What caused the problem? What are the symptoms ? Solution (if any) …?
-
Upload
alana-cervantes -
Category
Documents
-
view
37 -
download
1
description
Transcript of What caused the problem? What are the symptoms ? Solution (if any) …?
MP3 Mini WorkshopM. Zerlauth March 8th 2011
1v0
RSS.A78B1 quench-like events during PGC- Complementing Arjans slides for RCO.A45B1
• What caused the problem?• What are the symptoms ?• Solution (if any) …?• Potential existence of same problem in similar
circuits?
CERN
[email protected] MP3 Mini-Workshop – 8th of March 2011
Quench-like trips of RSS.A78B1/2…● RSS.A78B1/2 tripped twice during pre-cycling/heat runs, indicating an increase of current
just before tripping– RSS.A78B2 on 16th of February @ 04:16:56 at 400A during heat run/PGC– RSS.A78B1 on 18th of February @ 02:19:26 at 400A during heat run/PGC– RSS.A78B1 on 19th of February @ 01:07:48 at ~335A during heat run/PGC
● At first sight -> Looking like (and being mistakenly classified as) a quench, if not…
2
Looking familiar?
CERN
[email protected] MP3 Mini-Workshop – 8th of March 2011
Cross-talk @ 400A● During heat runs and PGCs, the normal cross-talk can be observed (different to
RCO.A45B1..)
CERN
[email protected] MP3 Mini-Workshop – 8th of March 2011
How to diagnose…● Although not always as clearly visible, V_MEAS is a better signal to
observe, suggesting sudden changes of impedance ● With the loosening of the connection, also the I_ERR_MA significantly
increases, mostly well before the circuit trip
1 min
CERN
[email protected] MP3 Mini-Workshop – 8th of March 2011
…and the guilty one is…
● Following the repetitive trips, EPC proceeded to exchange the power module and found…
● …a loose bolt on the DCCT head, right next to the thyristor…
● Note: EPC verified that even without the tightening bolt the circuit can NEVER completely open
CERN
[email protected] MP3 Mini-Workshop – 8th of March 2011
How to detect other case?
● Are similar cases possible in the machine -> Certainly YES
● Can we detect them? YES, but needs special runs (few hours for the whole machine)…
– The slower the loop is, the worse the rejection of perturbations will be
– By putting all loops slower than usual (10 or 100 times), and running the circuit in DC state (80% of nominal current) one should be able to detect bad connection through verification of the I.ERROR (to be confirmed for circuits with >> time constants).
-3.6-3.55-3.5-3.45-3.4-3.35-3.3-3.25-3.2-3.15-3.1
-100.005-100.004-100.003-100.002-100.001
-100-99.999-99.998-99.997-99.996-99.995
V.meas[V]
I.meas[A]
RCO.A45B2 I.MEAS
RCO.A45B2 V.MEAS
-4
-3
-2
-1
0
1
2
3
4
-120-100-80-60-40-200
20406080100120
V.meas[V]
I.meas[A]
RCO.A45B2 I.MEASRCO.A45B2 V.MEAS
0.1V