Minutes of the 3rd LHC Insertions Upgrade Working Group held on 18th October 2007

Present: V. Baglin, S. Fartoukh, A. Ferrari, M. Giovannozzi, J. Kerby, J.-P. Koutchouk, K.-H. Meß, R. Ostojic, V. Parma, L. Tavian, E. Todesco, E. Wildner

Excused: O. Brüning, H. Mainaud Durand, R. Van Weelderen

Invited: R. De Maria

    1.    News and approval of the minutes of the last meeting

 

    2.    Space constraints for the cryostats of the triplets (V. Parma, pdf file)

Vittorio presented  different aspects related to the cryostating, the lowering, the transport conditions and the integration of the new inner triplet quadrupoles into the LHC tunnel. He then summarized the implications this has on the  dimensioning of the new low-beta quadrupoles.

Cross-section:

The cryostat dimensions are determined by the transport conditions in the tunnel and the space available around the triplets in IP1 and 5. Discussions with TS/IC indicated that there is no direct access from the surface to IP1/5. As a consequence, the diameter of a long cryostat cannot be much bigger than that of the current LHC dipoles (OD914 mm) even if the transport vehicles were to be redesigned. The size of  the tunnel where Q1 stands also limits the cryostat OD to this value. The existing LHC support posts and cryostating technology (instead of the spiders in the present MQX magnets), with a cryostat OD914 mm, limit the outer diameter of the cold mass to about 600 mm. In addition, if the available iron plates are used for the magnet yoke, then there is a preference for a cold mass OD of 570 mm (LHC dipole cold mass), which would imply that the triplet magnet aperture should be less than 130 mm, to limit the field in the yoke to 2 Tesla. These dimensions assume that the heat exchanger (OD96 mm for the present MQX magnets) is inside the cold mass (with all implications yet to be quantified).

Weight and length:

There are three possible access points from which the new equipment can be lowered: point 2 (PM25), point 6 (PX64)  or SMI2. While the first two access points are rather close to point 1 and point 5 respectively, they do not allow to lower equipment longer than 10-13 m. This limit is 15 m for SMI2, possibly increased by a few meters if the equipment is inclined and light (e.g. 23 m long QRL sections were lowered from SMI2). It is nevertheless worth mentioning that the solution via SMI2  makes even harder any increase of the transverse dimensions of the triplet cryostats due to the size of the TI2 tunnel.

Minimum inter-distance between quadrupoles:

For the LHC triplets, the interconnect lengths were determined by the initial layout and its subsequent modifications, and are larger than those in the LHC arcs. As a starting point for the new design, Vittorio suggested the MB-MB interconnect which is  500 mm long. This length has been optimized for the tooling and cannot be reduced further. In addition, a space of 2*250 mm should be reserved for the end-dome area, and another 2*150 mm for the difference between the magnetic and mechanical lengths. As a result, Vittorio proposed the value of 1.3 m for the minimum distance between magnetic lengths of two adjacent quadrupoles. If two magnets are mounted in a single cold mass, then the distance between the adjacent magnetic lengths should be taken as 300 mm. These lengths do not include the additional space needed to accommodate other equipment such as orbit correctors (slot length of about 800 mm) or BPMs (slot length of about 300 mm).

As a general comment, Karl-Hubert Mess insisted that in the radiation hard area of the inner triplets one must rely on automatic cutting and welding tooling. Implications of this condition on the dimensions suggested by Vittorio should be further checked. He also suggested to "think differently", and consider the possibility of completing the cold masses in the tunnel, thus avoiding the complications introduced by the interconnects. The audience reacted to this suggestion in a mixed way, noting that more time is needed to develop this idea and its consequences.

 

    3.    Various optics solutions for Phase I (R. De Maria, pdf file)

Ricardo presented different triplet layouts and associated optics solutions presently under investigation to fulfill the goals of the LHC insertion upgrade phase I. The latter can be divided in three categories, namely:

Mechanical aperture (n1):

The "modular" and "compact" options assume rather low quadrupole gradients and lead therefore to longer magnet assemblies (total length of 55-70 m to be compared with about 40 m both for the symmetric triplet and the low-beta-max options and 30 m for the nominal LHC triplet). These two options tend therefore to increase the beta functions in the triplet but also in the LSS magnets. As a result, while low gradient quadrupoles lead to an increase of the beam clearance in the triplet itself, aperture bottle-necks appear in the LSS magnets if the latter are not changed (which is one of the present bounding conditions for the LHC upgrade phase I). These possible limitations were not taken into account when designing the "modular" and "compact" triplet layouts and, indeed, aperture checks a posteriori gave large aperture margins in the inner triplet quadrupoles (n1=13-20), but severe limitations in the LSS (n1=4-5 in D2, n1=6-7 in Q4 and n1=4-5 in Q5, see also the prediction by Stephane at the second LIUWG meeting). In their present optics version, the low beta-max and symmetric triplet options still exhibit aperture limitations in D2/Q4/Q5. These restrictions are however much less severe and are a priori correctable by a retuning of the triplet parameters (and changes in the beam-screen orientation, see Stephane's talk at  the second LIUWG meeting).

=> ACTION: ABP

For each of the proposed options, the D1 magnet induces a severe aperture limitation assuming  a vertical beam-crossing  at the IP and gap height less than 110-120 mm (i.e. 100-110 mm available for the beam).

JPK commented that in the low-beta-max option, the reduced aperture of Q1 may present a weak point, not present in the symmetric triplet, because this magnet is the first magnet exposed to the debris coming from the IP and requires a priori a substantial aperture margin (e.g. to accommodate a dedicated absorber as this the case in the present LHC layout). Another drawback of an optimized Q1 is the additional cost for R&D and spares, while an advantage (probably modest compared to the cost) is an increase of aperture margins (about 5% reduction of peak beta-functions in Q2 and Q3) .

Strength limitations:

Both for the "modular" and "compact" options, the optics matching constraints require two MQML magnets for Q6 instead of one in the present LHC layout. This Q6 upgrade is not needed for the other two options. A crossing angle of 2*225 murad (corresponding to a 10 s beam-beam separation for a b* of 25 cm) can be obtained with the current version of the MCBC or MCBY orbit correctors at Q4, Q5, Q6 (the orbit correctors at Q7 and Q8 might also be needed in some cases to generate the crossing scheme).

Chromatic aberrations:

Higher beta functions in the triplet quadrupoles results in an increase of the non-linear chromaticity (Q'' and Q''') and a strong chromatic dependence of the twiss parameters (off-momentum beta-beating).  These chromatic aberrations are therefore higher for the "modular" and "compact" options. For the "low beta-max" and symmetric triplet options, the off-momentum beta-beating already reaches 30% and 100% for a momentum deviation of 0.0003 (bucket height) and 0.0008, respectively. While the non-linear chromaticity can be minimized by imposing a pi/2 phase advance between  IP1 and IP5, this remedy cancels the off-momentum beta-beating in only half of the machine. This half  of ring can then either contain  the momentum cleaning insertion IR3 or the betatron cleaning insertion IR7. The final choice will depend on the collimation requirements which are not yet finalized. The proposed remedy also cancels the off-momentum beta-beating at the two IP's,  in the inner triplet and up to D2-Q4.  The impact on the luminosity and the mechanical acceptance of the LSS, except perhaps for Q5 and Q6, is therefore minimal. Jean-Pierre Koutchouk suggested to try using the lattice sextupole families to cancel the off-momentum beta-beating all around the ring. Stephane expressed his doubts on how it could work (this is like trying to correct globally a strong dipolar kick without any perturbation on the closed orbit).

Dynamic aperture (w/o beam-beam):

For the "modular" and "compact" options, the triplet is transparent for the LHC dynamic aperture in collision (17 and 22 sigma respectively for triplet only) and depends only on the field imperfections of the LSS magnets, in particular Q4 (DA=11 and 16 sigma's, respectively with all field errors included). This is due to the fact that the decrease with aperture of the field imperfections in the triplet quadrupoles is expected to be faster than the increase with the beta functions of the non-linear resonance driving-terms. On the other hand, the LHC dynamic aperture for the symmetric triplet and the low beta-max options is dominated by the field imperfections of the inner triplet and is 12 and 14 sigma's respectively (all errors included). Ricardo did not report on any tracking studies taking into account the beam-beam effect but Jean-Pierre Koutchouk quoted a value of 4 sigma's obtained by Ulrich Dorda (AB/ABP) for the "compact option", which, a priori, should further reduced for the "modular option" (even onger quadrupole assemblies), due to the number of parasitic beam-beam crossings which is  larger for these two layouts (see presentation by Ulrich made at the ABP/LCU meeting for more details).

Summary and discussion:

In addition to several  limitations of the "modular" and "compact" options summarized above (aperture restriction in the present LSS magnets, need for doubling Q6, DA with beam-beam),  and in view of the hardware constraints presented by Vittorio, it seems that quadrupole magnets with an aperture much larger than 130 mm might represent a real challenge for integration in the LHC ring. This limitation might also be valid for the LHC insertion upgrade phase II. While a recommendation to give low priority to the "modular" and the "compact" options was not explicitly given during the meeting, the general impression was that the low beta-max and the symmetric options were more promising and should be pursued further. Ranko requested that these two options should be further studied, and if needed their optics retuned, in the light of the constraints discussed in the meeting. In particular, orbit correctors should be introduced, correct position for the BPM found and the updated interconnect lengths implemented.

=> ACTION: ABP

 

    4.    Follow-up of actions: Beam-screen geometry (V. Baglin, ppt file, and C. Rathjen, ppt file) and maximum aperture of two-in-one quadrupoles (E. Todesco)

    4-a)    Due to lack of time, the slides of Vincent and Christian Rathjen were not presented during the meeting but only summarized by Vincent. The clearance of 2*0.7 mm between the beam-screen outer dimensions and the cold bore inner diameter can be kept unchanged  (needed to accommodate the so-called "sliding rings"). A priori, the beam-screen could also keep its  race-track shape with, as present, a difference of 2*4.8 mm between gap height and diameter. This clearance given by the diameter of the beam-screen capillaries has however to be confirmed by ACR. It depends on the beam-screen operating temperature (5-20K currently), the cooling capacity and the expected heat load; all these factors need to be further addressed.

=> ACTION: ACR

          Vincent reported that the thickness of the beam screen may need to be increased to 2.5 or 3 mm (compared to 0.6 mm in  the current MQX magnets) such that the beam screen can resist the transient forces during magnet quenches. This thickness has been obtained using pessimistic assumptions (beam-screen inner diameter of 150 mm and a gradient of 200 T/m). As the forces induced by the eddy currents during quench scale with the fifth power of the beam-screen radius and quadratically with the quadrupole gradient, a thickness of  1.5-2  mm may be suitable but has to be validated (confirmed by Christian after the meeting).

        Vincent also summarized his first estimates of gas density assuming an effective beam-screen transparency of 5% (2% in the nominal LHC). The gas load in the LSS is estimated to be a factor of 10 higher than the nominal LHC (i.e. 1014 H2-equivalent molecules per cubic meter), and is dominated by the electron multipacting effect. However, for a fully conditioned surface (as assumed for the nominal LHC), a gas load of a factor of about 2 higher is expected, and could be further reduced by increasing the beam-screen operating temperature to 40-60K.

    4.b)    Due to lack of time, the presentation of Ezio (maximum aperture of a two-in-one quadrupole as a replacement for Q4 in a second phase, see minutes of the second LIUWG meeting) was postponed to the next meeting. As a preliminary information, Ezio said that an aperture of 110 mm looks reasonable for  a two-in-one quadrupole  with an aperture separation of 194 mm, based on a single layer design. However,  its operating gradient would be about 85 T/m (compared to 160 T/m for the present Q4). Stephane commented that this gradient could be acceptable for optics solutions presently under investigation as there is space around D2 for longer Q4 magnet. In addition, he commented that if Q4 had an aperture of 110 mm, and with the aperture of D2 consistently larger, the aperture of the LSS quadrupoles would no longer be a criteria for defining the inner coil diameter of the phase I inner triplet, a least up to an aperture limit of around 170 mm (to be compared to the first guess estimate of 145 mm, see minutes of the second LIUWG meeting).

 

    5.    A.O.B.

Next meeting scheduled for 1 November 2007: Triplet aperture and collimation issues (R. Assmann).

S. Fartoukh