shaft and bearings
• Some odd asymmetric wear on lower shaft/bearing
  not much wear though
• Circular bearing stop collar was loose
• Lower section of shaft was bent approx. 10º off
  axis
• Bottom 8.0 0.5 mm of target tip melted
• Abnormal shape of pulses during failure

9
Initial Comments and Questions

• DLC coating failure was NOT the cause of failure
• The jamming occurred due to the circular collar
  coming loose and sliding on the lower shaft
• The link between the melting and failure is not
  clear
• What caused the bending ? no record of
  mechanical damage during removal and inspection
• What caused the melting of the target tip ?
  titanium melts at 1960K
• Normal operation should not let this happen 3
  full pulses before ISIS trip T1100K

10
(No Transcript)
11
11
12
1429
12
13

• Start-up after incident. Same depth and delay
  produces little beamloss. Target steered deeper
  but later into beam.
• BL 200 50 mV
• T 10.5 0.5 ms
• D 29.0 0.3 mm
• Delay and depth tweaked to increase beam-loss to
  desired value (300 mV)
• BL 360 50 mV
• T 7.0 0.5 ms
• D 30.9 0.3 mm
• The target is now going into the beam early and
  deep should be producing large beamlosses
  (1000 mV)

• Start of shift .Target started deep and was
  raised out of beam
• BL 450 50 mV
• T 8.0 0.5 ms
• D 38.0 0.3 mm
• Main run until incident. Target was again raised
  out of beam until desired beamloss achieved. The
  delay was altered later in period by 1 ms.
  Presumably the beam was wandering a little. Last
  recorded pulse at 1429.
• BL 300 50 mV
• T early 8.0 0.5 ms
• T late 9.0 0.5 ms
• D 39.1 0.3 mm

14
14
15
15
16
16
17

• Plots from the next MICE shift 19-12-08
• Despite target pulsing at maximum reach only
  modest beamloss achieved
• The effect of the incident permanent

17
18
29th Nov Incident

• At 1435 something happened to the target
• Last normal recorded pulse is 1429
• ISIS log entry at 1435
• Subsequent pulses needed to have the target 8 mm
  lower to produce the same beamloss
• This is clear evidence of the time at which the
  target was melted

19
Conclusions

• The target tip was melted during an incorrect
  procedure on 29th November 2008 (now not possible
  to do this)
• The target was temporarily parked in the beam and
  the trip was inhibited in some way needs 12
  full pulses to melt
• The target did not jam due to DLC coating failure
  no dust/wear on upper bearing (remember the
  previous R76 failure)
• The target jammed because the circular collar
  came loose and slid up and down on the shaft,
  intermittently jamming
• Not clear what caused the wear on the lower shaft
  could be from the detached collar or caused by
  the bend ?
• Unclear if melting caused collar to detach but
  possible
• Still unclear what caused the bending. Could the
  heating have also done this ?