EU contract number RII3-CT-2003-506395 CARE-Note-2006-014-HHH
High Energy
High Intensity
Hadron Beams
A Possible Upgrade of the LHC Injection Lines to 900 GeV using HERA Dipoles
Karl Hubert Meß, David Smekens
CERN, Geneva, Switzerland
Abstract
The injection lines TI2 and TI8 between the SPS and the LHC are designed for 450 GeV.
Upgrade scenarios with higher injection energy will require new injection beam lines. Either
completely new tunnels have to be created with freedom to choose the magnet system or,
more likely, superconducting magnets will be needed in the existing tunnels.
In this report, we examine the possibility of using HERA magnets, which may become
available with the shutdown of HERA, as transfer line magnets.
-1-
EU contract number RII3-CT-2003-506395 CARE-Note-2006-014-HHH
Introduction
One of the possible scenarios to improve the future performance of the LHC foresees an
increase of the injection energy of the LHC. This will improve the field quality in the
LHC magnets at injection and hence increase the dynamic aperture. At the same time, a
larger beam emittance could be tolerated, leading to greater tolerance of beam losses or
higher luminosity.
Obviously, such a scheme also requires a higher-energy injector. In principle, this
injector could be placed in the LHC tunnel or the high-energy proton beam could be
produced already in the SPS tunnel. In this case, new transfer line magnets in TI2 and TI8
would be required, because the bending strength of the present injection line magnets is
optimally matched and limited to the present SPS energy of 450 GeV.
It happens, that there exist two superconducting accelerators, the Tevatron at Fermilab
and HERA at DESY, with magnets of appropriate field strength, length, sagitta and
aperture to be used in this new injection line. Fortuitously, both accelerators are about to
be decommissioned in a few years. This note studies the implications of the installation
of magnets from HERA, which has seen less radiation than the Tevatron. For simplicity,
as a first approximation, it is attempted to keep the beam-line lattice as it is and simply to
replace the various normal conducting magnets by a superconducting one of twice the
strength.
Because the quench protection schemes for the Tevatron and HERA are incompatible, the
magnets of the two accelerators are not mixed in this study. Maybe a mixture of magnets
could lead to a better solution. The injection lines contain also C-magnets, septa and
kicker. These magnets cannot be exchanged with any existing superconducting magnets;
hence, R&D will be required to develop such devices. There are three QT magnets in
TT40. We propose to keep them, because they are presently excited at only half their
nominal strength, or very nearly so.
Present TI8/TI2 Layout
The layouts of the TI8 and TI2 Transfer Lines are described in the LHC Project Note 128
[1] and in the LHC Design Report, Vol III [2]. TI2 has a length of about 2.9 km. It
consists of one 48° horizontal bend and three vertical bends of 61, 42 and 9 mrad to avoid
the underground valley below St. Genis. The steepest slope is 2.6%. TI8 is somewhat
shorter (2.7 km) but steeper (3.77%). It consists of a horizontal bend of 103° in the
descending part, preceded and followed by vertical bends of 38 and 35 mrad respectively.
The lines use a FODO structure with a half-cell length of 30.3 m and 4 dipoles per half-
cell for the horizontal bending part. The vertical bends are made of a different type of
bending magnets. The main features of the injection lines are shown in Figures 1 to 4,
taken from the LHC project note 128. Note that the proton beam is bent counter-
clockwise in TI8 and clockwise in TI2. Note further that in both cases the magnets are
placed at the inner radius of the injection tunnel.
Table 1 gives a list of the bending magnets with their respective deflection angle.
-2-
EU contract number RII3-CT-2003-506395 CARE-Note-2006-014-HHH
Figure 1. The injection lines TI2 and TI8
-3-
EU contract number RII3-CT-2003-506395 CARE-Note-2006-014-HHH
Figure 2. The vertical deflections in TI2
-4-
EU contract number RII3-CT-2003-506395 CARE-Note-2006-014-HHH
Figure 3. The vertical deflections in TI8
-5-
EU contract number RII3-CT-2003-506395 CARE-Note-2006-014-HHH
Figure 4. The TI FODO cell structure
-6-
EU contract number RII3-CT-2003-506395 CARE-Note-2006-014-HHH
Table 1 List of bending magnet families in the injection channels
Deflection Total
Dipole group
Magnet type Number per magnet deflection
(optics name)
[mrad] [mrad]
TI 2
MBI MBI 112 7.4716 836.82
BH1 MBB 2 6.0000 12.00
BH2 B280 6 2.9993 18.00
BH3B MSIB 3 2.7620 8.29
BH3A MSIA 2 1.8570 3.71
BV1 B340 17 3.5947 61.11
BV2 B340 12 3.5216 42.26
BV3 B340 4 2.2619 9.05
BV4 MKI 5 0.1700 0.8
TT 40
BH1 BHC 3 5.0000 15.00
BH2 B340 4 3.7000 14.80
TI 8
MBI MBI 228 7.6110 1735.31
MBIT*) MBI 8 9.6742 77.39
BH3 B190 1 0.2590 0.26
BH4 B340 7 3.5000 24.50
BH5B MSIB 3 2.7620 8.29
BH5A MSIA 2 1.8570 3.71
BV1 B340 12 3.1634 37.96
BV2 B280 5 3.0038 15.02
BV3 MKI 5 0.1700 0.85
-7-
EU contract number RII3-CT-2003-506395 CARE-Note-2006-014-HHH
HERA Magnets
A HERA half-cell consists of one dipole on either side of the dipole-corrector and
quadrupole assembly. The FODO cell has a length of 47.012 m. The dipoles contain
beam-pipe corrector windings, as indicated in Figure 5. A dipole corrector and a beam
position monitor are also integrated in the cryostat of the quadrupole. A few shorter
quadrupoles and vertical dipoles exist to adjust the optics and to deflect the proton beam
vertically. The key parameters of the various magnets can be found in Table 2 [3], [4].
In HERA the superconducting main magnets are connected in series. The current flows
clockwise through the dipoles and counter-clockwise back through the quadrupoles.
Hence the optical lattice is fixed. Adjustments to the tune are made by varying the
relatively strong quadrupole correctors, wound around the beam pipe inside the dipoles.
All dipoles are curved to follow the local bending radius of the beam of r = 588 m. The
proton beam travels counter-clockwise in HERA. The magnets are placed on the outer
side of the tunnel with the quench relief valves also pointing to the outside. (Looking
with the beam: towards the right hand side.)
Table 2 HERA magnet types
Dipole Dipole Quadrupole Dipole Sextupole s Quadrupole Decapole Dodecapole
BL, BR BV Q d q 10 12
horizontal vertical Part of Q BL, BR BL, BR beam BL beam Q beam pipe
assembly beam pipe pipe pipe
Nominal 5027 A 5027 A 5027 A 35 A 65 A 85 A
Current
Nominal 4.649 T 4.649 T 90.18 T/m 1.17 T 46.4 T/m2 1.8 T/m 2*10-3 B* 44*10-4 B*
Field/
Gradient
Nominal 0.68 Tm 0.17 Tm @ 0.24 Tm 0.2 Tm
integrated @ 25 mm 25 mm @ 25 mm @ 25 mm
Field
Magnetic 8.886 m 3.4 m 1.874 m, 0.61m 5.9 m 5.830 m 3.0 m 2.0 m
Length 1.7m, 1.5m
Overall 9.766 m
length
Overall 0.61 m 0.61 m 62.06 mm 64.86 mm
diameter
Coil 75 mm 75 mm 75 mm 75 mm
inner
diameter
Beam pipe 55.3 mm 55.3 mm 55.3 mm 55.3 mm
inner
diameter
Loss @ 6.7W 9W N. a. N. a.
4.2K
Loss @ 27.5W 28W N. a. N. a.
60K
-8-
EU contract number RII3-CT-2003-506395 CARE-Note-2006-014-HHH
47.012 m
D QF D D QD D
deca dodeca deca dodeca
23.506 m
qf, sf df qf, sf qd, sd dd qd, sd
Figure 5. The HERA FODO cell contains 4 dipoles (D) and two quadrupoles (QF,QD).
All dipoles contain a quadrupol and a sextupol corrector (qf, qd, sf, sd). Every second
dipole has also a decapole (10-pole) corrector. Quadrupoles are accompanied by dipole
correctors (d) and contain a dodecapole (12-pole) corrector.
A HERA dipole causes an deflection at 820 GeV of 2.9599 mrad at the nominal
excitation with 5027 A (4.649 T). The beam pipe is bent correspondingly. Note however
that HERA has been operating for a number of years at 920 GeV with a field of 5.216 T.
This was made possible by lowering the temperature of the coils.
Fitting the HERA magnets into TI2 and TI8
Tunnel space
The HERA tunnel is much wider than the injection channels TI2 and TI8. This is easily
visible in figure 6, which shows the HERA tunnel cross-section overlaid in purple with
the cross-section of the injection channels. Clearly that it will be difficult to
accommodate HERA magnets in the LHC injection channels.
The situation is particularly difficult in TI2, which is also used to transport LHC dipoles
into the LHC tunnel. Figure 7 shows this situation as it is now on the left and a possible
solution with the HERA magnets at the right. Note that the beam is presently on the inner
curvature of the tunnel in both cases. In TI2 the bending is also to the right, i.e. a
clockwise bending. The HERA magnets would need to run with inverted polarity, which
requires a change of the polarity of the protection diode. This can be achieved by
removing the diode stacks, opening them, inverting the polarity, testing them under
cryogenic conditions and reassembling. This is a tedious, but possible, operation.
Alternatively, adapter pieces could be envisaged, which change the polarity inside the
cryostat. This seems possible, because both magnet ends (end covers) will have to be
opened in order to fulfil the conditions set by the cryogenics (see below). In addition, the
HERA dipoles have their quench relief valves and the quench exhaust pipe at the outer
curvature (i.e. in the transport space in Fig 7), which would clearly obstruct the transport
zone. The magnet line can not be moved easily to the outer curvature, because the
position of PMI2 was chosen to lower LHC magnets into the space at the outer curvature
of TI2. One could presumably install a transfer table at the lower end of the shaft, such
-9-
EU contract number RII3-CT-2003-506395 CARE-Note-2006-014-HHH
that the TI2 magnets are in fact installed underneath the shaft at the outside curvature.
Components for the LHC could then be lowered to the transfer table and moved sideways
and lowered into TI2, to pass on the inner curvature. In this way one could avoid
dismounting the vacuum pipe, which presently blocks the transport path. In summary,
the HERA magnets must be reworked to fit into TI2. Note that fig. 7 does not show any
cryogenic line. The number of cable trays, however, can not be reduced drastically (see
below). It seems extremely difficult to fit the HERA magnets into TI2, as is illustrated by
figure 8. The 500 mm quench relief pipe does not fit. The quench relief valves would
need to be seriously reworked and still the cables would not find space. Even worse, as
can be seen in figure 2, the deepest point of TI2 is underneath the creek Lion. According
to the studies on cryogenics for the injection lines (see below), the single-phase helium
will have to enter the string of magnets at the lowest point. However, the study does not
take the actual configuration into account and will have to be repeated.
Figure 6. The cross section of the HERA tunnel. Overlaid is the cross section of
TI2 or TI8 (red)
The situation in TI8 is slightly better. The bend is counter-clockwise, as in HERA. Hence
the magnets can run with the original polarity. The magnets are however also here on the
wrong side of the tunnel. Again they would have to be moved to the outer curvature to
give space to the quench relief valves and quench pipe, which looks impossible, as can be
seen in Fig. 9. Certainly the exhaust valves (“Kautzky valves”) need to be reworked.
- 10 -
EU contract number RII3-CT-2003-506395 CARE-Note-2006-014-HHH
However, the cable ladders still do not find space and water cooling have also to move.
While the cables are still needed, the water cooling could maybe be reduced. Presumably
the beam line could be lowered somewhat, which seems possible at this stage. It is
unclear, if and where the cryogenic recoolers could be placed, unless they can be part of
the connection cryostats or the magnets themselves.
Structure
It might be possible to rearrange the optics to make optimal use of the properties of the
HERA magnets and achieve slightly higher energies. In this study we restrict ourselves to
a one-to-one adaptation. HERA magnets are mapped onto the existing structure. The new
energy is assumed to be 900 GeV. Because of the higher bending power of the HERA
dipoles the present cell length of 60.6 m is sufficient. The space between the dipoles will
be filled with connection cryostats, containing the quadrupoles, the current leads and
cryogenic feed-boxes.
Figure 7. The TI2 and TI8 cross-sections in the present state (left) and with
HERA dipoles (right)
- 11 -
EU contract number RII3-CT-2003-506395 CARE-Note-2006-014-HHH
Figure 8. HERA dipoles in TI2 including two versions of the quench relief pipe.
The 500 mm version does not fit. There is also no space for cable trays.
Figure 9. The HERA dipole in TI8 on the tunnel outer curvature (looking with the
beam). Although there is slightly more space, the 500 mm quench relief pipe and
cable trays do not fit.
Note that the limitation to 900 GeV is given by two constraints: the optics chosen as
baseline and the bending radius of the magnets. Both constraints are somewhat flexible.
The density of dipoles could be increased and hence the total bending power. However,
- 12 -
EU contract number RII3-CT-2003-506395 CARE-Note-2006-014-HHH
in this case, the aperture would be reduced, due to the poorer matching sagitta of the
beam pipe. This seems possible comparing the HERA aperture with the present beam line
aperture. In conclusion, a 1 TeV beam line of sufficient aperture could presumably be
made with a new optics design.
The proposed structure has, however, a very serious problem. In HERA the dipoles and
quadrupoles are connected in series, containing only one bus-bar pair in the bypass. As
can be seen in Tables 3 and 4 below, this is not compatible with the existing optics in TI2
and TI8 and the corrector quadrupoles inside the dipoles (2* 10.62 T integrated gradient)
are insufficient to replace real lattice quadrupoles. They could, however be used for
adjusting the optics.
Quadrupoles
HERA has three types of superconducting quadrupoles with a length of 1.9 m, 1.7 m,
and 1.5 m respectively, and a nominal gradient of 90.2 T/m. All these quadrupoles are in
series with the main dipoles in HERA. In the LHC injection lines the quadrupoles are
excited quite differently. This poses a serious problem for the interconnection of the long
regular sections. The bus-bars can not simply be connected through, as in HERA. In fact
the helium pressure vessel has to be opened and the internal cabling of the bus-bars has to
be changed and additional current leads have to be introduced, to connect the quadrupoles
in the proper families. Furthermore, 6 kA cables would have to be laid warm or as
superconducting link in a separate cryostat, external to the dipoles, unless one is willing
to make major modifications to all dipoles. It is not clear, whether such a modification
can be accomplished, since the HERA the dipoles have only one pair of main bus-bars
passing at the outside of the flux return iron.
In summary, the use the HERA quadrupoles would require massive rework on the
magnets, many new high current power converters and new cables or superconducting
links.
It seems more favourable to build instead new quadrupoles with about the strength of the
LHC-MQTL [5]. The MQTL are with a length of 1.3 m even shorter than the HERA
quadrupoles, and have a nominal current of 400 A at 4.5 K (90 T/m) or 550 A at 1.9 K
(129 T/m), which fits well to the existing cables and possibly to the power converters. As
will be discussed below, the operating temperature of the dipoles must be below 4.0 K.
Hence a nominal current (linear interpolation) of 430 A can be anticipated for MQTL
type quadrupoles at 4.0 K. The magnets, only one aperture of course, could easily fit into
the “empty” cryostats, to be built anyway. Note that circuits QTLF 4004, MQID 8030
and MQID 8730 (see Table 3.) will need 2 MQTL type magnets in series. There will be
space enough in the (not so “empty”) connection cryostats. Even the HERA quadrupole
is not strong enough to replace the QTL in the QTLF 4004 circuit. These magnets are
marked in dark grey in Table 3, which shows the list of all quadrupoles needed. The
majority of the magnets would have to be operated up to 470 A (12% above nominal).
This might be possible. Alternatively two magnets in series with the MQTL type of
- 13 -
EU contract number RII3-CT-2003-506395 CARE-Note-2006-014-HHH
winding with magnetic length of 1 m (each) could be envisaged. Maybe the easiest and
safest solution is to use the corrector quadrupoles in the adjacent dipoles, which have
each 6.3% of the kick strength of a MQTL (thin lens approximation), to support the
lattice quadrupoles. Additional 85 A current leads and power converters would be
needed. Solutions exist for these items in the LHC.
Bending magnets
The following Table 4 lists all families of existing bending magnets and their angle of
deflection. The next columns show the number and types of HERA magnets as well as
their excitation for 900 GeV and an indication of the aperture decrease, caused by the
mismatch between bending radius and the built-in sagitta of the beam pipes.
In total 196 dipoles of the BL or BR type are needed. Eleven of them will have to be
turned by 90 degrees around the longitudinal axis. This involves de-cryostating, turning
the support fixtures and cryostating. Six magnets will have to be tilted. Again the fixture
will have to be adjusted. The magnets will, as a rule, be installed at a much steeper angle
with respect to the horizontal than in HERA. It might be necessary to make use of the
longitudinal transport restraint, build into the magnets.The connection to the quadrupoles
could be done with additional “empty” cryostats. They can at the same time serve as
cryogenic feed-boxes with additional current leads.
Correctors
The HERA dipoles contain quadrupole and sextupole corrector windings. Their
integrated strength is 10.6 T and 273.8 T/m respectively. The corrector quadrupoles
might be useful in the long arcs, where the main quadrupoles are in series.
The HERA quadrupole cryostats contain superferric orbit correctors with an integrated
field of up to 0.68 Tm. This corresponds to more than twice the required value (0.243
Tm). The correctors could be built into the cryostats for the MQTL type quadrupoles. The
current is however more than 20 times the current used for the MCIA. New power
converters and cable would be needed. However, the magnets would most likely be
damaged in the attempt to get the correctors out of the cryostat. Alternatively, one could
design a corrector of 0.25 Tm with a current of 3.5 A, which would match the existing
power converters, provided an energy extraction system can be added.
- 14 -
EU contract number RII3-CT-2003-506395 CARE-Note-2006-014-HHH
Table 3 List of all quadrupoles in TI2 and TI8 (including TT 40)
Quadrupoles in TI2
Converter Magnet Number Inom [A] I max [A] int. Field at 900 Current in Current in
in Series T GeV HERA Q MQTL
MQIF 2580M MQI 26.0 490.0 600.0 69.4 138.8 4070.0 460.6
MQIF 2840M MQI 13.0 490.0 600.0 69.4 138.8 4070.0 460.6
MQID 2850M MQI 40.0 490.0 600.0 69.4 138.8 4070.0 460.6
MQID 2010 MQI 1.0 370.0 500.0 52.4 104.8 3073.2 347.8
MQIF 2020 MQI 1.0 290.0 400.0 41.1 82.1 2408.8 272.6
MQID 2030 MQI 1.0 270.0 400.0 38.2 76.5 2242.6 253.8
MQIF 2040 MQI 1.0 330.0 500.0 46.7 93.4 2741.0 310.2
MQID 2050 MQI 1.0 460.0 600.0 65.1 130.3 3820.8 432.4
MQIF 2060 MQI 1.0 470.0 600.0 66.5 133.1 3903.9 441.8
MQIF 2860 MQI 1.0 500.0 600.0 70.8 141.6 4153.0 470.0
MQID 2870 MQI 1.0 510.0 600.0 72.2 144.4 4236.1 479.4
MQIF 2880 MQI 1.0 410.0 600.0 58.0 116.1 3405.5 385.4
MQID 2890 MQI 1.0 460.0 600.0 65.1 130.3 3820.8 432.4
MQIF 2900 MQI 1.0 390.0 600.0 55.2 110.4 3239.4 366.6
MQID 2910 MQI 1.0 220.0 400.0 31.1 62.3 1827.3 206.8
MQIF 2920 MQI 1.0 320.0 500.0 45.3 90.6 2657.9 300.8
MQIF 2930 MQI 1.0 380.0 500.0 53.8 107.6 3156.3 357.2
MQIF 2940 MQI 1.0 150.0 300.0 21.2 42.5 1245.9 141.0
MQID 2950 MQI 1.0 280.0 500.0 39.6 79.3 2325.7 263.2
TT 40
QTMD 4001 MQI 1.0 470.0 500.0 66.5 133.1 3903.9 441.8
QTRF 4002 QTR 1.0 100.0 400.0 40.6 81.3 2384.4 269.8
QTRD 4003 QTR 1.0 70.0 300.0 28.5 56.9 1669.1 188.9
QTLF 4004 QTL 1.0 250.0 400.0 101.6 203.2 5961.1 674.6
TI 8
MQIF 8700M MQI 34.0 500.0 600.0 70.8 141.6 4153.0 470.0
MQID 8710M MQI 34.0 500.0 600.0 70.8 141.6 4153.0 470.0
MQID 8010 MQI 1.0 400.0 500.0 56.6 113.3 3322.4 376.0
MQIF 8020 MQI 1.0 410.0 500.0 58.0 116.1 3405.5 385.4
MQID 8030 MQI 1.0 600.0 600.0 85.0 169.9 4983.6 563.9
MQIF 8720 MQI 1.0 500.0 600.0 70.8 141.6 4153.0 470.0
MQID 8730 MQI 1.0 600.0 600.0 85.0 169.9 4983.6 563.9
MQIF 8740M MQI 2.0 300.0 400.0 42.5 85.0 2491.8 282.0
MQID 8750 MQI 1.0 490.0 600.0 69.4 138.8 4070.0 460.6
MQIF 8760 MQI 1.0 500.0 600.0 70.8 141.6 4153.0 470.0
MQID 8770 MQI 1.0 360.0 500.0 51.0 101.9 2990.2 338.4
MQIF 8780 MQI 1.0 320.0 500.0 45.3 90.6 2657.9 300.8
MQID 8790 MQI 1.0 290.0 400.0 41.1 82.1 2408.8 272.6
MQIF 8800 MQI 1.0 360.0 500.0 51.0 101.9 2990.2 338.4
MQID 8810 MQI 1.0 290.0 400.0 41.1 82.1 2408.8 272.6
- 15 -
EU contract number RII3-CT-2003-506395 CARE-Note-2006-014-HHH
Table 4 List of all dipoles
Deflection Total Change in
Dipole Current
Magnet per magnet deflection HERA Magnet Sagitta Field Radius
group Number # A
type [mrad] [mrad] type mm T m
TI 2
MBI MBI 112 7.4716 836.82 BL, BR 56 5458.99 -0.25 5.05 594.65
BH1 MBB 2 6 12 BL v BR 1 4383.79 -4.20 4.05 740.50
BL, BR 2 3287.84 -8.23 3.04 987.33
BH2 B280 6 2.9993 18
BL+Bvturned 1 4755.94 -2.83 4.40 682.56
BH3B MSIB 3 2.762 8.29 NA
BH3A MSIA 2 1.857 3.71 NA
BV1 B340 17 3.5947 61.11 BL, BR turned 5 4464.89 -3.90 4.13 727.05
BV2 B340 12 3.5216 42.26 BL, BR turned 3 5146.08 -1.40 4.76 630.81
BV3 B340 4 2.2619 9.05 Bv 2 4320.30 -0.74 4.00 751.38
BV4 MKI 5 0.17 0.8 NA
TT 40
BH1 BHC 3 5 15 BL v BR 1 5479.74 -0.18 5.07 592.40
BH2 B340 4 3.7 14.8 BL v BR 1 5406.67 -0.44 5.00 600.41
TI 8
MBI MBI 228 7.611 1735.31 BL, BR 114 5560.84 0.12 5.14 583.76
MBIT MBI 8 9.6742 77.39 BL, BR tilt 6 4711.96 -3.00 4.36 688.93
BH3 B190 1 0.259 0.26 Dipole corrector 1 tbd 0.02 1.30 2307.69
BH4 B340 7 3.5 24.5 BL, BR 2 4475.12 -3.86 4.14 725.39
BH5B MSIB 3 2.762 8.29 NA
BH5A MSIA 2 1.857 3.71 NA
BV1 B340 12 3.1634 37.96 BL, BR turned 3 4622.46 -3.32 4.27 702.27
BV2 B280 5 3.0038 15.02 BL v BR 1 5487.04 -0.15 5.07 591.61
BV3 MKI 5 0.17 0.85 NA
- 16 -
EU contract number RII3-CT-2003-506395 CARE-Note-2006-014-HHH
Cryogenics
At present HERA is running under the following cryogenic conditions [6],[7]:
The magnets are cooled with supercritical helium with p> 2.5 bar, T=4.0K (!). The
supercritical helium is cooled by the counterflow of two-phase helium. The expansion is
done at the lowest point of an octant, which is in the middle or at one end of an octant.
There is a “DFB” containing the current leads and all the valves including a Joule
Thomson valve. The two-phase flow is always directed uphill to avoid the capture of
bubbles. To run at 920 Gev at HERA (as would also be required in the LHC injection
lines) the temperature must be as low as T=4.0K. This is achieved by lowering the
suction pressure of the screw compressors to 650 mbar. The pressure drop over the 620 m
of one octant is about 100 mbar. The inclination of the HERA tunnel (max 10 mrad) is so
small that the resulting pressure drop due to gravity can be neglected. The stationary
mass flow is 35 g/s. A study has to be made of how to achieve similar conditions in the
steep LHC transfer line tunnels.
In 1993 N. Delruelle et al. [8] studied a possible cryogenic system for the injection lines.
At that time the ideas about the injection lines had not yet converged to the present
design. Hence not all conclusions in this study can be applied to the present case.
However, the slopes were planned to be even higher. The authors assume HERA or UNK
like magnets of only 5.7 m length at 5.4 T. The preferred solution foresees single phase
helium with re-coolers. The helium is fed in at the bottom of the arc (of which 3 were
planned at that time) and proceeds through the magnets at a rate of 60 g/s. The liquid is
re-cooled at the end of each cell by a heat exchanger in a bath of boiling helium. The gas
is returned through the magnets using the holes in the iron laminations of the magnets,
which in the case of the HERA magnets is either used for the heat exchanger or blocked
(lower orifice). Thermally insulated pipes have to be inserted, to prevent heat propagation
between the cells. Note that HERA quadrupoles do not have these heat exchanger holes.
In addition a 500 mm quench relief line is needed.
Alternatively a two-phase cooling scheme with phase separators at the end of the cells
has been considered. This scheme offers many advantages. However the authors request
further tests before the solution can be seriously pursued, because “its feasibility is still
doubtful”.
The study does not include the very special actual geometry of TI2 with its up-hill and
down-hill slopes. The narrow tunnel will not be able to accommodate a refrigerator. A
study has to be made on the basis of the actual geometries, whether and how stable
conditions can be achieved at 4.0K.
Protection Issues
The HERA dipoles come with a quench protection system [9], which is based on
magnetic amplifiers. The magnetic amplifier acts as a discriminator in a bridge,
consisting of the two half-coils of the magnet and two adjustable resistors. As a backup,
four dipoles are combined in a group and treated with a similar bridge. Higher resistor
- 17 -
EU contract number RII3-CT-2003-506395 CARE-Note-2006-014-HHH
values insure a higher threshold for the so called “group detection”. The electronics
triggers the firing of two heater power supplies per magnet and heating up of two heater
strips. This system works reliably; however it is not compatible with any CERN control
system. Moreover, few experts remain at DESY, who could transfer the system to CERN.
The CERN quench protection for LHC is, with slight adjustments of the capacitance and
voltage of the heater power supplies, adequate for the HERA dipoles.
The magnets are protected against the energy of the other magnets in the string by cold
diodes. The diodes and the heat sinks are constructed to survive a decay time of 20 s from
6 kA. As the maximum voltage during the extraction has to be limited to below ±530 V,
leading to an extraction resistance of less than 175 mΩ, the maximum inductance per
protection block is limited to 3.3 H or 55 magnets of 60 mH each. This is close to the 56
or two times 57 magnets needed in the long arcs, but a bit too low. The resistors in
HERA are simple bifilar stainless steel pipes, which could be reused adding some
electrical protection. The switches are laterally of the size of the magnets and should fit.
The same holds for the electronics.
Note that a number of dipoles will need its own power converter. In these cases the diode
might create a problem for getting the energy out fast.
The quench protection and energy extraction for the quadrupoles depend on the choice of
the quadrupole system. The series connection of 40 quadrupoles of the MQTL type has
an inductance of 5 H. Bypass resistors, as in the LHC implemented for this magnet type,
will be necessary. The resulting time lag, while ramping the magnets current, is not
important for the application as injection line magnets.
Summary
The special geometry of the TI2 and TI8 transfer lines poses serious problems for
upgrading them into the 1 TeV range. The HERA dipoles, with the required cryogenic
pipe and cable trays, will not fit in, unless heavily reworked. A major rework of the
magnets is also necessary to accommodate to the different cryogenic conditions (and the
varying field direction in TI2). The magnets will have to be taken out of their cryostat,
the end-covers will have to be removed, the heat-exchanger pipes will have to be
replaced and new connections will be necessary. Eventually only the collared coils with
their flux iron can be reused. In any case, the end covers need to be closed again, after
rerouting the pipe for the exhaust valve. Finally, new cryostats will have to be
constructed.
The HERA quadrupoles can in all probability not be used, unless the optics is completely
changed and the cryostats, the cold bus-bars and the internal helium pipes are redesigned.
As a result around 180 new quadrupoles will have to be made.
The cryogenics has to be extraordinary slim in order to fit into the tunnel. The steep slope
puts additional constraints. In particular TI2 with positive and negative slopes presents
problems. This should be addressed in a separate study again.
- 18 -
EU contract number RII3-CT-2003-506395 CARE-Note-2006-014-HHH
Overall, the difficulties are considerable and must not be underestimated. Different
optical solutions, including superconducting combined function magnets [10], should be
investigated, before seriously trying to integrate the HERA magnets and their cryogenics
into TI2 and TI8.
Acknowledgment
We acknowledge the support of the European Community-Research Infrastructure
Activity under the FP6 "Structuring the European Research Area" programme (CARE,
contract number RII3-CT-2003-506395)
References
[1] A. Hilaire et al, The Magnet System of the LHC Injection Transfer Lines TI2 and
TI8, August 2000, LHC Project Note 12
[2] LHC Design Report, Vol III, Dipole & Quadrupol parameter
[3] C. Daum et al, Superconduction Correction Magnets for the HERA Proton
Storage Ring, IEEE Transactions on Magnetics, Vol. 24, No 2, March 1988, p
1377 ff
[4] P. Schmüser (University Hamburg), S. Wolff (DESY), private communication
[5] LHC Design Report, Vol. I, Overview of superconducting corrector magnets in
the insertions
[6] H. Lierl, (DESY), private communication
[7] G. Horlitz et al., Computer Calculation on Steady-State Operation and Different
Modes of Cool Down and Warm Up of the HERA Superconducting Proton Ring,
Advances in Cryogenic Engineering, Vol 31, 1985, p 723 ff
[8] Design criteria of the cryogenic system for the CERN LHC injection lines /
Delruelle, N; Kouba, G; Passardi, G; Tischhauser, J, CERN-AT-93-24-CR; LHC-
NOTE-240
[9] K. H. Meß, Quench Protection at HERA, PAC Washington 1987, Washington
1987, IEEE Particle Accelerator Conference, p 1474 ff
[10] T. Ogistu et al, IEEE Transactions on Applied Superconductivity, Vol. 15, No. 2,
June 2005, p 1175 ff
- 19 -