Choose timezone
Your profile timezone:
MP3 meeting
Participants
Adriaan Rijllart
Amalia Ballarino
Andrea Musso
Andrzej SIEMKO
Antonio Perin
Christian Giloux
Christine Darve
Gianluigi Arduini
Giorgio D'Angelo
Howard Pfeffer
Hugues Thiesen
Jean-Philippe Tock
Jens Steckert
Jim Strait
Laurent Deniau
Mike Koratzinos
Nuria Catalan Lasheras
Richard Mompo
Rob van Weelderen
Sandor Feher
Sandrine Le Naour
Walter Venturini Delsolaro
Zinur Charifoulline
MP3 meeting 14 October 2009
The leak is at the level of 10-2 mbar l s-1 at 4.5 K.
Andrzej Siemko added that two turbomolecular pumps have been installed, incrementing the total pumping speed by 400 l s-1. This stabilized the pressure at levels of 10-4 mbar, which is sufficient for normal operation.
According to Andrzej, in the light of past experience, the leak might become much bigger when cooling down at 1.9 K.
Should this happen, he said the possibility of operating this cell at 4.5 K will have to be discussed, expecially concerning ELQA (the magnet performance at 4.5 K, instead, is not a problem)
The subsector with the leak is 27L4 (Going from Q27L4 ti Q31L4)
It was found as well that teh A29L4 magnet fiducial has moved.
A couple of RCD circuits in sector 45 had two retraining quenches
They were declared non conform according to the rule.
Gianluigi confirms that 400 A is a suitable new I_PNO for beam operation
However one of them trained below 400A
Walter reminded that it is very unlikely that the quenches are in the same magnet
Decision is to have an additional ramp, which will probably solve the problem as the observed current gain between training quenches is of the order of tens of A.
Issue with the splice mapping in sector 78 (one segment giving 11 nOhm): Howard Pfeffer clarified that the problem was traced to the measurement chain and is not in the splice itself.
The board will be changed and the measurement repeated
Howard confirmed that the noise floor is much lower at night, he advocated performing the splice mapping overnight.
A couple of RQT circuits have high PCS resistance values
taking into account the measurement uncertainty the values are still compatible with the threshold, i.e. they are above threshold by less than one sigma.
Walter suggested approving the steps as the apparent resistance is decreasing at high current, as the resistance is measured again at I_PNO, where the uncertainty will be lower
After the second displacement (lowering of the low point of the link), the ELQA will be redone.
Richard Mompo asked whether the busbars of the D2 magnet go through the same link
Antonio answered yes
Mike Koratzinos asked whether the theory of a "dry joint" could be checked by a boiloff whereby one should observe the gradual uncovering of the other joints with similar effects on the resistance
J. Strait cautioned that this experiment would not be conclusive as the problem is probably not due to a single free surface but rather to localised gas bubbles
Walter asked how the duct is centered and how rigid it is
Antonio showed a plastic spacer on the radiograph. The spacer is not perpendicular to the ducts and does not touch the bottom surface.
Walter asked what happens if the duct touches the envelope
Antonio said that it might give rise to an additional heat load
Amalia warned to check the geometry before reconnecting the warm cables.
She was concerned that the cables might pull and damage the current leads
Andrzej was worried about the bellows at the magnet side, which might be overloaded.
Jean Philippe said that the main load is resulting from torsion
Gianluigi recalled that a longitudinal displacement of 3 mm was observed at the latest survey
Andrzej said that all the displacements have to be recorded and taken into account before deciding a new correction
Gianluigi said that the diplacements of the magnet and of the DFB are all tracked in the database of D. Missiaen
Concerning the test voltage to apply when requalifying the system, there is not yet a figure agreed by all the actors,
and formidable confusion was manifest in the discussion that followed Richard presentation.
No attempt is made to summarize it here.
Nuria asked to have a table with all the test voltages and conditions applied to these objects.
A dedicated meeting between Nuria, Andrzej, Jim, Sandor, Christine and possibly others will tackle the issue and report back to MP3
He argued that high voltage tests after a repair should aim at qualifying the circuits for the needed performance, with reasonable embedded margin.
He showed statistics of the voltages of charged quench heaters, indicating an average of about 910 V with 10V spread among channels
The global maximum is only 4% above the average
He then proposed to apply 20% safety margin, coming up with the figure of 1100V
A proposal to overcome this was presented, consisting of removing the overlap between crates, which is, together with the applied logic, at the source of the problem.
An alternative solution was advanced by Howard Pfeffer (and independently by Nuria)
This would modify the decision logic, resorting to a "voting" mechanism
Changing the logic, however, would take several weeks, as it would require extensive firmware updating. This cannot be done now
The same can be said of the other possible solution of correcting the heater polarities that give problems
However, at the next IST, the heater discharges will be all measured and recorded to trace the problematic ones
Answering to a remark by Walter, Howard recalled that although the simmetric quench detection is probably not needed to protect the magnets below 6 kA, it is still necessary to provide redundancy of the UPS.
Matters arising
Jean Philippe Tock informed about the leak between cold mass and insulation vacuum which has appeared in sector 34 (B29L4).The leak is at the level of 10-2 mbar l s-1 at 4.5 K.
Andrzej Siemko added that two turbomolecular pumps have been installed, incrementing the total pumping speed by 400 l s-1. This stabilized the pressure at levels of 10-4 mbar, which is sufficient for normal operation.
According to Andrzej, in the light of past experience, the leak might become much bigger when cooling down at 1.9 K.
Should this happen, he said the possibility of operating this cell at 4.5 K will have to be discussed, expecially concerning ELQA (the magnet performance at 4.5 K, instead, is not a problem)
The subsector with the leak is 27L4 (Going from Q27L4 ti Q31L4)
It was found as well that teh A29L4 magnet fiducial has moved.
A couple of RCD circuits in sector 45 had two retraining quenches
They were declared non conform according to the rule.
Gianluigi confirms that 400 A is a suitable new I_PNO for beam operation
However one of them trained below 400A
Walter reminded that it is very unlikely that the quenches are in the same magnet
Decision is to have an additional ramp, which will probably solve the problem as the observed current gain between training quenches is of the order of tens of A.
Issue with the splice mapping in sector 78 (one segment giving 11 nOhm): Howard Pfeffer clarified that the problem was traced to the measurement chain and is not in the splice itself.
The board will be changed and the measurement repeated
Howard confirmed that the noise floor is much lower at night, he advocated performing the splice mapping overnight.
A couple of RQT circuits have high PCS resistance values
taking into account the measurement uncertainty the values are still compatible with the threshold, i.e. they are above threshold by less than one sigma.
Walter suggested approving the steps as the apparent resistance is decreasing at high current, as the resistance is measured again at I_PNO, where the uncertainty will be lower
Update on Q4.R2 (A. Perin)
Antonio described the findings and the correcting actions on the position of the DFBMB and of the link (see slides)After the second displacement (lowering of the low point of the link), the ELQA will be redone.
Richard Mompo asked whether the busbars of the D2 magnet go through the same link
Antonio answered yes
Mike Koratzinos asked whether the theory of a "dry joint" could be checked by a boiloff whereby one should observe the gradual uncovering of the other joints with similar effects on the resistance
J. Strait cautioned that this experiment would not be conclusive as the problem is probably not due to a single free surface but rather to localised gas bubbles
Walter asked how the duct is centered and how rigid it is
Antonio showed a plastic spacer on the radiograph. The spacer is not perpendicular to the ducts and does not touch the bottom surface.
Walter asked what happens if the duct touches the envelope
Antonio said that it might give rise to an additional heat load
Amalia warned to check the geometry before reconnecting the warm cables.
She was concerned that the cables might pull and damage the current leads
Andrzej was worried about the bellows at the magnet side, which might be overloaded.
Jean Philippe said that the main load is resulting from torsion
Gianluigi recalled that a longitudinal displacement of 3 mm was observed at the latest survey
Andrzej said that all the displacements have to be recorded and taken into account before deciding a new correction
Gianluigi said that the diplacements of the magnet and of the DFB are all tracked in the database of D. Missiaen
Consolidation of DFBX (R. Mompo)
Richard recalled the problem (naked conductors in the DFBX connectors) and the solution implemented.Concerning the test voltage to apply when requalifying the system, there is not yet a figure agreed by all the actors,
and formidable confusion was manifest in the discussion that followed Richard presentation.
No attempt is made to summarize it here.
Nuria asked to have a table with all the test voltages and conditions applied to these objects.
A dedicated meeting between Nuria, Andrzej, Jim, Sandor, Christine and possibly others will tackle the issue and report back to MP3
Inner triplet quadrupoles quench heater voltages (J. Strait)
Jim presented the rationale behind the proposed figure of 1100 V as test voltage for the QH circuits.He argued that high voltage tests after a repair should aim at qualifying the circuits for the needed performance, with reasonable embedded margin.
He showed statistics of the voltages of charged quench heaters, indicating an average of about 910 V with 10V spread among channels
The global maximum is only 4% above the average
He then proposed to apply 20% safety margin, coming up with the figure of 1100V
The nQPS Symmetric Quench Detectors; Recent measurements and proposed parameter settings.(Jens Steckert)
Jens illustrated the recently discovered "feature" of the SimQPS of quenching unwanted magnets as a result of the coupling between the quench heater circuits and the magnet coilsA proposal to overcome this was presented, consisting of removing the overlap between crates, which is, together with the applied logic, at the source of the problem.
An alternative solution was advanced by Howard Pfeffer (and independently by Nuria)
This would modify the decision logic, resorting to a "voting" mechanism
Changing the logic, however, would take several weeks, as it would require extensive firmware updating. This cannot be done now
The same can be said of the other possible solution of correcting the heater polarities that give problems
However, at the next IST, the heater discharges will be all measured and recorded to trace the problematic ones
Answering to a remark by Walter, Howard recalled that although the simmetric quench detection is probably not needed to protect the magnets below 6 kA, it is still necessary to provide redundancy of the UPS.
There are minutes attached to this event.
Show them.
The agenda of this meeting is empty