MP3 Meeting

Wednesday 30 Jun 2010, 10:00 → 12:30 Europe/Zurich

874/1-011 (CERN)

Agenda: - Presentation on QPS_OK Logic and Strategy (by R. Denz) - AOB

Slides QPS_OK.pdf QPS_OK.pptx

(Minutes taken by M. Zerlauth)

-) By default the QPS_OK cannot be masked, only it’s influence on the injection permit at the SIS level

-) In principle any SIS channel can be made UNMASKABLE, probably should be looked at as by now we always go to higher currents

-) Loss of QPS_OK is mostly LOCAL (the masking could be allowed), however losses of e.g. complete WorldFIP segment must be cured prior to re-injection

-) Losses of QPS_OK that need fixing before the next fill

                -) QHPS is discharging (<810V) for any kind of circuit; current fill can continue but needs to be fixed for next run > injection current

                -) Loss of Quench Loop controller of the main circuit (current fill can continue)

                -) Loss of EE systems of 600A and 13kA circuits (current fill can continue)

                -) Problems with Voltage Taps on current leads (will trigger after RESET), current fill can continue but needs immediate fix afterwards

 

-) Losses of QPS_OK that probably need immediate action even during the fill (to be decided)

                -) If there is >=1 QHPS of the SAME magnet lost, strongly recommended to perform Slow Power Abort and fix the problem

                -) Massive losses of communication with e.g. complete WorldFIP segments or multiple elements in the circuit  e.g. UPS malfunctioning (causing charging of QHPS)

 

-) Possible improvement needed

                -) Disable Masking of SIS channel

-) Capture automatically the loss of >=1 QHPS and perform a SlowPA (probably through the SIS)

-) Detect massive losses of QPS_OKs -> new policy to be established and endorsed by MP3 (programmed dump + Slow Power Abort, e.g. via a QPS_MAINTAIN)

 

AOB:

Bernhard:

QH discharges as they occurred yesterday on Q10 and Q9 should be added for completeness into the MP3 DB.

 

Reiner:

FastPA of the RB now looks like a SlowPA due to the delay of the EE switches by 350ms

LV tools having some issues with the RBAC login              

Q10 is most exposed circuit as the cabling (including the mid-point) is the longest one. Possibility to cure is to reduce the voltage range these detectors can read (currently +-200V are readable), could be changed to +-40 as for the IT and would increase precision of detection and decrease sensitivity.