Summary notes of the forty-eighth meeting of
the LHC Commissioning Working Group
Tuesday June 17th, 14:00
CCC conference room 874/1-011
Minutes of the Previous Meeting and Matters
Arising
The minutes of the 47th
meeting had not yet been distributed at the time of the 48th meeting
(due to the change from a 2-week to 1-week cycle and the 23rd
LHCMAC).
Roman Pot Operation at
the LHC (Ernst Radermacher)
Ernst Radermacher’s
presentation was structured as follows: the Roman Pots, their position during
data taking, control and interlocks, commissioning, and summary.
The TOTEM experiment is installed at
IP5, near CMS. It consists of two inelastic telescopes and two sets of Roman-Pot
module-pairs on each of the outgoing beam lines. The Roman Pots are located at
147 m and 220 m distance from the collision point. A single Roman-Pot module
contains 3 pot units – one horizontal and two vertical ones, interconnected by
simple bellows. Each module also features an integrated beam position monitor.
The horizontal and vertical units of a module are separated by 4 m. The Roman Pots
are equipped with individual stepper motors. They were designed by TS. All
units have been installed and baked out. A thin window of 150 micron sits
between the silicon sensor of the detector and the beam vacuum. Ferrites in the
vicinity minimize possible interference with the beam and protect the detector.
The inside of the pot is in vacuum. The detector can be installed inside the
pot after removing a flange.
Ernst Radermacher
pointed out that there is a 200 micron distance between the detector and the
window. The window flatness is better than 50 micron. The window will be
displaced by 50 micron in case of a vacuum loss at 100 mbar.
The Roman-Pot motor control is based
on the Central Collimation Application (CCA). The interface with the low-level
control is taken care of by the Collimator Supervisory System (CCS). The Roman
Pots are controlled as much as possible in the same way as the collimators. The
LVDTs (linear variable differential transformers) are read at 100 Hz, and the
philosophy is “safety first”. The procedure is to agree on the ultimate position
limit, and to store this limit in the CCA database. The limit must then be
corrected for the actual BPM reading. The LVDT values are compared with the
limit. If the limit is exceeded, the beam is dumped.
Bernhard asked whether the Roman
Pots would be set only as a function of nominal position or also as a function
of measured data rates. Ernst replied that he would come to this topic later.
Bernhard next inquired why the beam was dumped. An emergency exit procedure
could be an alternative for the dump. Stefano responded that one must avoid the
type of problem encountered at the Tevatron, where a
Roman Pot moving into the beam had caused dramatic damage. Jorg
clarified that the beam was only dumped when the limit was exceeded.
Ernst presented two running
scenarios. Before injection the Roman Pots are in their retracted positions,
i.e. at the end switches, and the LVDT values at zero. After injection, one
would wait for stable beams, and then approach the beam to the agreed position.
In case of UNSTABLE_BEAMS or IMMINENT_BEAM_ABORT the Roman Pots would be
retracted. A flow diagram (slide 14) illustrates the position interlock logic
(CIBU 1 - Control Interlock Beam User Interfaces
1), which involves the end switches and the LVDT resolvers, the latter being
read out at 100 Hz. The central (pink) part of the interlock, related to LVDTs
and limits from FESA, depends on the motor rack power. TOTEM would like to have
this part supported by UPS’s (uninterruptible
power supplies). The injection inhibit (CIBF – Controls Interlocks Beam Fibre-link) is set according to the states of all end
switches. At low intensity the retraction for the “unstable beam” mode can be
avoided by a manual override from the DCS (Detector Control System). This
possibility is provided since the collimation team would like to operate the
Roman Pots also with unstable beam for joint collimation and Roman-Pot
calibration studies. An alternative for the manual override would be the “safe
beam flag”, if the safe beam intensity were sufficient
for the collimation tests (this does not seem to be the case however).
Ernst now outlined the commissioning
of the Roman pots: Before the start of beam operation, the stepping motors are
commissioned, LVDTs calibrated, and the C3F8 cooling over long distances put
into operation. One assembly will be installed in a horizontal pot at 220 m in
Sector 4-5 before the end of June and in Sector 5-6 during July. Two vertical
assemblies at 220 m in Sectors 4-5 and 5-6 will be installed during a technical
stop later this year. Interlocks will also be tested.
After the start of a beam operation,
in a typical TOTEM run, under “stable beam” conditions the pots are moved
together with the CCC. In particular the BPMs are read and the beam position is
verified together with the CCC. The safe limits are corrected with the actual
BPM values. The pots are then moved towards the beam down to a distance of 35
mm. Next the horizontal pots are moved nearer to the beam to see the
“diffractive peak”. And also the vertical pots are inserted closer to the beam
such that the rates are equal in both pots. The final distance could be 10-15 sigma. BLM and RP signals are compared. If the situation is
good for data taking the MOVEMENT_INHIBIT will be set and data taking can
begin.
The early physics includes the
measurement of elastic and diffractive protons. The elastic protons are seen in
the vertical pots. The diffractive peak is observed only in the horizontal
plane. In 3 hours of running at 1e32
cm-2s-1, one expects to see 1e6 elastic events and 6e7 centrally diffracted
(DPE) events within the Roman Pot acceptance.
Ernst summarized that TOTEM will
install two horizontal Roman pots before LHC closure, and two vertical ones in
a technical stop. TOTEM is copying as much as possible the control of the LHC
collimators, which requires a close collaboration between TOTEM and LHC. In
2008 TOTEM plans to take data for diffractive physics. A common effort from LHC
and TOTEM will deliver a safe position for the best physics results.
Stefano remarked that a collaboration
with IT was underway. Helmut asked when the TOTEM Roman Pots would be in OUT
position, and whether they would not be IN during squeeze. Ralph wanted to move
both collimators and Roman Pots for certain studies. Therefore there are two
beam modes where Roman Pots can be IN. Helmut also asked for the beta function
desired in 2008. There was no request for any extreme TOTEM-specific beta
values. Ernst also added that the aim was to complete the 220 m station in the
first shut down, as well as half of the pots at the 147 m position (to learn
about the background situation there).
Jorg commented that the logic for movement of the Roman Pots had
been agreed upon quite a while ago. Roger asked whether it was safe to set the
limits according to the BPM readings. Jorg answered
that this might need some experience. The same issue existed for the
collimators. We will need to learn to safely control these devices. Bernhard
asked whether TOTEM and LHC believed in BPM measurements. Ernst commented that
when approaching the beam one could also observe the data rates, though beam
halo could complicate the interpretation. Stefano briefly explained the
motivation behind the request of the collimation team. Concerning the earlier
question of Bernhard regarding the alternative exit procedure, Jorg pointed out that there were dump limits and also alarm
limits, and that an alarm limit could be used to trigger an emergency retraction.
Stephane asked whether TOTEM wanted to move the Roman Pots before or
after establishing collisions. The 35 mm distance was possible only with stable
beams. Mario Deile replied that TOTEM also wants to
measure in single beam operation. Stephane commented
that the orbit at the Roman Pots will change before and after bringing the
beams into collision.
Ralph S. asked how TOTEM would verify a 20 sigma distance
between pot and beam. A distance in mm is measured while moving in. How would
one know the beam size at the Roman Pot. Ernst replied that the theoretical value was
used. Ralph S. asked whether a dynamic beta beat possibly as large as 20% was
included in the distance number. Ernst answered that the physics was
independent of sigma. Ralph S. remarked that the physics would depend on the
alignment of Roman Pot with respect to the beam. Ernst explained that the rates
on both vertical sides would have to be equal, and that the beta beating was
not important for the data taking. Bernhard commented that once a Roman Pot had
become a secondary collimator at HERA. The distance was crucial. Ernst
commented that the Roman Pots would go as close as possible. Mario clarified
that the setting limits were obtained from the collimation team, quoted in
units of mm or sigma. Frank asked whether the limits from the collimation team
included dynamic optics errors. Stefano replied no. Jorg
pointed out that one may get some clues from the cleaning efficiency.
Ralph S. asked how one would distinguish physics from beam scraping.
Mario explained that for the physics trigger, there must be a coincidence in
both arms.
Massimiliano asked whether the empty or the equipped Roman pots would be
first commissioned with beam. Ernst responded that only the pots with detectors
would be commissioned. Jorg commented that otherwise
one really might now know where the beam is. Massimiliano
asked whether the nearby BLM was part of the pot safety. Ernst replied that
only the LVDT reading was used for the interlock. He added that the beam would
move only a few microns during a fill, and the motion was measured directly.
Mario pointed out that TOTEM did not know which system would
first see a problem, the BLM or the detector. All signals will be measured. The
pot would be moved in very slowly.
Massimiliano asked whether one could use the current on the detector
sensor as an additional hardwired safety interlock. Ernst and Mario replied
that currently there was no detector-based interlock, and that such interlock
would not be implemented for the first year. One reason is that one does not
know which settings this interlock should correspond to. Mario added that the
readout chip provides the trigger information.
Massimiliano asked whether one could try to find a more robust detection
signal. Mario replied that there are other ideas, e.g. more BLMs, but all left
for later phases. Ernst remarked that
the Roman pots set at 10 sigma should not be a
problem. Stephane commented that the safest distance
is 15-16 sigma in the straight sections. Jorg clarified that
the 10 sigma value had been specified for TOTEM dedicated runs. Stefano
commented that the standard BLMs should protect the Roman pots. Daniela pointed
out that simulations had been done to optimize the TOTEM dedicated BLMs. The
results had been given to Bernd, who had asked for an analysis of the signals
that would be observed if a Roman pot was moved too close to the beam. A new
fellow would complete this task. Stefano clarified that the TOTEM BLMs were
dedicated ones, but of the standard type.
Massimiliano now came back to the installation schedule. On each side one
horizontal pot would be ready. Access was needed for installing the vertical
pots. Such information should be passed to him. He asked how much access time
would be needed exactly. Ernst replied one day. Roger and Massimiliano
commented that probably there would be many accesses in 2008 albeit no
scheduled stops.
Gianluigi asked whether the two vertical Roman Pots were located in
the secondary vacuum. Ernst answered, yes, all the Roman Pots were. The window
would blow up at 50 bar pressure only. Pots with detectors are equipped with a
pump providing a pressure of a few mbar. One could not go to lower pressure due
to glow discharge at the detector.
Sector 56
Commissioning (Roger, Gianluigi, Mike)
Roger started with a brief overview
of the status of the hardware commissioning. The Sector 81 powering part of HWC
had started one or two weeks prior to the meeting. Sector 78 HWC was about 90%
complete, the triplet still outstanding. The dipole 30 right of point 7 had
shown some unusual behavior, and had subsequently become the subject of many
investigations. According to the latest findings this magnet – an early series Alstom magnet - might be OK, however. Possibly there was a
need to increase the pressure in the QPS system from 100 mbar to 250 mbar to
cope with the diagnosed problem. Sector 56 was commissioned up to about 5.5 TeV. During an incidence in the previous week one of the
dipoles quenched in a symmetric fashion, which was not detected by QPS,
indicating a hole in the QPS for particular quenches. The training campaign was
stopped following this incidence, and the sector was being handed over to
operation. In preparation for the operation commissioning, discussions had been
held in the control room and a meeting with a few people had also taken place
at the end of the preceding week.
Gianluigi now discussed the preliminary plans for tests that had been
worked out by the machine check out team consisting of Mike, Reyes, Rossano and himself. He first reviewed the status of Sector
56. The same afternoon a final test of the overheated bend was being performed
at 2 kA. If this last test was successful, operations would take over the
sector. The power converters in the arc and in UA63 would be unblocked either
this very afternoon or the following morning.
A small number of converters were not available, primarily
the triplet correctors. A few converters were still under discussion: the 3
circuits at the bottom of slide 3 which should become available. Gianluigi pointed out that there was still quite some
activity in this sector, concerning e.g. optical fibres,
alignment of the dump-line vacuum chambers, concrete shielding of Q5, bake out
of CMS chamber, activities for TOTEM and ZDC, TIM
installation (modification of access doors for camera system), MSD (septa) gaussmeter, installation of protections for power
converters, and BLM tests.
Daniela commented that the shielding was not for the Q5
magnet itself, but for the protection of nearby power converters.
An important issue to be addressed in
the tests is the 600 A converters, which were commissioned with ramp rates and
acceleration rates lower than had been requested (approximately 40 circuits were
concerned). Gianluigi informed the working group that
the QPS software had been upgraded so that in the future one should be able to
increase the acceleration rate. He clarified, however, that it was Sector 78 which
was being commissioned with the new QPS software (more filtering; not yet the very
latest request), and not Sector 56, so far. Stephane remarked
that, more precisely, the 0.02 units of acceleration had been changed to 0.1
units in Sector 78. But the number asked for was 1 (or 0.3), and this had not
yet been accommodated anywhere. He confirmed that the limited acceleration rate
was a concern for about 40 magnets.
Gianluigi made the proposal to leave the mornings to complete the
works already ongoing in the tunnel, and to use the afternoons to go ahead with
the operations tests. This time sharing was to be reviewed at the end of the
following week, when, as he hoped, tests could continue through the entire day.
One would then try to collect and bundle the remaining interventions.
An important object of the tests was
the generation and preparation, which included sorting out of the optics,
injection tests including FIDEL, 5 TeV ramp, 5 Tev squeeze, programmed ramp down, pre-cycle, context-independent
parameters, knobs & bumps (e.g. separation bumps, crossing angle bumps, lumi scan, Q, Q’ coupling, coupling compensation,
harmonics), and rf test devices.
Roger asked whether all information
available from FIDEL was taken into account. The answer was yes and this
question smoothly led to Gianluigi’s next slide (no.
9), according to which the transfer functions for all magnet circuits, the harmonic
errors for quads and bends, decay correction and snapback corrections were included.
Slide 10 contained more details on the cycles to be used for the tests.
Roger posed the question whether
this was the first time that that these transfer functions from FIDEL were
tried. Mike replied that a number of
models had been used before, but that this was the first real release.
Trim tests addressed the full
parameter space. CTRIM (cubic trim) – the main trim drive module – was going to
be tested thoroughly. Trims would be tested in various possible situations: at
injection, on the ramp & squeeze functions (including checks of the limits),
flat top, and during squeeze after squeeze. Circuit acceleration limits would
be tested at the same time. Slide 12 presented a list of available trim knobs.
Other items to be tested included
the SIS, MPS, BI devices, logging, alarms, fixed displays, trigger acquisition,
sequencer, run control, timing, experiment instrumentation & signals, experiment
communication – DIP (the list was to be extended).
Gianluigi next presented the detailed test programme
for the days from 17 June through 26 June. On Wednesday and Thursday (17-18.06.)
one would look at the 600 A acceleration rates and ramp rates [Reyes, Rossano, Gianluigi], with support
from MPP (which might be needed with low priority), and at LSA trim tests. Friday
19 June would be devoted to the generation of ramp and squeeze with the achieved
parameters [Mike and Reyes], as well as to alarm and SIS verification
(parasitically) [Gianluigi, Luis, Markus Rossano]. On Monday, 23 June, Mike, Massimo, Ralph, and Stephane would test the pre-cycle, injection decay
corrections, snapback correction, ramp, bumps and knobs etc. Wednesday 25 June
was designated to RB-QF-QD tracking (1/2 day), and to SSI tests for power
converters. On Thursday 26 June the RB performance w/o active filter would be
studied.
Gianluigi announced that regular
daily meetings would be held in the glass box at 17:00 starting the
following day. These meetings would
allow the plan for the next day to be communicated in the HWC meeting the
following morning. Gianluigi mentioned that Mirko Pojer was the HWC contact
person for the planned tests.
In week 27 cryogenics tests were
scheduled. No powering was possible in Sector 56 during these tests, which
included the cool-down and tests of the Sector 56
quench lines, cold compressor recovery tests, and a large quench recovery test.
Some pending interventions could be done during this week, provided they were
compatible with the cryo activities.
Stefano asked whether there would be restrictions on the access, and
implications for patrols. Gianluigi answered that the
plan was to keep the sector closed during the test, if the patrol was not lost.
Otherwise one would need to re-patrol. The smooth transition in this first week
might require some compromise.
Brennan and Verena
commented that for the next couple of weeks, one would better find a systematic
way to minimize the interference. Clearly there was a need to learn.
Walter came back to the topic of the
PO tracking test. Gianluigi explained that there
would be two tests. The first was a ramp, feeding the nominal current in QF and
QD, and in RB (to get a reference). Later one would wish to correlate ramps in
different sectors. The test would become trickier if several sectors were
concerned.
Roger asked in which form we would
record what we do. Gianluigi replied that for the
time being spread sheets were being used. Later one could add copies to the LHC
OP logbooks. The easiest and most flexible approach should be followed. Jorg asked whether this recording activity should perhaps
be coupled with a cleanup of the database, e.g. removing old MAD and layout
versions, including BPMs which did not exist anymore. Stefano remarked that
there was the same type of problem for collimators. LHC layout version 503 was
under approval. Massimo commented that the sequence was ready, but the aperture
model was taking time. The optics was not different from that of the previous
version. The only difference was some hardware changes. Jorg
elaborated that the real layout of the machine differed from the MAD version
requiring artificial bridges. Database and actual machine should be made
identical. Massimo agreed, citing the example that the undulator
position was still not correct. Massimo and Mike almost agreed to use version
500 for a start, and that the open issue concerned the hardware. Mike said he
would download the latest sequence that had been declared ready. However, it
was then remarked that the ultimate aim was to run with version 6.503, and that
obsolete or outdated items from version 500 had not yet been cleaned up. Jean-Jacques
added that even in the last version released wrong BPMs still persisted. Jorg agreed, and he highlighted that one list from orbit
feedback and one from the database showed about 100 mismatches, partly due to the
use of old names. Mike suggested that this could be the check no. 1. Jorg said he would require the correct BPM list in the database
with an associated optics sequence. Massimo said he would discuss these
problems with Sami, emphasizing that the layout version 6.503 would be the
version used for the LHC commissioning. He reiterated that the aperture model
part concerning the vacuum chambers was not finished.
Jean-Jacques stated that BI was interested
in the daily glass-box meeting at 17:00, and could join it to discuss their
plans. Gianluigi highly welcomed this initiative and he
extended a warm invitation to Jean-Jacques and his team, recalling that the
first meeting would be held on the following Wednesday and that some HWC
persons would also participate.
Injection of first beam (Mike)
Sector 56 (Gianluigi)