Transcript From MARIC minutes of 22.11.2006

600A circuits: splices qualification
W. Venturini Delsolaro for MPP
Hardware Commissioning “Debriefing” Workshop
11 July 2007
Outline
•
•
•
•
•
Original motivation
Procedure: PCS1, 2, 3, and 4
Procedure changes on the fly
Findings and results
What we have learned
From MARIC minutes of 22.11.2006:
PROBLEM OF BAD SPLICES IN LINE N of SEC. 7-8
Tests at Cryo-Lab R ~ 20 nΩ instead of expected 3 nΩ
Calculations by M. Calvi + EEWG + MPP 
integration of new steps in the HCP (EDMS 519716) for
7-8 only, initially
PCS steps
•
•
•
•
“20 mV or as low as possible…”
PCS1 tuning below 20 A:
PCS2 programmed trip at 50 A
PCS3 programmed trip at 100 A
PCS4 splice verification at ± 200 A
Indeed most powering details were defined during
commissioning to cope with the 20mV threshold
above 20 A. In some cases thresholds had to be
raised to 40 mV (lattice sextupoles), for Q6 we
used 50 mV
RSD1.A78.B1; cycle ± 200 A

V V
Rs 
2 I

Noise in level in RSD.A7.B2 circuit
URES=RS · I
(resistive voltage computed by QPS)
RSF1.A78.B1 from 0 to 200 A
RCS.A78.B2 from 200A to 0A
•
•
•
•
Resolution ~1μΩ, up to a few μΩ with noise. We can only measure the total R
If consider 20 nΩ as “bad”, single bad splices are not detectable (this we knew)
Tolerance from cryogenics  7 nΩ (LHC design report, vol. 1, p. 315)
So, if we are strict, we cannot even say that cryogenics tolerances are met
CIRCUIT
MAGNET
TYPE
Number of
magnets
number of
connections
Sum of Connections
Resistance if R=Rmax
measured (20nOhm)
[μOhm]
Sum of Connections
Resistance if R=R
estimated in average in
7-8 (8nOhm) [μOhm]
RQTxx
MQT
8
90
1.80
0.72
RQTxx
MQT
1
14
0.28
0.11
RQSxx
MQS
2
42
0.84
0.34
RSxx
MS
9
44
0.88
0.35
RSxx
MS
10
46
0.92
0.37
RSxx
MS
11
48
0.96
0.38
RSxx
MS
12
50
1.00
0.40
RQTLxx
MQTLI
1
10
0.20
0.08
RQTxx
MQT
1
14
0.28
0.11
RQSxx
MQS
2
42
0.84
0.34
RQSxx
MQS
4
80
1.60
0.64
RSSxx
MSS
4
68
1.36
0.54
RQTLxx
MQTLI
1
10
0.20
0.08
RQTLxx
MQTLI
2
6
0.12
0.05
RQ6
MQTLH
6
4
0.08
0.03
ROxx
MO
8
68
1.36
0.54
ROxx
MO
13
80
1.60
0.64
HWC
RESULTS
< 1 μΩ
in almost all
cases, 2.5 μΩ
in RQS.L8.B1
Conclusions (1)
• We had to learn how to power the 600A circuits circumventing
the pitfalls of non linear inductances and parallel resistors
(PCS1-3). This took the best part of the time (because of the low
threshold). PCS4 was the easiest step
• We verified the absence of “very bad” (>>1μΩ) splices for the
“priority 1” short list of circuits
• Tolerances from cryogenics (almost) OK
• These tolerances are distinct from quench limits (local)
• Time was very little: analysis was done “on the fly”, TIMBER
data could not be used (filtering), as originally foreseen
• For the future, access to “good” data would be an asset (METER
on request for specified signals and time windows? Alternatively
PM can be used, but it is more time consuming)
• In other sectors the quality assurance of splices was reinforced,
the likelihood of bad splices is much less (we could take the risk)
Conclusions (2)
• Goal for 7-8 was to avoid damage due to over voltages
• There is still concern on the consequences of the
thermal cycle, the same procedure (or a very similar)
will be applied in 7-8, during 2nd run
• After successful PCS4 the threshold can be brought to
nominal
• Concerning the “forbidden zone” between 20A and 200A
we have no more reasons today than we had before to
ignore it, apart from the lower likelihood of problems in
other sectors
• The same applies to the 20 mV threshold
• MPP will issue a new standard procedure for Sect. 4-5