SB2009 Proposal Document
Release 1.1
December 2009
ILC Global Design Effort
Prepared by the Technical Design Phase Project Managers and
the Accelerator Design and Integration Team Leaders
Project Managers:
M. Ross, N. Walker, A. Yamamoto
AD/I Team Leaders:
C. Adolphsen, D. Angal-Kalinin,
W. Bialowons, A. Brachmann,
J. Carwardine, J. Clarke, A. Enomoto,
T. Himel, S. Fukuda, P. Garbincius,
R. Geng, C. Ginsburg, S.Guiducci,
J. Kerby, V. Kuchler, T.Lackowski,
A. Latina, C. Nantista, J. Osborne,
M. Palmer, J.M. Paterson, M. Pivi,
A. Seryi, T. Shidara, N. Solyak, N. Toge,
J. Urakawa
Table of Contents
SB2009 Proposal Document .................................................................................................................... 1
1 Introduction.......................................................................................................................................... 1
2 SB2009 Overview.................................................................................................................................. 3
3 Layout ................................................................................................................................................. 16
4 SB2009 Proposal ................................................................................................................................. 20
4.1 Issues of Main Linac Accelerating Field ....................................................................................... 20
4.2 Electron Source ........................................................................................................................... 22
4.3 Positron Source ........................................................................................................................... 25
4.4 Damping Rings ............................................................................................................................. 33
4.6 Main Linacs .................................................................................................................................. 43
4.7 Beam Delivery Systems ............................................................................................................... 62
4.8 Conventional Facilities................................................................................................................. 67
5 Cost Studies ........................................................................................................................................ 80
6 Risk Analysis ....................................................................................................................................... 85
Appendix A. Availability Task Force Report ........................................................................................... 89
SB2009 Proposal Document
1 Introduction
This report documents the proposal from the ILC Project Management Design Team for major
changes to the published RDR [Intro1] baseline, as a result of the effort in accordance with the GDE
R&D Plan [Intro2]. It represents the culmination of approximately one-year of study and includes
new results from the on-going risk-mitigating R&D, as well as a critical review of the existing
Reference Design.
At the end of the Technical Design (TD) Phase in 2012 the GDE will publish a Technical Design Report
(TDR). One of the primary goals of the TD Phase is to constrain the VALUE estimate to the RDR value
of 6.6 BILCU. The TDR will contain a description of ILC design in sufficient detail to support an
updated VALUE estimate. The updated reference design will incorporate the results from the ongoing risk-mitigating R&D and the Accelerator Design and Integration (AD/I) studies, but will remain
strongly based on the existing published Reference Design (RDR).
With these goals in mind, the Project Management considered it most beneficial to focus the
available (and limited) design resources on a re-examination of the reference design during 20082009, with a view to establishing a better and more cost-effective baseline before commencing TD
Phase 2 (Summer, 2010). The intention is to use the modified baseline as the basis for detailed design
and cost work foreseen for the TDR in the two years to follow. During TD Phase 2 the baseline will be
maintained and evolved under strict change control.
The adopted approach to re-baselining is based on an assumption that the RDR design – although
sound – is both conservative in many of its design decisions, relatively immature from a detailed
engineering standpoint, and is “performance-driven” as opposed to cost optimised. Conventional
Facilities and Siting (CFS) was identified early on as a strong focus for design optimisation; in
particular the reduction of underground civil construction, achieved by a critical re-evaluation of the
criteria driven by the accelerator design assumptions.
With nearly all ILC worldwide resources focused on risk mitigating R&D (most notably SCRF), a full
study of system engineering is not practical: Therefore, a strategic decision was made by the Project
Management to focus the available design resources on a few top-level design elements which have
a large cost leverage (at the level of a few % each), and where alternative designs exist which are
both feasible and more cost-effective. The Project Management believes that the resulting proposed
baseline is:

a simplification of the scope of the construction project with a potential for cost reduction of
up to one-billion ILCU (cost-reduction at this level will supply a margin to accommodate
possible cost inflation incurred during TD Phase 2, most likely at the unit cost level);

achievable while maintaining a level of technical risk which is comparable with the RDR
baseline.
Further to the basic (and in many respects generic) accelerator layout, the study has also begun to
address the impact of potential site constraints in a more realistic way. Particular attention is paid to
allowing flexibility in design decisions in support of future possible host sites. This aspect is
considered fundamental to the proposed approach, and will be reflected in the Project
Implementation Plan (to be published as part of the TDR). Several layout and design flaws with the
1
reference design have also been identified as a direct result of these studies; these have begun to be
addressed as part of this re-baseline proposal. The Project Management believes this process will
naturally lead to a better, more cost-optimised design, which in many respects is more complete
than the RDR.
The report is divided into six sections:

Section 1 – this introduction.

Section 2 gives an overview of SB2009

Section 3 gives a bird's eye view description of the overall system layout.

Section 4 discusses the proposed design changes in more detailed, including additional studies
and R&D that remain to be done (TD Phase 2). This section is sub-divided into the following subsections:
 main linac (single tunnel solution, high-level RF, and accelerating gradient);
 electron and positron sources;
 damping rings;
 bunch compressors;
 beam delivery systems;
 conventional facilities.

Section 5 summarises the impact of the proposed modifications on the RDR VALUE estimate.

Section 6 gives a brief outline of risk analysis.
Finally, a report on the system availability studies made is given in an Appendix.
The following sections are not intended to be a design report. Rather, they contain a relatively
conceptual description of the proposed modifications to the RDR baseline. It is acknowledged that
not all questions raised have been answered at this time, but the Project Management Team believes
that enough of the key issues have been addressed so that these modifications are proposed based
on sound judgement. Many of the remaining issues (details) – as well as new R&D required as a
result of the proposal – will feedback into the GDE R&D Plan, to be updated at the end of TD Phase 1.
This Proposal document will be submitted to the Project Director for analysis and review in early
2010, before launching TD Phase 2. In due course, reviews, high-level approval, and possible
refinements are expected before the establishment of the new baseline configuration.
References
[Intro1] ILC Reference Design Report Volume 3 - Accelerator, August, 2007,
http://www.linearcollider.org/cms/?pid=1000437
[Intro2] ILC Research and Development Plan for the Technical Design Phase,
http://ilcdoc.linearcollider.org/record/15442/files/TD_Phase_R%26D_Report.pdf?version=2
2
2 SB2009 Overview
In this section, we provide an overview of the key SB2009 elements and their rationale. Section 2.1
introduces the overall approach and goals of the Accelerator Design and Integration (AD/I) process.
Section 2.2 introduces the key proposed top-level baseline modifications and then summarises each
of them. Sections 2.3 and 2.4 briefly summarise the impact on cost and physics related performance
respectively.
2.1 Rationale and Methodology
The baseline modifications described in this document were arrived at with several specific goals in
mind:

Overall cost reduction - Any opportunities for cost reduction should be taken, in as much as
they do not unacceptably impact performance or increase technical risk.

Improved cost balancing - Cost margins created as part of the cost-reduction exercise, can be
made available for other subsystems which incur increased (estimated) construction costs.

Improved understanding of system functionality - Attempts at understanding of any
performance impact force a careful analysis of systems' functionalities, strengths and
vulnerabilities; this has a critical value on its own beyond cost-reduction.

More complete and robust design - Revisiting many of the design and implementation details
that were not completely covered during the RDR design phase. These efforts, when made
appropriately, will improve the overall robustness of the ILC systems design.

Re-optimised R&D plans - Improved understanding of the system functionalities and
performance issues will help the Project Management in producing a re-optimised and more
effective global R&D plan to pursue in TD Phase 2.
The key design elements for the studies were primarily identified by their cost impact (based on the
RDR VALUE estimate). The list of proposed design modifications were iterated and reviewed during
the second-half of 2008, resulting in a final set of proposals for study that were published in January
2009 [Over1]. Further studies and refinements evolved into an integrated proposal by the Project
Management for a straw-man baseline (SB2009) in May 2009[Over2]. The baseline modifications
proposed in this document are the end result of iteration by the Accelerator Design & Integration
(AD/I) Team (see below), as well as input from various external reviews and the broader ILC
community.
The AD/I Team, under the direct leadership of the Project Managers, includes:




A set of key leaders who organized the efforts and led the overall discussion and direction.
Individual “area” and “global” groups who are responsible for specific accelerator subsystems or
accelerator-wide facility systems.
The costing group who are responsible for collecting and organizing the cost implications.
Ad-hoc groups formed under the leadership of the Project Managers to address the risks in
terms of technical development toward construction of the ILC and the systems availability.
The work was orchestrated around several GDE workshops and special face-to-face focus meetings,
as well as numerous weekly teleconferences.
3
2.2 Summary of Proposed Changes
Figures 2.1 shows an approximate comparison of the overall ILC layout as documented in RDR and as
put forth in the proposed new baseline.
Figure 2.1: RDR layout (left) and the re-baseline layout in proposal (right).
Both the RDR layout and the proposed new layout require a site footprint of approximately 31km
length, consistent with maximum operational beam energy of 250 GeV (500 GeV centre-of-mass
energy). Exact lengths still remain to be made consistent, but any future adjustments are expected to
have a negligible impact on the estimated cost increments.
The following changes are proposed for the ILC design baseline configuration as described in the RDR,
are summarised in the following seven primary top-level Working Assumptions (WA):
WA1. A Main Linac length consistent with an average accelerating gradient of 31.5 MV/m and
maximum operational beam energy of 250 GeV, together with a High-Level RF distribution
scheme which optimally supports a spread of individual cavity gradients.
4
WA2. A single-tunnel solution for the Main Linacs and RTML, with two possible variants for the
High-Level RF (HLRF):
a) Klystron cluster scheme (KCS);
b) Distributed RF Source scheme (DRFS).
WA3. Undulator-based positron source located at the end of the electron Main Linac (250 GeV), in
conjunction with a Quarter-wave transformer as capture device.
WA4. A lower beam-power parameter set with the number of bunches per pulse reduced by a
factor of two (nb = 1312), as compared to the nominal RDR parameter set.
WA5. Reduced circumference Damping Rings (~3.2 km) at 5 GeV with a 6 mm bunch length
WA6. Single-stage bunch compressor with a compression ratio of 20.
WA7. Integration of the positron and electron sources into a common “central region beam
tunnel”, together with the BDS, resulting in an overall simplification of civil construction in
the central region.
The above WAs are the reference for the recent design work, specifically for the CFS and cost efforts.
It is however important to note that the WAs are not entirely independent of each other –
particularly WA4 and WA5, and WA3 and WA7. These dependencies are explained in the following
sub-sections, which are organised more thematically than the above list.
The key top-level design parameters for the SB2009 proposal are summarised and compared to the
RDR nominal parameters in Table 2.1.
5
Table 2.1: Selected design parameters for SB2009 compared to the RDR nominal parameter set.
RDR
(nominal)
SB2009
Beam and RF Parameters
No. of bunches
2625
1312
Bunch spacing
ns
356
670
beam current
mA
9.0
4.8
Avg. beam power (250 GeV)
MW
10.8
5.4
Accelerating gradient
MV/m
31.5
31.5
Pfwd / cavity (matched)
kW
294
156
3.5E+06
Qext (matched)
6.6E+06
Tfill
ms
0.60
1.12
RF pulse length
ms
1.57
2.00
RF to beam efficiency (matched)
%
61
44
Norm. horizontal emittance
mm.mrad
10
10
Norm. vertical emittance
mm.mrad
0.040
bunch length
mm
0.3
0.3
horizontal β*
mm
20
11
horizontal beam size
nm
640
470
IP Parameters
0.035
no trav.
focus
with trav.
focus
vertical β*
mm
0.4
0.48
0.2
vertical beam size
nm
5.7
5.8
3.8
19
25
38
2
4
3.6
Dy
δEBS/E
%
Avg. PBS
kW
Luminosity
cm -2 s -1
6
260
200
194
2.00E+34
1.50E+34
2.00E+34
2.2.1 Main linac – SCRF: choice of operational accelerating gradient (WA1)
The most highly leveraged R&D that has the greatest potential return in terms of performance and
cost is the development of cavity gradient. The cost impact is seen primarily in the linac length
required to achieve 500 GeV in the centre of mass. In 2005, a goal of an operational accelerating
gradient of 31.5 MV/m per cavity with Q0 ≥ 0.8 x 109 in pulsed operation was adopted for the
Reference Design, with a cavity production specification of ≥35 MV/m with Q0 ≥ 1010 in a low-power,
CW vertical acceptance test. These parameters were adopted within the paradigm of developing a
forward-looking design, and assumed a worldwide R&D effort to routinely establish these
parameters. This R&D is currently on going and represents the largest fraction of the global ILC R&D
resources.
For the current SB2009 proposal – and in lieu of expected developments in the SCRF R&D – these
specifications for the accelerating gradient and Q0 are left unchanged. Thus the global CFS
requirements (linac length, cryogenic requirements) are left as per the RDR. This is primarily to allow
the CFS groups to continue with their design and cost plans. It is noted that any subsequent changes
in accelerating gradient, arising from review of the R&D status will (to the first-order) result in a
scaling of the main linac which is relatively straight-forward, and will not require major re-evaluation
of the design [Over3].
In contrast to the RDR baseline, however, we propose the HLRF systems support a spread of
operational gradients from cavity to cavity. This will allow lower performance cavities to be utilised,
assuming that high-performing ones maintain the average. The approach has impact on both the
definition and accounting of cavity production yield (together with acceptance criteria), as well as
required operational overhead in the accelerator (RF power and gradient). The global R&D groups
are currently evaluating these issues. As of this writing every indication is that support of a spread of
cavity gradients will provide a better cost optimised solution, albeit at the expense of a more
complex HLRF distribution system and lower overall RF-to-beam power efficiency.
The final choice of average operating gradient will be taken during TD Phase 2 after review of the
worldwide R&D status. (An interim review is currently on going.) The choice of parameters must
necessarily include an analysis of cavity and cryomodule production statistics, as well as results of
operational overhead under full beam-loading conditions. The choice will also be influenced by
production models, assumed yield distributions and must be cost-optimised and be technically
justified. Finally, prospects in the R&D beyond the TDR should also be taken into consideration, given
the uncertainty in the formal start of project construction. The ongoing SCRF R&D will be tailored in
TD Phase 2 to address these critical cost and performance issues.
For SB2009 it is acknowledged that working assumptions for several key margin and overhead
parameters still need to be addressed. The exact amount of gradient and RF power overhead
required for operational margin remains to be specified. The choices made at this juncture should be
consistent with the goals proposed for the TD Phase 2 R&D Plans, but should maintain as realistically
as possible the ‘forward looking’ paradigm established during the RDR.
2.2.2 Main linac – Single tunnel configuration (WA2)
Foremost among the proposed changes is the adaptation of the main linac tunnel configuration to
one with only a single, accelerator-enclosure tunnel, thereby eliminating the support equipment
tunnel proposed in the Reference Design. Several key studies and strategic initiatives make it
attractive to propose this change for the following reasons:
1. The single tunnel configuration is a simpler underground construction, effectively removing
~26 km of underground (support equipment) tunnel.
7
2. Safety studies commissioned in each region (Asia /Japan, Americas/US and Europe / CERN)
found that valid strategies could be realized for single-tunnel life safety egress in each of the
regional locations studied (allowing for some variations according to local regulations).
3. Recent studies on High Level RF power sources and distribution have resulted in feasible new
concepts for these systems that are much more suited to a single-tunnel solution than those
proposed in the RDR (namely the Klystron Cluster Scheme (KCS) and the Distributed RF
Source Scheme (DRFS)).
4. Availability studies of the Main Linac on the proposed single-tunnel and new HLRF systems
configurations show that an acceptable performance can be achieved with appropriate
engineering of sub-system designs.
The two proposed novel HLRF solutions (KCS, DRFS) are described in more detail in Section 4.6. These
different approaches provide options for specific sites where local constraints may favour one
solution over the other. Allowing such flexibility in the designs (multiple configurations) goes beyond
the “generic site” approach used in the RDR. Dealing with the reality of the requirements of potential
national hosts and their proposed host-sites (so-called “bid-to-host”) is an important aspect of the
Project Implementation Plan. However, the desire to maximally maintain a single design in both the
layout and the core technologies remains paramount, in an attempt to make the most efficient use
of the worldwide available resources.
2.2.3 Reduced beam-power parameter set (WA4, WA5, WA2)
The proposed baseline change with largest anticipated cost saving (both construction and operation)
is the reduction of the beam-power by 50%. The beam-power scales as N nb, where N is the number
of particles per bunch, and nb the number of bunches per pulse. Since the luminosity L ∝ N2 nb it is
proposed to halve nb, while keeping N the same as the 2007 Reference Design. A factor-of-two
reduction in nb allows:

a reduction in the number of klystrons and modulators (peak current/power reduction), in
this case by ~50% over the RDR number;

the possibility to halve the circumference of the damping rings to approximately 3.2 km.
The luminosity can be recovered in general by pushing the beam-beam parameters, which are not
considered a major cost driver. The more demanding parameters are however a potential increase in
performance risk, and have potential issues for the physics and detectors.
The significant reduction in RF power source together with the smaller damping rings offer a
considerable cost reduction and overall simplification in terms of construction scope. For the RF
power source at least, the option is attractive as it opens up the possibility to restore the “missing
klystrons” at a later stage, allowing operation with an increased beam current as a possible
luminosity upgrade path. The amount of infrastructure, such as conventional facilities, included
during the construction phase to support this upgrade still requires study to be examined for further
cost-optimisation. However, for the smaller circumference damping ring, the upgrade would require
increasing the current by up to a factor of two (beyond the Reference Design specification); collective
effects in the damping rings (most notably e-cloud) may well be a bottleneck to increasing the
number of bunches, an issue which is currently under study (part of the on-going e-cloud R&D
programme). Hence an upgrade along these lines represents an increased risk over the Reference
Design parameters.
8
The obvious cost and construction scope benefits for this option must be weighed against a
perceived higher risk in achieving the design luminosity due to the more demanding beam-beam
parameters. The impact of the higher Beamstrahlung on the detectors and physics programme must
also be reviewed. Finally, the effective lower beam current reduces the overall RF to beam power
efficiency, and so the reduction in average RF power is not a full factor-of-two but closer to 38%.
There has been a conscious decision to maintain the same power handling capability as assumed in
the RDR for the beam dumps, positron target and capture, etc, and other systems which might be
difficult to upgrade later and after some time in operation. The benefits of initially over-rating these
systems (reduced technical risk), while supporting possible future beam-power upgrade, are
considered to out-weigh any possible cost saving.
2.2.4 Central Region Integration (WA7, WA3, WA5)
The motivation for the changes in the Central Region Integration as proposed in this document is the
simplification of the central region tunnelling and civil engineering. By combining their functions
within a single beam tunnel, the complexity of the underground excavation in the central region can
be substantially reduced. A more detailed description and historical evolution of the current scheme
is given in Section 3.
The modifications compared to the Reference Design include:

The Damping Rings have been moved vertically into the same plane as the BDS and shifted
horizontally to avoid the Detector Hall (see Figure 2.1 right). This removes the need for the
long (~2 km) vertically sloped beam tunnels (so-called escalator), which can be replaced by
much shorter horizontal transfer tunnels. The Damping Rings tunnel can now also share one
shaft with the Detector Hall.

Since the BDS magnets do not require a large amount of transverse tunnel space, it is
feasible to house the electron source and the 5 GeV injector linac in the same tunnel as the
positron BDS, thus removing the need for a separate beam tunnel.

Similarly, the undulator-based positron source and associated 5 GeV booster linac can be
more efficiently accommodated at the exit of the main electron linac (electron BDS, see
Sections 2.2.5, 4.3 and 4.7 for more details).

An additional ~450 m beam path length is maintained in the positron system for bunch
timing.

Finally, an additional 500 MeV linac can be incorporated into the e+ source region which,
when used in conjunction with the e+ source photon target, can facilitate a low charge
auxiliary source for commissioning and tuning etc.
In addition to the simplifications in civil construction and the associated cost saving, the proposed
scheme further consolidates sources of high-radiation hazard into the central region, which is
considered beneficial with respect to environmental impact.
Although a significant simplification in the scope of the civil construction and overall layout, the need
to support multiple beamlines and their support infrastructure in a single tunnel is challenging.
Availability and operational issues (impact on personnel protection and access zoning) needs to be
reviewed. Use of 3D-CAD is essential in understanding both the topology of the accelerator
beamlines and issues such as installation, safety etc. Initial studies in support of this proposal indicate
that acceptable solutions exist.
9
Because of the multiple beamlines – including the two superconducting 5 GeV booster linacs – it has
been initially decided to retain a parallel service tunnel in this region to house the power supplies,
klystrons etc. As the 3D-CAD work evolves into more details of the hardware implementation, this
will be reviewed and cost-optimised.
2.2.5 Changes to the undulator-based e+ source (WA3, WA7)
The motivation for moving the undulator-based positron source from the 150 GeV point to the end
(250 GeV) point in the electron Main Linac is primarily to consolidate technical systems and radiation
sources as already discussed briefly in Section 2.2.4. However, there are several additional impacts of
this particular modification that merit additional attention.
Several “system integration” aspects should be noted as direct benefits of moving the source to the
high-energy end of the Main Linac:

Machine protection of the undulator is combined with machine protection of the Beam
Delivery system, resulting in a net beamline length reduction of about 400 m

The low-energy positron transport system is shortened by several kilometres.

The positron source, which represents a high-radiation environment, moves within ±2.5 km
of the centre of the accelerator complex where other high-radiation sub-systems (e.g. the
main dumps) are located. This is a preferred feature in some potential sites, since it
consolidates most of the radiation hazard into a single “central campus area”.

The beam-line ‘chicane’, which was required in the RDR design can be replaced by a dogleg,
and better integrated into the BDS lattice. The narrow energy acceptance associated with the
RDR chicane is also moved to the BDS where there is already a natural bandwidth limit of a
few percent; this greatly facilitates beam-based alignment in the linac.

The RDR solution (i.e. Locating the source at the 150 GeV point in the electron Main Linac)
requires the upstream linac to perform to its full specifications. Failure to achieve these
highest performance goals during early stage commissioning could result in poor source
performance, as there is little or no safety margin. Initially achieving 80% full acceleration in
the main linac would still allow physics reach to 400 GeV centre-of-mass: However, this
would come at a cost of reduced positron yield below 1.0.
An additional proposed modification, which is independent of the central region integration, is the
adoption of a simpler capture magnet (Quarter Wave Transformer, QWT) immediately after the
positron-production target. This simpler magnet is definitely feasible, and represents a more
conservative (reduced-risk) option in comparison with the Flux Concentrator (FC) assumed in the
RDR. However, use of a QWT reduces the capture yield by a factor of 2 when compared to the Flux
Concentrator – a difference that must be mitigated by a significantly longer undulator, resulting in
higher incident power on the production target. For this reason, the FC is still a very desirable option
and continues to be a priority R&D item.
One fundamental ramification of the relocation of the source is the need to run it at varying electron
energies, depending on the required centre-of-mass energy. Specifically, the impact on luminosity
below centre-of-mass energies of 250-300 GeV, where the positron yield drops rapidly. In this region,
alternate-pulsing schemes can be implemented to regain the luminosity (to within a factor of two) of
the 1/ scaling value (See section 4.3 for details). For detector calibration at the Z mass, the
proposed low-charge auxiliary source will be sufficient.
10
At higher energies, the higher-energy electron beam becomes an advantage, offering very large
production margins (up to a factor of 5 at 250 GeV), giving further risk reduction at the higher centreof-mass energies where luminosity is likely to be at a premium. For the same reasons, this opens up
the potential for very high-polarised beams at these top-range energies.
Finally, it should be noted that the original RDR concept for Ecm< 300 GeV was based on decelerating
the electron beam after the positron source. While conceptual feasible, this mode of operation is not
without challenges and still requires careful study.
2.2.6 Single Stage bunch compressor (6 mm bunch length) (WA6, WA5)
Adoption of a new lattice for the Damping Rings has allowed the extracted bunch length to be
reduced from the RDR value of 9 mm to 6 mm. (This modification has been implemented both for an
earlier 6.7 km ring and the currently proposed 3.2 km ring.) The shorter bunch length opens up the
possibility to adopt a simpler single-stage compressor with a compression ratio of 20, still achieving
the nominal RDR value of 0.3 mm at the Interaction Point.
The benefit of this is overall simplification of the bunch compressor system, as well as a saving in cost
(mostly tunnel length). Beam dynamics studies have also shown that the emittance degradation is
reduced. The disadvantage is the loss of flexibility in bunch length tuning and the ability to reduce
the bunch length to 150 m. These aspects can be seen as reducing the operational risk margin.
2.3 Cost Increment
One of the primary motivations for the proposed baseline modifications is the reduction in the
overall cost. This is expected to help constrain the TDR updated VALUE estimate to the RDR value of
6.6 billion ILCU, allowing some margin for incurring increased estimated costs during TD Phase 2.
During the early analysis of the proposed modifications, a ballpark estimate suggested a possible
total saving of approximately 15%.
During the last six-months (May - December, 2009), a more detailed design and costing effort has
been made. The cost increments associated with the proposed modifications have been scaled
directly from the RDR VALUE estimate basis. No attempt was made to provide new updated unit cost
estimates at this time, with the exception of those components that are new to SB2009 (notably the
HLRF system components). Thus, the increments reported here are direct comparisons to the
reported RDR estimate.
To date, the total reduction of the modifications proposed in this document amount to ~13% of the
RDR value estimate. It is expected that this will increase as further cost-driven iterations and
refinements are made during the course of TDP-2.
A high-level breakdown and explanation of these increments is given in Section 5.
2.4 Availability studies
As for the published Reference Design, availability (and design for high-availability, HA) remains an
important issue for SB2009. To-date, the focus has been on the removal of the Main Linac service
tunnel, where availability considerations have played a key role. An availability task force was set-up
by the Project Managers with a goal of finding a realistic and workable HA solutions for the Main
Linac single-tunnel configuration, including the two new proposed HLRF solutions (KCS and DRFS).
The studies were substantially based on availability modelling using the purpose-written Monte Carlo
tool AVAILSIM, which has been used to model and compare various configurations and to provide HA
specifications for the Main Linac components. To-date, the results indicate that satisfactory HA
engineering solutions exist for a single-tunnel, which add approximate 1% unscheduled downtime to
11
the overall availability during simulated luminosity runs at the maximum centre-of-mass energy of
500 GeV. (This seemingly small number actually corresponds to a doubling of the downtime of the
Main Linac as compared to the RDR two-tunnel solution.) Of primary importance is the energy
overhead required to maintain the top-most energy of the machine. The simulations show that a
3.5% overhead is typically required for both KCS and DRFS, providing the meantime between failure
(MTBF) specifications for the RF hardware are achieved. The availability simulations have provided
key input for the design of both HLRF systems, where acceptable HA solutions have been found
corresponding to the MTBF requirements.
Other aspects of SB2009 – in particular the central region integration – have not as yet been
modelled in detail. This remains to be done in TD Phase 2. However, early cursory examinations
indicate that no major differences with the RDR solutions exist.
It should be noted that the simulations and subsequent requirements on MTBF are based on a set of
specific assumptions for the required uptime (15% budgeted), maintenance models (including
preventive maintenance) during operations as well as the impact of various failure modes. Achieving
the required component MTBFs based on these assumptions remain in many cases an R&D
challenge. However, the differences between the RDR and SB2009 appear marginal in this respect.
A full report from the Availability Task Force can be found in the Appendix.
2.5 Risk assessment
An important aspect of evaluating the soundness of any design is an assessment of the technical risk
associated with it. A Risk Register was produced for the Reference Design that catalogued the top 50
or so critical design decisions, and attempted to rate them in terms of their perceived risk to the
success of the project. Those design decisions with the highest risk (accelerating gradient, electron
cloud effects in the DR etc.) are being addressed with the highest priority by the on-going worldwide
R&D.
As we approach the end of TD Phase 2, it is important to review and re-assess the technical risk, in
view of progress in R&D, and in particular the design modifications proposed in this document. Such
a review and re-assessment requires a well-defined methodology (beyond what was done for the
RDR).
At this time, it is not possible to provide a full quantitative and formal risk assessment, as this takes
time. Such an assessment is currently being planned and should be complete by the end of TD Phase
1 (mid-2010). A methodology is currently being developed by the Project Management, which will be
applied during the assessment.
What is perhaps more pertinent to this proposal at this time is the perceived change in risk with
respect to the Reference Design. The ‘increased risk’ associated with the design modifications are
documented throughout technical descriptions given in Section 4. In many cases, this perceived risk
is difficult to quantify, and by nature the assessment of whether the risk has increased (or decreased)
tends to be subjective.
In Section 6, an attempt is made to summarise the key incremental risk elements of SB2009, as
compared to the Reference Design. A brief description of the methodology being developed is also
given.
12
2.6 Physics scope impact
While all of the proposed modifications have at various levels an impact on the overall performance
risk of the machine, two working assumptions have a direct and quantifiable impact on the physicsrelevant machine parameters:


WA3 Locating the undulator-based positron at the end of the main electron linac
WA4 Reduced beam-power parameter set
In this section we briefly summarise the physics related parameters. More specific details can be
found in the relevant subsection of Section 4.
2.6.1 Operation at 500 GeV Centre-of-mass Energy
All of the work to-date, including the RDR phase, has been focused on an optimised parameter set at
the maximum energy of 500 GeV centre-of-mass. The basis for design during the RDR was the socalled “parameter plane” which attempted to give some redundant scope and safety margin to
achieve the design luminosity of 2×1034 cm-2s-1. The requirements for the machine were based on
being able to accommodate the complete parameter ranges specified by this plane.
Table 2.2: RDR parameter plane ranges compared to SB2009 specifications
(TF refers to Travelling Focus).
Bunch population
Number of bunches
Linac bunch interval
RM bunch length
Normalized horizontal emittance at IP
Normalized vertical emittance at IP
Horizontal beta function at IP
Vertical beta function at IP
RMS horizontal beam size at IP
RMS vertical beam size at IP
Vertical disruption parameter
Fractional RMS energy loss to
beamstrahlung
Luminosity
min
1
1260
180
200
10
0.02
10
0.2
474
3.5
14
1.7
x 1010
ns
m
mm-mr
mm-mr
mm
mm
nm
nm
%
x 1034cm-2s-1
RDR
nominal
2
2625
369
300
10
0.04
20
0.4
640
5.7
19.4
2.4
2
max
2
5340
500
500
12
0.08
20
0.6
640
9.9
26.1
5.5
SB2009
no TF
with TF
2
2
1312
1312
530
530
300
300
10
10
0.035
0.035
11
11
0.48
0.2
470
470
5.8
3.8
25
38
4
3.6
1.5
2
For 500 GeV centre-of-mass operation the most direct impact is from the WA4, the reduced beam
power. From simple scaling, a reduction in the number of bunches per pulse by a factor of 2 reduces
the luminosity by the same factor (luminosity per bunch crossing remains the same).
The RDR parameter plane allows us to regain some if not all of this luminosity in principle. Table 2.2
gives the ranges of key parameters from the RDR compared to SB2009.
The approach is to push harder on the beam-beam parameters to increase the specific luminosity.
The largest modification is a reduction in the horizontal and vertical beta functions. These
modifications have impact on beamstrahlung as well as the required beam-halo collimation depth (or
conversely on the allowed apertures within the IR).
However, all of these parameters (with the notable exception of the vertical disruption parameter for
the Travelling Focus parameters) are still within the parameter plane for which the BDS and IR was
13
designed. It is acknowledged, however, that the risk margin in these parameters is decreased,
although the impact of this is difficult to quantify.
Currently SB2009 has two feasible parameter sets for 500 GeV centre-of-mass:

Pushing the beam-beam into a high-disruption regime (Dy ~ 25), which results in a luminosity
of approximately 1.5×1034 cm-2s-1 – a reduction of ~25% over the RDR value. The reduction in
the peak luminosity (top 1% in E/E) is slightly worse at ~37%. Beamstrahlung (and therefore
beam-beam backgrounds) at 4% is higher than the RDR nominal value by a factor of 1.6 but is
still significantly less than the maximum RDR specified value of 5.5%. The higher vertical
disruption parameter leads to tighter tolerances on stability at the collision point (beambeam based feedback). The same value was however proposed for TESLA, which has been
extensively studied. The approach requires virtually no additional hardware or modification
to the BDS/IR design.

A second approach relies on a technique known as a travelling focus, where the focus at the
interaction point is adjusted along the bunch, effectively compensating the hourglass effect
and allowing the vertical beta function to be reduced below the bunch length. In principle
such a scheme achieves higher luminosities, and can compensate completely for the factorof-two reduction due to the lower bunch number. The beamstrahlung is higher than the RDR
nominal parameter value, but again within or close to limits of the RDR specification. The
very high disruption parameter does have implications for luminosity stability and tuning
issues; these still remain to be resolved and will require further beam-beam simulation. The
technique also incurs a cost as additional hardware (crab-cavities) is foreseen to produce the
effect, however this cost is relatively small.
The only direct impact of relocating the undulator-driven positron source to the end of the main
electron linac (WA3) for 500 GeV centre-of-mass running is an increase in the single-bunch RMS
energy spread. The RDR solution (Flux Concentrator, 147 m undulator at 150 GeV) results in an RMS
energy spread of 0.12% at 250 GeV beam energy. Adoption of the more conservative quarter-wave
transformer (QWT) as capture device, requiring an increase in undulator length to 231 m, increase
the energy spread at 250 GeV to 0.14%. With the undulator at the end of the linac (71 m undulator,
QWT), this increases further to 0.2% at 250 GeV - still well within the bandwidth of the BDS.
2.6.2 Operation below 500 GeV Centre-of-mass Energies
During the RDR phase no attempt was made to produce optimised luminosity parameter sets for
centre-of-mass energies below 500 GeV. A basic assumed requirement was a 1/scaling below
500 GeV down to 200 GeV centre-of-mass.
As of writing, there are still no formal optimised parameters for specific intermediate energies,
although possible scaled parameter sets do exist. Generation of possible formal parameter sets for
key centre-of-mass energy scenarios still remains to be done in TD Phase 2.
However, WA3 and WA4 do have immediate implications for lower energy running. Specifically:

For the reduced bunch number parameters at lower centre-of-mass energies, the luminosity
compensating effects from the beam-beam become somewhat weaker, and so the
luminosity reduces faster than 1/

The generation of positrons from the undulator source reduces approximately quadratically
with the beam energy, and rapidly decreases to zero below a threshold at ~125 GeV
(250 GeV centre-of-mass).
14
Recent estimates have suggested an approximate luminosity of 3×1033 cm-2s-1 at Ecm = 250 GeV for
SB2009. Further studies and optimisations are expected in TD Phase 2.
The positron source is currently specified to produce a yield of 1.5 positrons per electron at a beam
energy of 150 GeV. (A yield of 1 is required, and the specifications assume a 50% production
overhead margin.) Above this value, the yield increases rapidly, providing significant overhead and
safety margin. With these design parameters, the positron yield drops to ~1 (no production
overhead) at about 125 GeV. Should the physics demand high-luminosity running below 250 GeV
centre-of-mass, there are several conceptual schemes that could be considered. One such scheme is
to use alternate 2.5 Hz pulses for generating positrons and luminosity respectively. These schemes
require additional hardware at the end of the linac (bypass lines, kickers, dumps etc). Although these
schemes require further detailed design (and cost) studies, there appears to be no showstoppers.
One such possible scheme is outlined in Section 4.3.
Finally, we should note that the proposed auxiliary positron source should provide sufficient
luminosity (few ×1032 cm-2s-1) for calibration of the detector at the Z-pole.
References
[Over1] ILC Minimum Machine Study Proposal, ILC-Report-2009-019, January, 2009, http://ilcedmsdirect.desy.de/ilc-edmsdirect/file.jsp?edmsid=*865085
[Over2] Summary report of the first meeting on Accelerator Design and Integration, ILC-INT-2009-035,
June, 2009, http://ilc-edmsdirect.desy.de/ilc-edmsdirect/file.jsp?edmsid=*879845
[Over3] This is based on the assumption that the choice of gradient is unlikely to change by more
than the order of 10%.
[Over4] In part, as the number of klystrons and modulators required depends on this option.
15
3 Layout
In this section we present an overview of the changes that are considered for the ILC layout. While
some of these changes are consequences of the proposed changes in the subsystem designs and are
integral part of the SB2009 proposal, some have come about through continuing system design and
development prior to the SB2009 re-baseline studies.
Referring to Figure 2.1 in Section 2, the most visible changes are in the central region, specifically the
Damping Rings (DRs). In the Reference Design, the 6.4 km six-fold symmetric rings[Layout1] were
located around the interaction region and elevated by 10 m above the Beam Delivery systems along
with the electron injector and positron KAS (keep alive source). The arrangement allowed some
independence in operation of the injectors and rings from personnel access to the Beam Delivery
Systems (BDS), but at a cost of a relatively complex layout of tunnels and shafts.
When considering the detail component layout of the six-fold ring, a different philosophy was
explored, resulting in a proposal for a two-fold design ‘racetrack geometry’ with long straight
sections to accommodate RF, wigglers as well as injection and extraction. The lattice of the arc
sections was changed to a new very flexible one, where the emittance and momentum compaction
could be separately adjusted. These changes were accepted in 2008 before the SB2009 formal
process was begun. For further detail regarding DR component layouts, see Section 4.5.
The changes in geometry stimulated discussions regarding alternate layouts for the central region.
The racetrack design has the beam injection and extraction occurring at nearly the same point in the
centre locations of one of the two long straight sections; this allows the DR to be located in the same
plane as the BDS, but offset horizontally from it (clear of the detector hall at the IR). Short (in-plane)
transfer line tunnels can then connect the DRs to the main BDS tunnel. Further layout considerations
led to the concept of a compact and efficient central region[Layout2], which incorporated all of the e±
injection systems, DRs, BDS and Interaction Region in a single central campus. The long linac tunnels
containing both the main linac and RTML beamlines would then extend outwards in either direction
without interruptions.
These ideas were combined with the construction scheme of the main linacs on the basis of the
single-tunnel scheme, the low beam power option and single-stage bunch compression into the AD&I
studies and the SB2009 Design.
The single-tunnel linac layout incorporates the proposed single-stage bunch compressor. This singlestage bunch compressor ends with the beams at ~ 5GeV as opposed to 15 GeV in the two-stage
design. The overall design results in a net reduction of some ~300 m in comparable machine length.
A major change in the electron linac is the re-location of entire undulator-based positron source from
the 150 GeV point to the central region. Electron and positron main linacs are now identical, except
for an additional 4.1 GeV energy reach in the electron linac to drive the positron source (the
Reference Design has 3 GeV). This allows changes in accelerating gradient, gradient distribution or
maximum design energy to be easily accommodated. Changes of gradient of the order of 10% would
require little in the way of fundamental changes in the layout or of infrastructure support from CF&S.
Figure 3.2 shows the concept of beamlines which share tunnels in the central region. Figure 3.3
shows a cross section of the positron main linac near the beam delivery section. Accurate to-scale
tunnel schematics can be found in the ILC EDMS and at
http://ilc.kek.jp/SB2009/TUNNEL%20DRAWING%20SET%20(11-20-2009).pdf .
16
Figure 3.2: Topological diagram of the beamlines in the ILC central area.
Figure 3.3: Cross section of the positron main linac near the beam delivery section.
Many of the design details and changes of these systems are described in Section 4. The following is
a discussion of the general layout and the design philosophy.
The SB2009 design goal was to minimize the underground volume, while developing the individual
system designs and interfaces with CF&S beyond where they were at the publication of the RDR.
Since the injectors and beam delivery systems are very varied in their support requirements (power
supplies, RF, instrumentation etc), it was decided at an early stage of the studies that a single support
tunnel would be maintained for the central region. Even with the support tunnel, there is a reduction
of ~5 km out of ~14 km in tunnel length in central region, not including the DR’s.
The reduced beam power option, with half the number of bunches, enables (but does not require) a
half-circumference damping ring. The central campus layout can accommodate either the 3.2 or 6.4
km ring designs, since both utilise the same design for the 1 km straight sections, one of which is
used for injection and extraction for both electron and positron rings are located (the other
accommodates the RF and damping wigglers).
17
Figure 3.4: Schematic Directional Beamline Layout
18
Much of the effort to date has been on the region where the positron production systems and the
beam delivery systems co-exist. In the Reference Design, the positron production system was
incorporated into main electron linac at the nominal 150 GeV location. At the publication of the RDR,
there were only conceptual designs for this system, whereas today the designs now address many
more practical engineering issues. For example, there are short sections of tunnel requiring
“alcoves”, which will require different tunnel cross-sections, the exact cost-optimised design of which
will be dependent on local geology. Many of these issues were duplicated in the “Keep alive source”
(KAS) in the RDR. The KAS has now been replaced by an “Auxiliary Source”, which uses an
independent electron beam in conjunction with the same target, capture, and booster systems as the
gamma beam from the undulator. As of writing, the SB2009 design has only one vault, which is
designated to handle high radiation environment. This region will require further detailed
development in TD Phase 2.
The SCRF Main Linacs have a relative large acceptance in both energy and transverse phase space as
compared to either the undulator source or the BDS. Hence the from one to the other require
protection in the form of collimator and fast beam-abort systems. Because the undulator is located
at the end of the linac leading into the BDS in SB2009, a single section of collimation and abort
system will protect both in an integrated cost-effective fashion. Consolidation of the positron source
into the BDS also moves all the bandwidth limiting beamline sections to the central location, which
will simplify the main linac commissioning and tuning.
The electron injector easily fits into the SB2009 layout, whose geometry choice is primarily
determined by the requirements from the positron injector side.
It is desirable (but not mandatory) that a bunch is injected into (fills) an empty DR bucket that has
just been emptied by extraction. This allows maximum flexibility in the bunch patterns, but also puts
tight constraints on the difference in path lengths travelled by the electrons and positrons through
the machine. The issue was well known at the time of the RDR, but a fully worked-out lattice solution
was not presented, because at that time the path length correction had to be kilometres. In the
SB2009 layout and with the smaller 3.2 km rings, the correction can now be a few hundred metres. A
natural asymmetry in the length of electron and positron sides is exploited to incorporate this delay
drift length.
There are other examples where the SB2009 designs and layouts are more complete and detailed
than the RDR. This is due to both the continuing design work over the time since the RDR but also
driven by the AD&I studies of the SB2009 proposal. In some places this complicates comparisons.
The detail system designs and their pros and cons are based on these layouts and are discussed in
detail in the following sections. Figure 3.4 gives the schematic directional beamline layout.
References
[Layout1] Adjusted from the 6.7 km circumference quoted in the reference design to accommodate
inconsistencies with bunch timing.
[Layout2] Efficient in this context refers to minimizing the necessary underground volume of
tunnels, caverns and shafts.
19
4 SB2009 Proposal
4.1 Issues of Main Linac Accelerating Field
4.1.1 Status of Cavity Field Gradient:
The expected production cavity gradient and yield are two of the most important parameters in the
ML-SCRF technical area to be provided in the evaluation process, as well as in the studies of the
overall construction cost and plans in relationship with the Conventional Facilities & Siting Group and
Accelerator System areas. In 2005, the S0 performance goal for each of the 9-cell cavities to use in
the ILC main linacs was defined as ≥35 MV/m with Q0 ≥ 8 x 109 in low-power vertical acceptance
tests. This number included an estimated 10% for the ensuing assembly steps and operational
margin, so as to ensure an average accelerating gradient of 31.5 MV/m with Q0 ≥ 1 x 1010 in cavities
installed in horizontal cryomodules. The world-wide R&D effort ensued with these goals. This R&D is
currently still ongoing, and it represents the largest fraction of the global ILC R&D effort.
As part of the TD Phase-1 baseline review and subsequent re-baselining effort (including SB2009
proposal), it was agreed to review the choice of accelerating gradient, in the light of present and
prospective outcome from the R&D programs, together with the specific system implementations
and operational models of the main linacs. Test results from approximately sixty 9-cell cavities
processed and vertically RF tested are expected to be included in the TDP-1 cavity gradient
evaluation. One of the critical exercises, therefore, is to catalogue and analyse the production
statistics of the 9-cell cavities and their performance by putting together all the available data from
all three regions in a scientifically consistent fashion.
For this purpose a taskforce was established to create a global database for ILC cavity test results.
The mandate of this taskforce is to:
1.
2.
3.
4.
Collect and verify the global cavity R&D information;
Create and maintain a global database for the ILC-related cavity performance;
Analyse the database to understand the global progress in improving cavity performance;
Assist the Project Managers through analysis of these results and in re-evaluation and projection
of future cavity performance during TDP-II and the future production phase.
The team produced an interim, progress report at the Linear Collider Workshop of the Americas
(ALCPG) and GDE meeting held in Albuquerque, New Mexico in the fall of 2009. Figure 4.1.1 shows
the current (Dec 2009) status of, cavity gradient performance in the first and second passes including
the chemical process and cold vertical tests for the cavities manufactured by vendors and processors
who have currently reached the S0 goal with at least one cavity. In addition, the database effort
allows tracking of the rapid progress made in cavity processing at ANL/FNAL and KEK and cavity
manufacturing at new vendors.
In addition to evaluations of vendor progress based on hard cuts in cavity performance, another
important production statistic is the scatter in production cavity performance, and the effect on the
requirements of the RF systems. The cavity database will allow for a systematic means to complete
this evaluation. During actual machine construction, methods such as sorting or grouping cavities
before installation in a cryomodule may be used to soften the range requirements on the RF system.
20
Electropolished 9-cell cavities
Electropolished 9-cell cavities
JLab/DESY (combined) up-to-second successful test of cavities from qualified vendors - ACCEL+ZANON+AES (25 cavities)
100
100
90
90
80
80
70
70
yield [%]
yield [%]
JLab/DESY first successful test of cavities from qualified vendors - ACCEL+ZANON+AES (30 cavities)
60
50
40
60
50
40
30
30
20
20
10
10
0
0
>10
>15
>20
>25
>30
>35
>40
>10
max gradient [MV/m]
>15
>20
>25
>30
>35
>40
max gradient [MV/m]
Figure 4.1.1: Production yield as a function of gradient after the 1st (left) and 2nd (right) pass process
and vertical test as of December 2009. Note all cavities with a gradient exceeding 35MV/m meet the
ILC specification of Q0 ≥ 8 x 109 at 35MV/m.
4.1.2 Issues of Main Linac System Design
In conjunction with the (GDE and AAP) review process in 2010, based on the current R&D results we
propose to keep the cavity gradient goals at 35MV/m in vertical test,S0, and 31.5MV/m in operation
in an installed cryomodule, S1. We note that as the R&D progresses, including horizontal testing of
individual dressed cavities, and full cryomodules and strings of cryomodules through the ILC S1 and
S2 programs, and efforts at the XFEL, a better understanding of the components of the margin
between the cavity gradient goals will emerge. Ultimately for the ILC accelerator the gradient will be
reviewed based not only on this information, but also on the economics of mass production, and
production yield. Furthermore, as design of the RF systems progresses, full systems analyses of the
integrated power / cavity system, allowing for appropriate variability in cavity gradient and system
operational overhead, will have to be completed.
We propose that TD Phase 2 we define the specific magnitude of the spread (+/- 10 – 20%, pending
more discussion) of operational gradient to support with the HLRF systems. This allows lower
performance cavities to be utilised and increases the production yield, assuming that highperforming ones maintain the average. This approach has impact on both the definition of cavity
production yield and acceptance criteria. Currently every indication is that this will provide a better
cost optimised system, albeit at the expense of a more complex HLRF distribution system and lower
overall RF-to-beam power efficiency. The exact amount of RF power and gradient overhead required
for operational margin remains to be determined. Within the TD Phase, the FLASH beam tests with
high-gradient cryomodules will give us further details.
In preparation for industrial production, we will investigate the assumption of a usable cavity fraction
of 80 % utilized in the RDR, with the possibility that it may be improved to be a level of 90 % including
all effects described above.
Based on the S0, S1, S2 and related R&D programs, the operational field gradient for the ILC machine
will be confirmed in light of the experience and expectations related to margins required for
redundancy, operational margin, dynamic tuning, and availability.
References
[SRF1] C.Ginsburg, presentation at the GDE meeting, Albequerque, October, 2009,
http://ilcagenda.linearcollider.org/contributionDisplay.py?contribId=182&sessionId=29&confId=346
1
21
4.2 Electron Source
4.2.1 Overview:
The proposed changes in the SB2009 design have two notable implications to the electron source
system:

The low-P parameter set increases the bunch spacing by a factor ~2. Increased bunch spacing
will reduce the challenges for the source drive laser system, due to a more complete population
inversion of the Ti:Sapphire laser medium in between pulses. However, the goal of the R&D
remains unchanged and set to satisfy the full RDR specification.

Change in the damping ring design, combined with the revised layout of the central injector
complex, as shown in Figure 4.2.1, will affect geometry of the transport lines, including their
lengths.
By far the most significant modification is the re-location of the source from its own independent
underground housing (tunnel), into the tunnel system shared with the positron and beam delivery
systems. As explained in Sections 2 and 3, this has been motivated primarily by cost considerations,
and a desire to reduce the scope of the underground construction and maximise the use of common
shafts and utilities.
Figure 4.2.1: Topological layout of the positron side of the central region, showing the approximate
location of the electron source (in green).
At the level of subsystem design, however, the ILC SB 2009 results in no fundamental changes to the
polarized electron source. The source parameters are essentially the same. They are summarized in
Table 4.2.1. The fundamentals of the laser system design would also remain essentially the same.
22
Table 4.2.1: Source parameters
Parameter
Electrons per bunch (at gun
exit)
Electrons per bunch (at DR
injection)
Number of bunches
Symbol
ne
Value
4*1010
Unit
Number
Comments
Same as RDR
ne
2*1010
Number
Same as RDR
Ne
Number
bunch repetition rate
F µb
Low-P
parameter; new
baseline
Low-P
parameter
bunch train repetition rate
bunch length at source
Peak current in bunch at source
Energy stability
Polarization
Photocathode Quantum
Efficiency
Drive laser wavelength
F mb
Δt
Iavg
S
Pe
QE
~ 1312
(was ~2625
in RDR)
1.5
(was 3 in
RDR)
5
~ 1ns
3.2
<5
80 (min)
0.5
Hz
ns
A
% rms
%
%
Same as RDR
Same as RDR
Same as RDR
Same as RDR
Same as RDR
Same as RDR
nm
Same as RDR
single bunch laser energy
E
780-810
(tunable)
5
µJ
Same as RDR
Λ
MHz
4.2.2 Design Work to Pursue during TDP2
Although the studies to date have shown that acceptable solutions exist for the shared tunnel layout,
further optimisation is needed in the beamline design during TD Phase 2. For instance, for the
finalised tunnel design, a re-optimisation of the electron source transport lattice is required. One
critical item to note is the control of spin polarisation. Correct bend angles and arc radii for most
transport lines have to be established. The resultant constraints have to be an integral part of the
overall design of the central region, including the Source, RTML and BDS.
Figure 4.2.2: Schematic view of Wien filter proposal
Recently, a new option for the electron spin rotation from longitudinal to transverse has been
discussed, as shown in Figure 4.2.2. It utilizes a Wien filter, introduced immediately downstream of
the electron gun, replacing the two spin rotating superconducting solenoids in the electron to ring
transfer line. This would result in a small cost saving for the electron source. However, additional
space is required immediately after capture solenoid (see Figure 4.2.2), and it is desirable to keep the
total distance between the gun and the sub-harmonic buncher system as short as possible, so as to
23
reduce the growth of longitudinal emittance due to space charge. Further simulations are needed to
verify the feasibility of this option, and this work is expected to continue during TD Phase 2.
4.2.3 SB2009-specific R&D Work to Pursue during TD Phase 2
No additional work is required on the electron source system in addition to the already ongoing riskmitigating R&D such as those in the laser system, high-voltage gun and photocathode development,
and studies into cathode surface charge limits. Design work on layout and lattice optimisation will
continue naturally during TD Phase 2.
24
4.3 Positron Source
4.3.1 Overview
In addition to changes in some aspects of the beam parameters, the SB2009 proposes some major
changes to the system implementation of the undulator-based positron source at the ILC.
Relocation of the undulator: As part of the proposed central region integration (Section 3), the
complete undulator-based positron source is re-located from the 150 GeV energy point to the exit of
the main electron linac, and is integrated with the up-stream part of the BDS. This modification will
benefit several system integration issues:

The undulator source can be better integrated into the upstream area of the BDS, where
more tunnel space and more freedom of lattice design are available than in the main linac.

Machine protection systems for the smaller acceptance undulator and BDS beamlines can be
effectively combined into a single system, located immediately downstream of the main linac.

All sources of restricted energy bandwidth are now localised in the central region, leaving
both the main SCRF linac as systems with nearly identical contiguous high-bandwidth. This
greatly facilitates their commissioning and tuning.

Moving the positron target and capture system to the end of the linac consolidates all the
high-radiation environment systems within the central area, which is expected to be
beneficial for certain host sites (radiation safety and environmental impact).

A large energy overhead is available to drive the undulator source, which allows operational
margin for the early commissioning in the event that the maximum-performance of the main
linacs system is not achieved (i.e. maximum gradient at full beam loading).
An important consequence of the re-location is that the energy of the electron beam passing though
the undulator will vary with the required centre-of-mass operation. In general the positron yield from
the source scales approximately as  down to a beam energy of ~125 GeV, after which the yield
rapidly drops to zero. This clearly has ramifications for the physics related parameters (primarily
luminosity) at lower than 500 GeV centre-of-mass operation. This important issue is discussed more
detail, together with the proposed solution, in the latter part of this section.
Adoption of a simpler capture magnet (Quarter Wave Transformer, QWT): The QWT would be
implemented immediately after the pair-production target. This magnet is technically simpler and
more feasible than the flux concentrator that was previously considered during RDR. Therefore, this
change contributes positively to reduction of the overall technical risk of the system. This is
accomplished at the cost of reduced positron capture, however, and a subsequent increase in the
undulator length is introduced to make up for the loss. Our strategy is to maintain the more
conservative QWT-based scheme as the new baseline, while continuing the R&D effort on the flux
concentrator. The flux concentrator may be adopted again in the future, once a feasible design has
been established. It is noted that this proposal is independent of the location of the positron source.
Removal of an independent Keep Alive source: The positron Keep Alive Source in the RDR was based
on an independent 500MeV electron drive beam, target, capture RF, remote handling, etc., which
necessitated replication of many subsystems. For SB2009, it is proposed to replace the Reference
Design KAS with an Auxiliary Source, which will instead make use of a 500MeV electron drive beam
which shares the same target, capture RF, etc as the undulator based source. The Auxiliary Source
would be used in the commissioning or system diagnostic purposes for the downstream positron
25
systems, such as the positron damping ring, the positron RTML and the positron main linac. The
proposed Auxiliary Source should be capable of producing a few % of the nominal single bunch
charge (as compared to the ~10% specified for the original KAS, but with reduced construction scope,
complexity and cost).
In addition to the above immediate implications for the source, the low beam-power option also
reduced the demands on the target by reducing the average beam power. However, the working
assumption for the R&D is to retain the full Reference Design specification, which will supply further
operational (risk-mitigating) margin to the design, or maintain the option for possible future
upgrades in beam power.
Figure 4.3.1 shows the conceptual diagram of the positron production system and related beamlines.
The pertinent parameters for the SB2009 positron source system are summarized and compared
with the RDR in Table 4.3.1.
Figure 4.3.1: Conceptual layout of the positron source region. The electron beam is travelling from
left to right. Red lines indicate electrons, blue lines indicate positrons and black lines indicate
photons.
Overall, this revised positron system design is expected to offer a reasonably solid conceptual
foundation for detailed engineering study to pursue during TD Phase 2, in accordance with the
overall layout changes as put forth in Section 2 and 3. The impact on the luminosity performance is
discussed later in this subsection.
26
Table 4.3.1: Comparison of the RDR and SB2009 positron source parameters. Note R&D goals for TD
Phase 2 retain the RDR specification.
Parameter
Positrons per bunch at the IP
RDR
2 x 1010
Bunches per pulse
Pulse repetition rate
2625
5
Positron energy (DR Injection)
DR transverse acceptance
DR energy acceptance
Electron drive beam energy
Electron energy loss in undulator
5
0.09
±0.5
150
3
Required additional electron
linac overhead
Undulator period
Undulator strength
Active undulator length
3
Field on axis
Beam aperture
Photon Energy (1st harmonic)
11.5
0.92
147 (210 after
polarisation
upgrade)
0.86
5.85
10
Photon beam power
131
Target material
Target thickness
Target power adsorption
Ti – 6%Al – 4%V
14
8
SB2009
1 to 2 x 1010
(see Figure 4.3.3 for
details)
1312
5 (125 to 250GeV)
2.5 (50 to 125GeV)
5
0.09
±0.5
125 to 250
0.5 to 4.9
(see Figure 4.3.4 for
details)
4.1
11.5
0.92
231 (maximum, not
all used when
>150GeV)
0.86
5.85
1.1 (50 GeV) to
28 (250 GeV)
102 at 150 GeV
(less at all other
energies)
Ti – 6%Al – 4%V
14
8
Units
Hz
GeV
m-rad
%
GeV
GeV
GeV
mm
m
T
mm
MeV
kW
mm
%
4.3.2 Positron Yield and Operational Scheme for Low Energy
The main issue with the SB2009 positron source is the yield variation as a function of electron beam
energy.
Figure 4.3.2 shows how the yield varies for the RDR undulator parameters (0.86 T on-axis field and
11.5 mm period) as the electron energy changes. The design yield is 1.5 (i.e. 1.5 positrons will be
available within the damping ring acceptance at 5 GeV for every electron that passes through the
undulator). This is the number that is achieved at 150 GeV with a 147 m long undulator. In SB2009 an
undulator length of 231 m is required instead of 147 m for the RDR, so as to compensate for the
reduced positron capture with the QWT.
27
Figure 4.3.2: Positron yield versus electron energy in the RDR positron system.
Figure 4.3.2 indicates that as the electron energy increases beyond 150 GeV, the yield increases quite
significantly, reaching a value of ~5 at a beam energy of 250 GeV. In such cases, in practice, some
sections of the undulator would be switched off in order to bring the yield back towards 1.5, but
nonetheless the design parameters offer an inherently large safety margin[Posi1]. On the other hand,
as the electron energy reduces below 150 GeV the yield drops, reaching 0.75 at 125 GeV. At this
energy there will only be 1 x 1010 positrons per bunch at the IP instead of 2 x 1010 as at all energies
above 150 GeV.
The solution proposed for this issue of lower-energy operation is as follows. At beam energies of
125 GeV and below the operation of the ILC will be configured, such that alternate macropulses (this
corresponds to one train of 1312 bunches over 970 ms) will either be transported to the IP or be
used to generate positrons. This will be achieved by accelerating alternate macropulses to different
energies. One macropulse will operate at 150 GeV and will generate positrons as normal. Then the
next macropulse will be accelerated to the low energy required at the IP and will not be used to
generate positrons. With this pulse-switching scheme the number of positrons available at the IP will
be 2 x 1010 per bunch but the macropulse rate will effectively be 2.5 Hz.
The accelerator issues raised by alternate operation of the machine at different beam energies are
currently under study. Differences in beam trajectories at the exit of the Main Linac will certainly
require fast orbit correction at the entrance to the undulator [posi2]. (It is very likely that a fast intratrain feedback system will be implemented at this location in any case.) It is currently assumed that
no such correction is required in the Main Linac itself, however this requires further investigation.
Additionally, a new transport line for the high-energy beam will be required to carry the electrons at
150 GeV from the exit of the undulator to the electron BDS tune-up dump. The number of positrons
at the IP from the SB2009 as a function of electron energy is summarised in Figure 4.3.3.
The electrons will lose energy as they travel through the undulator and emit synchrotron radiation.
This energy loss needs to be compensated for by the upstream main electron linac. Figure 4.3.4
shows how the energy loss varies with the electron energy. At the top-most energy the loss is
28
approximately 4.1 GeV, which must be compensated for by additional linac (an addition of ~1.1 GeV
over the Reference Design solution).
A second consequence of the emission of synchrotron radiation is an increase in the energy spread of
the electron bunch. Figure 4.3.5 illustrates how the relative energy spread changes with electron
energy for the RDR and SB2009 cases. The RDR case assumes a relative energy spread of 1.5% at
15 GeV (at the entrance to the main linac) and the SB2009 case assumes a relative energy spread of
1.08%.
Number of Positrons at the IP
2.5E+10
2E+10
1.5E+10
1E+10
2.5 Hz at the IP
5 Hz at the IP
5E+09
0
50
75
100
125
150
175
200
225
250
Electron Energy (GeV)
Figure 4.3.3: Number of positrons at the IP vs electron energy for the SB2009 scheme. For the RDR
solution, the bunch charge is constant at 2×1010 particles at 5Hz operation for the entire range.
Figure 4.3.4: Energy loss due to emission of SR in the undulator vs electron energy.
29
Figure 4.3.5: Relative energy spread for the electron beam vs electron energy.
4.3.3 Polarized Positrons
Generation of polarised positrons is not part of the baseline solution. However it is a required
upgrade. Nevertheless the RDR baseline generates 34% positron polarisation, which will be of benefit
to the physics programme.
Figure 4.3.6: Positron polarisation vs electron energy for the RDR and SB2009 in the
baseline configuration. Much higher polarisation levels are achievable in both layouts
following a simple upgrade of the positron source.
In the SB2009, polarised positrons will also be generated but the rate of polarisation will vary with
energy, as shown in Figure 4.3.6. The upgrade path to the desired 60% polarisation level is to either 
Install an extra ~200 m of undulator and to then use collimation to remove the photons with the
wrong polarisation characteristics, or

Rely upon the development of the flux concentrator, which increases the yield to the equivalent
of the installation of extra undulator length.
30
The latter scheme of changing the capture magnet for a more efficient device is the presently
assumed option. The exact polarization level that will be achieved with this option has not yet been
evaluated as function of electron energy.
4.3.4 R&D and Design Work to Pursue during TDP2
As with the electron source, the primary baseline R&D remains unchanged by the adoption of the
SB2009 proposal:





Target system R&D
Undulator magnets
Flux concentrator
Remote handling
Engineering integration
The main area of risk for the SB2009 is with the target. The adoption of the QWT increases the length
of the undulator and this enhances the peak photon beam power on the target. Fortunately the
reduction in the number of bunches by a factor of two reduces the average power on the target,
which effectively increases the performance risk margin. Nevertheless, the identified Reference
Design issues still remain to be resolved and the solutions validated. They include: the pressure shock
wave impact to be confirmed as reasonable, the eddy current effect to be simulated (an experiment
is ongoing to confirm this), and the rotating vacuum seals to be confirmed suitable. Another
important SB2009-specific issue that needs careful assessment is the performance of the target
when used as the Auxiliary Source in conjunction with a 500 MeV electron beam.
The performance of the positron source with realistic undulator magnets also remains a subject for
further study during TD Phase 2. Now that two full-length undulators have been constructed and
tested [Posi3], it is possible to evaluate the actual spectral output from these devices. A simulation is
required of a 230 m long undulator with the level of errors measured, to determine how the
trajectory wander and phase error will impact on the positron yield and polarisation output, as well
as the emittance of the electron beam. A beam test with a full-scale undulator cryomodule would be
desirable to check for unexpected issues, such as vessel heating, as well as confirm the photon
spectrum. However, these are general identified issues with the undulator source, and not specific to
SB2009 (although the parameters are modified somewhat).
Since the best route to polarised positrons is through the flux concentrator, this device should
continue to be studied. A feasible solution is still to be generated, although the latest findings are
encouraging. Even better capture efficiency will be possible with a liquid lithium lens system. Studies
of such a device have been carried out and these should be continued, as the rewards of a working
system would be significant.
The remote handling region will be an important part of the positron source. This will contain the
upstream collimator, target, QWT, and capture linacs. The shielding thickness has been assessed and
the present assumption of 1 m appears to be more than adequate providing the correct grade of
concrete is used. The remote handling unit still needs careful design and the operating scenarios
need to be assessed in more detail. This issue is not SB2009 specific.
The engineering integration of the positron source with the BDS and RTML has started, and will
continue during TD Phase 2. The additional dump line (for the 150 GeV electron beam) due to the
pulse-switching mode during low energy operation below 125 GeV, and similarly, a dump line for the
500 MeV electron drive beam for the Auxiliary Source need to be integrated to the current design.
31
Although not specific to SB2009, R&D into some alternative approaches to the positron system
design will also be pursued during TD Phase 2, such as the use of liquid lead target or a hybrid target
system utilizing the photon channelling technique.
References
[Posi1] In reality the maximum yield will be somewhat less than 5, due to limitations with power
handling (losses) in the capture sections and downstream beamlines and collimators.
[Posi2] K. Kubo: Presentation at the GDE Meeting, October, 2009,
http://ilcagenda.linearcollider.org/getFile.py/access?contribId=247&sessionId=31&resId=0&
materialId=slides&confId=3461
[Posi3] J.Rochford, "The Superconducting Undulator for the ILC Positron Source", presented at PAC
2009 as WE2RAI01, May, 2009.
32
4.4 Damping Rings
4.4.1 Overview
This section describes the new design to adopt at the ILC damping ring systems in accordance with
the overall layout changes as put forth in Section 2 and Section 3 of this document. The so-called Low
Power option, with a factor of two reduction in the number of bunches, allows (but does not require)
a corresponding factor of two reduction in the circumference of the damping rings, leading to a
major cost saving. Thus, the SB2009 proposal includes a new damping ring design. However, as
discussed in more details later, from the view point of beam dynamics and most of the hardware
aspects, this change is not expected to cause major performance challenges or technical aggravation,
and should provide a firm conceptual basis for more detailed engineering studies during TDP2.
The proposed new design for the damping rings will have the circumference of 3.2km with a
racetrack shape [DR1]. Following the publication of RDR which assumed 6.7km rings with roughly
hexagonal shapes, in discussion at GDE meeting (TILC08) [DR2,3], a racetrack ring shape was adopted
with 6.4 km circumference and a different arc lattice. The main reason for this choice was tunnel
simplification and reduction of the number of shafts and caverns in order to improve operational
efficiency and reduce costs. The difference between the new 6.4km arc lattice and the RDR one has
been driven by the decision of reducing the rms length of bunches from 9mm to 6mm. This reduction
has major benefits for the bunch compressors downstream of the damping rings. To achieve a
shorter bunch length, for fixed RF voltage and frequency, a lattice with a smaller and flexibly
adjustable momentum compaction factor has been chosen [DR4].
ILC Damping Ring
0
-100
400 m
-200
-300
1000 m
-400
-500
-600
-400
-200
0
200
400
600
800
Figure 4.4.1: Layout of the 3.2km damping rings.
Figure 4.4.1 shows the layout of the 3.2km damping rings. The electron and positron rings are to be
stacked in a common damping ring tunnel located in the central part of the ILC accelerator complex.
This construction is fundamentally the same as that put forth in RDR.
Aside from the reduced circumference the fundamental technical design and implementation of the
damping rings remain the same as or similar to previous designs. For instance, the bunch separation
and the number of particles per bunch remain the same, and the beam current in the ring is the
same. Therefore, we expect similar overall performance from the beam optics or beam dynamics
viewpoint. Table 4.4.1 summarizes the pertinent parameters for the new damping ring design in
comparison with both the RDR and the TILC08 version of the design.
33
Table 4.4.1: Parameter list for the RDR and the TILC08 version of the damping ring compared with
the SB2009 3.2 km ring.
RDR
Circumference (m)
Energy (GeV)
Bunch number
N particles/bunch
Damping time x (ms)
Emittance x (nm)
Emittance y (pm)
Momentum compaction
Energy loss/turn (MeV)
Energy spread
Bunch length (mm)
RF Voltage (MV)
RF frequency (MHz)
B wiggler (T)
Total wiggler length (m)
Number of wigglers
TILC08
SB2009
6695
6476
3238
5
2625
2×1010
25.7
0.51
2
4.2×10-4
8.7
1.3×10-3
9
24
650
1.67
200
80
5
2610
2×1010
21
0.48
2
1.7×10-4
10.3
1.3×10-3
6
21
650
1.6
216
88
5
1305
2×1010
24
0.53
2
1.3×10-4
4.4
1.2×10-3
6
7.5
650
1.6
78
32
The technical work done for the long ring can be easily applied to the short one. The lattice still
needs some optimization work and then it will be used to evaluate the expected beam performance.
4.4.2 Ring Description
The electron and positron ring are arranged one on top of the other with counter-rotating beams.
Injection and extraction for each ring are located in the same straight section. The injection line
entering the electron ring is superimposed on the positron extraction line and vice versa. RF cavities
and wigglers are in the opposite straight section with respect to injection and extraction. The wiggler
straight is located downstream of the RF cavities in order to avoid damage by synchrotron radiation.
The RF cavities for each ring are offset from the centre of the straight so that the cavities for the two
rings are not superimposed on top of each other.
In the SB2009 lattice, the reduction of the energy radiated per turn and of momentum compaction
allows to achieve a shorter bunch length with less than half the number of RF cavities with respect to
the RDR. The number of wiggler magnets is 32 instead of 80, less than half, due to the higher field in
the arc dipoles.
The lattice in the arcs is based on the SuperB arc cells, two very similar adjacent cells, but with
different phase advance (see Figure 4.4.2): one is  and the other ~0.75. By tuning the phase
advance in the second cell, the emittance and momentum compaction of the lattice can be tuned.
The optical functions are shown in Figure 4.4.3 for the RF and wiggler straight section and in Figure
4.4.4 for the injection and extraction straight section. The lattice of both straight sections is made of
the same building blocks as used in the longer racetrack lattice, including phase trombone sections to
adjust the tunes and chicanes for path length tuning. The optical functions for the entire ring are
shown in Figure 4.4.5. The new lattice is still in a preliminary stage of development and requires
further optimisation of the dynamic aperture and evaluation of the effects of magnetic errors and
alignment errors. Based on the experience gained with the present reference lattice, we are
confident that by proper tuning of phase advances and the sextupole distribution, an adequate
34
dynamic aperture for the large injected emittance of the positron beam can be achieved. At the
same time, work is in progress at IHEP Beijing on an alternative lattice design using FODO cells. The
optimal lattice design will be selected similarly to how it was done for the previous longer lattice
[DR3].
Figure 4.4.2: Optical functions of the 3.2km damping ring arc cells.
Figure 4.4.3: Optical functions of the 3.2km damping ring RF and wiggler section
35
Figure 4.4.4: Optical functions of the 3.2km damping ring injection and extraction section
Figure 4.4.5: Optical functions of the 3.2km damping ring
4.4.3 Work to Pursue during TD Phase 2
The main challenges for the damping rings are with the fast kickers, low emittance tuning and
controlling collective effects, in particular electron cloud and fast ion instability. As mentioned
36
earlier, the bunch separation and the number of particles per bunch remain the same as or very
similar to RDR/TILC08. Therefore, the magnitude of technical challenges associated with them would
remain essentially the same. As a consequence, in TD Phase 2 it is considered adequate to
continually pursue the already ongoing technical R&D programs from the TD Phase 1 period into TD
Phase 2, as they are. They include the work being done at dedicated test facilities such as CesrTA and
ATF and at many laboratories involved in the ILC collaboration. An important goal of TD Phase 2 is to
evaluate the performance of the SB2009 damping ring design with respect to all the limiting effects,
on the basis of these experimental and theoretical efforts.
Kickers
All of the kicker specifications are the same for the SB2009 design as for the RDR, except for the
repetition frequency within the pulse, which is 1.5 MHz instead of 3 MHz and, hence, less demanding
for the shorter ring. The required kicker pulse length is given by the bunch separation, 6 ns, for the
nominal parameter set (1300 bunches) which is the same as the RDR. When considering ~2600
bunches as a possible luminosity upgrade, the bunch distance would be 3ns as in the “low charge”
parameter set in the RDR parameter plane. A kicker pulse with rise/fall times shorter than 3 ns has
already been measured at KEK ATF [DR5]. A test of multibunch beam extraction from the ATF
damping ring with 5.6 ns bunch distance, slightly shorter than the nominal ILC bunch distance, has
been performed in October 2009. This extraction scheme with fast kickers will be used for ATF2
operation in 2010.
As for the RDR, injection can be performed by using 22 stripline structures inside the vacuum pipe,
each 30 cm long operating at a voltage of 20 kV, provided as +10 and -10 kV pulses on opposite
electrodes. For the extraction only half the number of kickers is required due to the very small
extracted beam size. In the RDR the tolerance on horizontal beam jitter of the extracted beam is 0.1
σx which requires the extraction kicker amplitude stability to be 0.1% or better. Present R&D plans
aim at this objective. If a tighter limit on the amplitude stability is required, then a further
stabilization method, such as adding a feed-forward system in the turnaround, could be considered.
The R&D activity planned for the RDR is continuing to demonstrate kickers satisfying all the DR
specifications including repetition rate, amplitude stability, field uniformity, strength and operational
reliability.
Low emittance tuning
For the low emittance tuning we do not expect significant differences between the two rings even
though the sensitivity to alignment errors of the new lattice remains to be evaluated. For both rings
the steps to achieve the ultra-low emittance are the same: the definition of the required alignment
precision and stability, the precision and sensitivity of the beam position monitors and of the beam
size monitors, the effectiveness of the tuning algorithms.
Collective effects
Collective effects need to be re-evaluated for the SB2009 design, including the fast ion instability
space charge incoherent tune shifts and intrabeam scattering. We do not expect a big difference
from previous evaluations since these effects depend mainly on the ring currents that are the same
as for the RDR. The shorter bunch length poses more stringent requirements on the vacuum
chamber impedances. First estimates indicate that the nominal operating parameters are below the
thresholds for microwave instabilities [DR6]. Special attention, however, must be paid to the effect of
the electron cloud instability.
37
Electron cloud for 1300 bunches (6 ns bunch spacing)
For the nominal configuration with 1300 bunches and 6ns bunch spacing, electron cloud mitigation
techniques are needed both for the RDR and the SB2009 rings. R&D is in progress at the dedicated
test facility, CesrTA, and at other labs. Results are promising and a range of mitigation methods are
being tested. We have convened a working group to apply the results of the R&D to the DR design.
The findings will be used as input for the ring design that will be chosen for the new baseline. Given
the same current and bunch distance we expect similar or even higher instability threshold for the
shorter ring [DR7].
Electron cloud for 2600 bunches (3ns bunch spacing): luminosity upgrade
A possible luminosity upgrade scenario from the SB2009 baseline consists of 2600 bunches with
2×1010 particles per bunch, i.e. 3 ns bunch spacing and twice the nominal current. We expect the
electron cloud build-up to be more severe with the shorter bunch spacing. Thus this configuration is
much more challenging than the RDR low-p parameter set with double number of bunches which
employs 5200 bunches with 3 ns bunch spacing and 1×1010 particles per bunch, i.e. it has the same
current as the nominal RDR parameter set.
Achieving the performance of the SB2009 ring for the luminosity upgrade will require additional
simulation studies, improved mitigation techniques, a more expensive vacuum design, etc. Further
work on mitigation techniques is needed to significantly increase our level of confidence when
dealing with this parameter set. In the event that effective EC mitigations cannot be devised, a backup option would be to add a second positron damping ring.
At present we are confident that for the nominal parameter set with 6 ns bunch spacing a careful
vacuum chamber design with the proper mitigation techniques can assure the required damping ring
performance. For the luminosity upgrade, i.e. 3 ns bunch spacing, the EC working group will provide
evaluation of the instability thresholds and recommendations on mitigation techniques during TD
Phase 2.
References
[DR1] M.E.Biagini, “First Look at a 3 km Damping Ring”, LCWS08 ,November 16-20, 2008,
UIC Chicago,
http://ilcagenda.linearcollider.org/materialDisplay.py?contribId=516&sessionId=11&materia
lId=slides&confId=2628
[DR2] A. Wolski, Specifications for the ILC Damping Rings EDR Baseline Lattice,
https://wiki.lepp.cornell.edu/ilc/pub/Public/DampingRings/LatEvalPage/EDRLatticeSpecifica
tions.pdf
[DR3] M. Palmer, “ILC Damping Rings Lattices Evaluation”, GDE Meeting TILC08, 3-6 March
2008, Sendai, Japan
http://ilcagenda.linearcollider.org/getFile.py/access?contribId=187&sessionId=63&resId=2&
materialId=paper&confId=2432
[DR4] A. Wolski, M. Korostelev, “ILC Damping Rings Lattice DCO”, GDE Meeting TILC08, 3-6
March 2008, Sendai, Japan
http://ilcagenda.linearcollider.org/materialDisplay.py?contribId=185&sessionId=63&materia
lId=slides&confId=2432
[DR5] T. Naito, “Beam Test of the Strip-Line Kicker at KEK-ATF”, PAC09, Vancouver 4-8 May
2009
38
[DR6] G. Stupakov, “Status and future plans for instability studies for the ILC damping
rings,” presented at ILCDR07-KEK (December 2007).
https://wiki.lepp.cornell.edu/ilc/pub/Public/DampingRings/KEKWorkshopTalks/Stupakov.pd
f
[DR7] M. Pivi, presentation at LCWA09.
http://ilcagenda.linearcollider.org/sessionDisplay.py?sessionId=27&slotId=0&confId=3461
39
4.5 Bunch Compressors
4.5.1 Overview
This section describes the proposed changes in the bunch compressor systems (also called, the ringto-main-linac - RTML systems). Major modifications proposed for the RTML are:

Adoption of the single-stage bunch compressor scheme, removal of one (per side) of the 220 kW
dump and associated components, and removal of one section (per side) of beam diagnostics.

Redesign of the second extraction line downstream of the bunch compressor to accommodate
larger energy spread.

Redesign of the RTML lattice in the central injector complex, associated with new layouts of the
damping rings, particle sources and the beam delivery systems.
Replacement of the previous two-stage bunch compressor (BC) with a single-stage BC is motivated by
a significant cost saving and a desire to simplify the system. The current design facilitates a
considerable reduction in the overall beamline length by ~300 m. While this change limits the
available compression ratio to 1/20, it is considered adequate in the light of the bunch length of
6 mm from the exit of damping rings and the nominal bunch length of 300 m, as required at the
beam interaction point (IP). Beam dynamics issues (emittance preservation) already identified for the
Reference Design two-stage system continue to be focus of study for the proposed single-stage
system in TD Phase 2, although the new system is expected to be perform better in these respects.
Removal of one (per side) of the 220 kW dump and associated dump line components (located
downstream of the 2nd compressor in the RDR design) offers another set of cost savings and overall
system simplification.
The remaining extraction line and 220 kW dump, located after the new single stage compressor has
to be redesigned, allowing a large enough energy acceptance for tuning and emergency beam-abort.
Cost saving is also possible by shortening of the diagnostics and matching sections (5 GeV instead of
15 GeV in baseline design).
Proposed changes in the damping rings and the e+/e- sources will affect the design and length of the
RTML line from DR tunnel to main tunnel in non-significant ways.
Figure 4.5.1 shows a revised schematic diagram of RTML, indicating the various elements described
in the text.
4.5.2 Single-Stage Bunch Compressor
The baseline (RDR) design included a two-stage compressor, facilitating an overall maximum bunch
compression ratio of a factor of ~45. The main arguments in support of a two-stage compressor have
been support of the parameter plane (flexibility): assuming the RDR 9 mm damping bunch length, the
two-stage compressor system can achieve bunch lengths of 200 m (RDR low-P parameter set).
However, with the adoption of a damping ring lattice capable of achieving a 6 mm bunch length, it is
now possible to reconsider the possibility of a single-stage compressor with an overall reduction in
compression ratio. Figure 4.5.1 compares the geometry of the RDR two-stage system with a possible
single-stage system that is capable of a factor of 20 compression. The compression factor 20 is
sufficient for achieving the nominal bunch length of 300 m at the interaction point. This bunch
compressor will not support bunch length of 200 m.
40
Figure 4.5.1: Comparison of the geometry of the RDR two-stage system (top) with a possible singlestage system (bottom) capable of a factor of 20 compression.
The bunch exits the single-stage compressor with energy of 4.37 GeV after which it is accelerated in
the Main Linac, which must now accommodate an additional 10.63 GeV of acceleration. For the sake
of cost-comparisons, however, all components up to 15 GeV are included as part of the new singlestage compressor. The single-stage compressor system results in an overall comparable length
reduction of 314 meters of beamline and tunnel. The numbers of components in the two-stage
compressor (BC1 and BC2) and the single-stage (including the Main Linac up to 15 GeV) are
presented on Table 4.5.1.
Table 4.5.1: Beamline lengths and component counts in the two-stage and single-stage bunch
compressor systems.
Length, m
RF units/klystrons
Cryomodules
Cavities
Quadrupoles
BPMs
BC1+BC2
1114
16/17
48
414
88
84
BC1S+ML to 15 GeV
800
14
42
360
61
59
The parameters of the RF section and wigglers have been optimized to match the bunch length and
energy spread requirements. The cost advantages of the single-stage system are:

Reduction in beamline and associated tunnel length (314 meters)

Removal of the second 220 kW/15 GeV beam dump and extraction line components

Removal of one section of the beam diagnostics
4.5.3 New RTML Lattice in the Central Area
In the SB2009 rebaselining proposal, the damping ring circumference has been reduced to 3.2 km.
The RDR design foresaw the ring extraction point to be located at about 1 km from the central region
in the direction of the turnaround. Now, the DR extraction is expected to take place at about 100
41
meters from the central area. This change necessitates a complete redesign of the beamlines in this
area, affecting the required number of horizontal and vertical doglegs. Fortunately this change
simplifies the overall layout of the central area. Only two doglegs are used as shown in Fig. 4.5.2.
Lattice files includes extraction line, in the vertical dogleg, skew correction section beam diagnostics
(emittance measurement station and beam profile monitor) and the collimation section. The new
layout is not expected to affect performance risk in terms of low emittance transport. Continued
error analysis and development of beam tuning techniques are planned during TD Phase 2.
Figure 4.5.2: Central Area beamline footprint: the left hand plot shows the view from top. The slope
of the central straight section is determined by the spin rotator at injection. The right hand plot
shows the vertical dogleg along the central straight section.
4.5.4 Proposed Relevant Studies
Studies of single-stage bunch compressor will include the following packages:

Refined lattice design of a single-stage bunch compressor, diagnostics section and matching
section.

Beam physics simulation to study effect of coupler RF kick, alignment and phase/amplitude
stability of the RF system and provide requirements. The goal to demonstrate that RTML
emittance budget can be achieved and beam parameters at the exit of RTML system provide
acceptable emittance budget in Main Linac. This point is not SB2009 specific.

Re-design and evaluation of the extraction line and 220 kW dump with a suitably large energy
acceptance after compressor. This point is not SB2009 specific.

Development of CAD models and cost estimations for single-stage bunch compressors: wiggler
type design and ultra-short chicane type design. This point is not SB2009 specific.

Experimental studies of amplitude and phase stability, required for single-stage bunch
compressor at FLASH/DESY facility (so-called 9 mA studies). This point is not SB2009 specific.

Re-design of the RTML section from DR tunnel to ML tunnel. This task can be completed after
more detailed configuration of other area systems: DR and sources (FNAL).
It is the intention of the RTML area group to continue these efforts during TD Phase 2.
42
4.6 Main Linacs
4.6.1 Overview:
This section describes the proposed changes to the technical implementation of the main linac
systems, as associated with adaptation of the single-tunnel scheme.
At this moment two schemes are considered for RF power generation and distributions onto
individual cavities. One is called the "Klystron Cluster System" (KCS) in which all the equipment to
generate the L-band RF power is situated on surface buildings rather than in the underground
tunnels. RF waveguides are implemented to connect the power sources outside the tunnels and the
cavities inside the tunnels. The other is called the "Distributed RF System" (DRFS) in which all the
equipment to generate the RF power is situated inside the underground tunnel, together with the RF
waveguides.
In cases of both the KCS and DRFS schemes, the technical implementation of the cavity packages,
including those of couplers and tuners, within the cryomodules, would remain common. Likewise,
the segmentation of the cryomodules, cryomodule strings, and cryogenic systems remain
fundamentally common. However, there will be certain differences in ways that the low-level RF
systems (LLRF) operate, since the numbers of cavities driven under one RF power source group
would be different.
Key features of the two systems are summarized in Table 4.6.1.1.
Table 4.6.1.1: Summary of key features of KCS and DRFS which are considered for the proposed new
main linac RF power system implementation.
Klystron Cluster System
Distributed RF System
Klystrons / Modulators
RDR-like 10MW klystrons +
modulators on surface
Smaller ~750kW klystrons +
modulators in tunnel
Surface Buildings
Surface building & shafts every
~2.4 km, each housing 2
clusters of 19 klystrons
-
RF power delivery
From surface building into the
tunnels via circular TE01
waveguide (0.48m ); After
power splitting at circular tapoffs (CTO), RDR-like power
distribution with revised design.
Waveguide system inside the
tunnel local to each of the
klystrons and the cavities
associated with them.
Cavity / Klystron
population ratio
832 cavities to be driven by RF
power combined from 18 units
of 10MW klystrons.
4 cavities to be driven by one
unit of ~750kW klystron
Both the KCS and DRFS are considered to have good enough reasons to claim their technical
feasibility and worth for R&D into the TDP2 period. Both systems seem capable of fitting into the
43
single main linacs of an inner diameter of ~5m, as will be discussed in Section 4.9. In the following,
we discuss the specifics of the technical implementation, known issues and plans for their
development and system validation to pursue in TDP2.
4.6.2 Klystron Cluster System (KCS)
Concept of Klystron Cluster System
The Klystron Cluster System, or KCS, is one of two conceptual schemes under consideration for
powering a single-tunnel main linac layout, the other being DRFS. In KCS, the rf power is generated
on the surface and transported down to the linac tunnel. To minimize the number of additional
shafts required, multiple sources are clustered together and their power combined into a single,
overmoded waveguide (referred to as the main waveguide). This power is then tapped off in equal
portions at intervals along the linac equal to the RDR klystron spacing (see Fig. 4.6.2.1). In addition to
eliminating the need for a service tunnel, this scheme brings much of the heat load associated with rf
generation to the surface.
Figure 4.6.2.1: Basic schematic layout of the Klystron Cluster System (KCS) with inset of tunnel crosssection. High-power rf is produced in surface buildings. Power from many sources is combined into
an overmoded waveguide and brought down to the linac. There it is evenly distributed through
graduated tap-offs at 38 m intervals along lengths exceeding a kilometer. Two systems running in
opposite directions can share one building and shaft.
An attractiveness of the KCS comes from the fact that it allows relatively long distances, compared to
the Reference Design, between the power sources and the accelerator. In the RDR, the waveguide
length between the closest cavity and its power source was around 7 m (25 m for the farthest),
whereas for the KCS the farthest distance can be up to 1300 m. This is made possible because of the
low loss transmission characteristic of TE01 waveguide. In the KCS, the total system waveguide
power loss is only roughly doubled compared to the Reference Design. Very high power TE01 mode
systems function well in accelerator applications, such as those in the commonly used ‘SLED’ system.
Large-scale X-band transmission systems have been also developed on the basis of circular
waveguides, demonstrating their high-power applications with significantly smaller losses compared
to those based on rectangular waveguides. As a result, a large collection of design and fabrication
44
techniques exist for development of related components such as high-vacuum mode-converters,
bends, pump-outs etc which can be readily applied to KCS.
Design tools such as the simulation program ‘HFSS’ have been, by now, thoroughly validated. With
this and on the basis of existing computational and experimental data, the design criteria have been
established for the maximum peak electrical and magnetic field within each of the components, so as
to stay within a comfortable operational regime in the light of surface breakdowns and pulse heating
limits in varying conditions, such as the choice of the materials, pulse lengths, RF frequencies and gas
(or vacuum) pressures. For the KCS described here, we believe we are within a five to ten safety
margin. For instance, the peak surface electric field in the tap-off power coupler component is about
1.5 MV/m, while the expected breakdown threshold at the nominal pulse length is above 10 MV/m.
A similar margin is achieved with the pulse-heating damage threshold, where the maximum pulsed
temperature rise of about 3 degrees C is an order magnitude below levels associated with
breakdown in X-band cavities.
Another very attractive aspect of the KCS is the intrinsic simplicity of the accelerator tunnel that it
realizes. It contains no modulators and klystrons, and, therefore, substantially reduced water cooling,
compared to the Reference Design. This makes the KCS ideal for relatively open site topography with
rock-based geology, as in two of the RDR sample sites. The long waveguides allows a certain degree
of flexibilities on the locations to build the KCS cluster surface buildings, relative to the ML tunnels,
so as to accommodate site-specific constraints.
System Description and System Components
Power sources in the surface buildings:
Surface buildings containing power supplies, modulators, LLRF, rf drive, and klystrons are situated
roughly 2.4 km apart along each linac. The source specifications are essentially unchanged by this
scheme, except for a 38% increase in pulse duration to accommodate the increased cavity fill time
due to the lower linac beam current. The modulators, either of the TESLA design described in the
RDR or the more recently developed Marx design, provide 120 kV pulses. The klystrons, presumably
solenoid focused, six-beam MBK’s, (though possibly sheet-beam klystrons) are assumed to generate
at a 5 Hz repetition rate the nominal 10 MW, 1.3 GHz pulse with a 2.16 ms flat top. This 38%
increase over the RDR pulse width is necessary due to the longer cavity fill time with the reduced
beam current.
Waveguides into the tunnels:
Dual outputs from each klystron are coupled symmetrically into the main waveguide through
circulators and vacuum windows. Although loads on the circulators should largely absorb any
reflected power, waveguide valves must be included in each of these connecting arms to allow safe
replacement of a failed klystron during operation.
For power handling considerations, the main waveguide is evacuated and operated in the circular
TE01 mode, which has no electric fields on the wall. To couple into this overmoded circular
waveguide, an in-line component referred to as a coaxial tap-off (CTO) is used. Efficient coupling to
the circular waveguide requires that the power fed into the CTO’s two input ports have equal
amplitude and phase of the appropriate values compared to what’s already flowing in the waveguide
from the upstream klystrons. As the power builds up along the main waveguide, the CTO’s have
progressively smaller coupling ratios.
Two 90° bends are required to bring the rf power from a klystron cluster down to and along the
accelerator tunnel. Bending is non-trivial in overmoded waveguide. These must carry the maximum
45
power and preserve mode purity at the end. A design solution might use corrugated circular
waveguide in a carefully profiled bend (a prototype exists at X-band). Designing and demonstrating a
very high power TE01 bend is part of KCS R&D program envisioned in the next few years.
For the main waveguide, a diameter of 0.48 m is a good choice as it is below the cutoff diameter of
TE02 and the attenuation is modest, about 13% over 1.25 km in aluminum (or about 7% on average).
Due to the danger of transferring power into the 20 parasitic modes that propagate at this diameter,
a fairly tight tolerance (~1 mm) must be kept on the cross-section. Matched step tapers are used to
connect it to the 0.35 m diameter circular ports of the CTO’s.
The main impact on the tunnel cross-section (inset in Fig. 4.6.2.1) is the presence of this main
waveguide with insulated cooling, mounted near the ceiling. It is interrupted periodically by vacuum
pump-out sections with associated pumps and by CTO’s connecting to the local power distribution
waveguide systems. Additional hardware formerly located in the RDR service tunnel, such as beam
and rf diagnostic electronics, will also need to be fit in the main tunnel, with appropriate shielding.
This includes power supplies and electronics for magnets, cavity coupler actuators, BPM’s, piezos,
and cavity tuners, as well as LLRF front-end modules for down-mixing signals and relaying them to
the surface. Perhaps ¾ of the racks from the RDR, with associated power and cooling, will be
installed, preferably under the cryomodule supports. This equipment will have to be radiation
shielded with several centimeters of concrete and lead.
At intervals of about 38 m, equivalent to the three cryomodule rf unit length in the RDR, high-power
rf is tapped off from the main waveguide through a CTO. The linac CTO’s are the same as those used
for combining, but oriented oppositely as tap-offs and sequenced in order of increased coupling as
the power is depleted. Thus the output ports of the CTO’s replace the feeds from local klystrons in
the RDR design.
The CTO interior (see Fig. 4.6.2.2) has a diameter step up which couples nearly half the power into
the TE02 mode. A tube is then introduced at the original diameter to divide the volume between inner
circular waveguide and an outer coaxial region. The mix of circular modes causes a radial beating in
field intensity so the location of this tube relative to the step up determines the power split between
the regions, and hence the coupling. The coaxial guide is shorted and coupled through eight
azimuthally distributed slots to a wrap-around waveguide. The outer wall of the latter opens into two
WR650 ports. Here rf windows terminate the vacuum envelope.
Pump-out ports are easily incorporated at the CTO ports, just before the windows. Further pumping
through special pump-out inserts will be required directly on the main waveguide. Based on
experience with the overmoded SLED II system at SLAC, having a port with ion pumps every 10 m
should reduce the main waveguide pressure to below 10-6 torr initially and down to 10-8 torr over
time. A pressurized line would be preferable, but it is not clear that the expected peak field levels,
approaching 4 MV/m, could be sustained in nitrogen at one or two bars absolute pressure. As part of
the R&D program, operation under both vacuum and pressure will be evaluated.
46
Figure 4.6.2.2: TE01-mode Coaxial Tap-Off (CTO) with wrap-around power extraction. Only the gap
and matching ridge vary for different couplings (3 dB pictured). For combining, it is used in reverse.
Power distribution onto cavities:
From the CTO, power is fanned out through a local power distribution system to 26 cavities. Whereas
the RDR had a fixed coupling hybrid for each cavity feed, the acceptance of a relatively large spread
in cavity gradient capabilities (perhaps 24.5~38.5 MV/m) adopted to increase yield calls for the use of
adjustable coupling to tailor the local power distribution. The Variable Tap-Off (VTO) developed at
SLAC, or its equivalent, would be used to optimize the relative power delivered to individual or pairs
of cavities. Pairing would be based on similar gradient rating from pre-installation tests. Pair feeding
through a hybrid or magic-T will allow reflected power to be directed into a load, eliminating the
need for circulators. This savings should roughly balance the VTO cost. Also, the three-stub tuners of
the RDR will be replaced by phase shifters and tunable power coupler antennae, decoupling and
simplifying phase and QL adjustment.
If each cluster combines 17 ten megawatt klystrons and has a 93% average transport efficiency, then
it can power 32 rf units (1.2 km long) with 4.9 MW each (for half current operation relative to the
RDR). With the KCS, a klystron failure does not bring down an rf unit as it does in the RDR, but rather
reduces power uniformly along the fed linac section. However, due to the nature of the combining
circuit, the available power goes as the square of the fraction that are operational. For example, one
klystron failure out of 17 results in 150.6 MW in the main waveguide and 9.4 MW misdirected into
loads. For improved reliability, each cluster would include 19 klystrons with one left off as a reserve.
If one fails, the spare would be turned on, maintaining the full power of (18/19)21910 MW = 170.5
MW. A repair/replacement can presumably be made before a second failure occurs in a given cluster.
Component counts:
Two clusters collocated in a single rf power building would power about 2.4 km of linac, one feeding
upstream and the other downstream. Being on the surface allows easy air cooling of the rf sources,
and two or three floor levels could used to minimize the overall building footprint. With slight
adjustments to the current shaft locations and two additional small diameter shafts per linac, the
main linacs can each be powered through five shafts by nine KCS systems.
47
Table 4.6.2.1: Nominal parameters for the klystron cluster rf power distribution system.
Numbers in parentheses are for the upgrade 9 mA beam current.
# of KCS per main linac
# of rf units per system
# of cryomodules per system
# of cavities per system
# of klystrons/modulators per system
peak rf power per system (MW)
9
32
96
832
19 (36)
170 (340)
RF controls:
The beam-to-rf timing will vary by 0.9 microsecond for the downstream run and 8.8 microseconds for
the upstream run. Centered, the latter represents ± 0.4% of nominal cavity fill time. Flat gradient
along the bunch train can be maintained locally by using QL to tailor the cavity fill times accordingly.
With 18 klystrons serving as a single source, the low level rf system will have to work with vector
sums from 832 cavities, rather than 26, in controlling each cluster. These sums would be computed
locally for each cryomodule and then sent in real time to the cluster building, where a redundant
summing and processing system (2 of 3 majority logic) would be used to adjust the klystron drive
pulses
The following LLRF (low level rf) configuration would ensure high availability at low cost for the
Klystron Cluster System. For each cryomodule, the probe signals from the 8 or 9 cavities are each
down-mixed to baseband (I and Q) and digitized at 10 MHz in a ‘CM Sum Unit’ located near the
cryomodule in a radiation shielded enclosure. After calibration and filtering are applied, a digital sum
of the cavity voltages is computed in the CM Sum Unit and sent via a fiber optic cable to the
associated surface building at a rate of 1 MHz for feedback purposes. The raw digitized data for the
probe rf and other signals, such as reflected rf, are sent between pulses for diagnostics purposes.
If a CM Sum Unit fails to send this info, the cavities in that cryomodule would be detuned and the
power to them would be diverted to a load (both would take on the order of minutes). Note that
there would be some failures, such as of an individual digitizer, for which the cryomodule could
continue to operate. At the surface building, the fiber optic signal from each of the ~96 cryomodules
would be split three ways and sent to three identical Drive Processors.
The output rf for each of the klystrons would also be digitized at 1 MHz and split three ways for the
Drive Processors. The electronics to do this would be separate for each source, and if it failed, at
worst that klystron would be shut off. Others signals, such as the klystron reflected power and power
launched in the TE01 main waveguide, would be also digitized and used for monitoring and for slow
feed forward control. The system could run without these on the time scale in which they could be
repaired.
Each Drive Processor would use the CM sum signals, the klystron output signals and dynamic FF
tables to compute the digital drive signal to each of the klystrons. There would be one digital output
per klystron. At each klystron, the three digital signals from the Drive Processors would go through 2of-3 majority logic before being used to up-mix an LO source that gets amplified (to 100 W) to drive
the klystron.
With this logic, if any one of the Drive Processors fails (hard or soft), the results of the other two are
used. The clock signals and LO for this system would be provided by redundant sources so that if one
48
source fails, the signals would still provided by the others. Also, the AC power to the Drive Processors
would be provided through different breakers. The fiber optic splitters could perhaps have battery
backup. With this scheme, one may have to turn off individual cryomodules or klystrons due to LLRF
failures, just as in the baseline design, but rarely the entire cluster (e.g., in the case when all three of
the Drive Processors disagree for more than one klystron).
Heat load issues:
The additional heat load presented by the main waveguide varies along the tunnel as the power is
depleted from ~230 W/m to ~6.3 W/m, averaging ~114 W/m (for 170 MW-2.2 ms-5 Hz operation).
This does not include the heat load of the local distribution systems. Presumably, the main
waveguide will be cooled via water tubing attached to the pipe inside an insulating wrap. (The
cooling should be sufficient to handle the 45% increase for a 9 mA beam current upgrade.) To absorb
longitudinal waveguide thermal expansion, alignment-maintaining bellows must be included at
flange joints. Five 5 mm gaps (virtually invisible to the TE01 mode) per 38 m rf unit at room
temperature could absorb up to 50°F temperature change. Phase stability will nevertheless be at the
level of 10°/°F at the farthest point (1.2 km) due to diameter expansion. This can be tracked and
compensated by the local phase shifters (the circular waveguide water cooling system should
minimize the slow temperature variations to below 1°F).
For a high current ILC upgrade, the number of klystrons, modulators, and associated hardware per
KCS would be increased from 19 to 36, with the cluster buildings and their electrical and cooling
systems expanded accordingly. The tunnel layout would be unaffected by the upgrade in beam
power, always distributing the available power evenly among the rf units. In particular, the cooling
system in the tunnel would be initially sized for the high current operation.
Cost Considerations
The cost benefit of Klystron Cluster System is primarily realized in the elimination of the service
tunnel. A secondary cost benefit comes from the simplification of the electrical and cooling systems
that have been moved to the surface. The fiscal impact has been estimated to be about a 300 M$
savings. It does, however, entail added complications and risks beyond those of just combining the
service and beam tunnels in the RDR. These include:





Four more 4-m-diameter shafts are needed (two per linac), and the surface presence of the
machine is increased.
Efficiently combining tens of rf sources efficiently into overmoded waveguide is a challenge.
The waveguide bends may be difficult to build.
Although the surface electric fields in the KCS system are relatively low, if there is a breakdown,
the high stored energy (1.4 kJ maximum, based on the round-trip time to shutoff the rf sources
at upgrade beam current) may cause surface damage. However, waveguide arcs are usually
highly reflective.
The LLRF control architecture has to be more robust and the system provides less local control
of the energy variation along the linac
Operational and Availability Issues
The impact of the rf system on machine availability is considered negligible. Redundancy will be built
in to the low level rf, the high power systems contain spare sources, and the distribution waveguide
is assumed to be robust against breakdown.
49
R&D Plan and Schedule
As a first step toward demonstrating feasibility of the KCS, a 10 m run of the main overmoded
distribution waveguide and two CTO’s are being built. These components will first be configured to
measure the transmission through the system and the power handling capability of the wrap-around
waveguide and coaxial section. They will also be high power tested by resonating the line with a
short at one end and a CTO at the other end that is terminated by a shorted waveguide segment to
give the desired coupling. With input power of about a megawatt, standing wave peak fields
equivalent to those of a 330 MW travelling wave will be produced. This test will be done both with
the system under vacuum and filled with nitrogen at 2 bar absolute pressure. Future tests should
include demonstrating the tap-off and combining functions of the CTO, developing and testing a
bend, and perhaps building a 200 m resonant ring.
These issues are also addressed during TDP2.
50
4.6.3 Distributed RF System (DRFS)
Concept of Distributed RF System
The Distributed RF System (DRFS) is another possibility for a cost-effective solution in support of a
single Main Linac tunnel design. In the proposed baseline implementation of DRFS, one unit of 750
kW Modulating Anode (MA) klystron would drive four cavities. In case of the high current option, two
cavities will be driven by a 750kW klystron. Figure 4.6.3.1 shows the conceptual diagram of the DRFS.
The DRFS is based on much simpler and more compact HLRF and LLRF units than the RDR baseline or
the KCS, and it offers a good operational flexibility in coping with performance variations of individual
cavities. This section discusses the technical implementation scheme for the DRFS that have been
developed so far. Because of the increased number of components involved, it is recognized that the
total system cost and the total system reliability are potential concerns, although the unit cost of the
components is expected to be low and the reliability of individual components at their lower RF
power expected to be reasonably good. These issues have been examined, and will be discussed.
Finally, the ongoing R&D efforts on the DRFS and the system prototyping work is discussed, together
with the near-future development plans.
Figure 4.6.3.1 : Schematics of the DRFS concept.
System Description
In the RDR scheme, three units of ILC cryomodules, containing 26 cavities in total, are driven by the
RF power from one unit of 10MW L-band klystron. In the proposed new baseline scheme of DRFS,
four cavities are driven by one unit of 750kW L-band MA klystron. Therefore, one would see that
three cryomodules with 26 cavities will be driven by six and a half units of MA klystrons. In a practical
implementation, the proposed scheme of DRFS is to use 13 units of MA klystrons to drive six
cryomodules, containing 52 cavities.
The DC power and anode modulation for a group of 13 units of klystrons are provided by one
common DC power supply and one common anode modulator (MA modulator). In order to realize
high reliability, each of the DC power supplies and MA modulators is associated with one backup unit,
which will be designed and implemented to be "hot-swappable". Each of the power and voltage
distribution circuits will have a high-voltage SW, which switches off the line when over current
failures are detected.
51
A DC power supply has a bouncer circuit which compensates for the dropping of the pulse flat top.
This can be accomplished by a relatively small capacitor bank. The charger of a DC power supply
comprises of a bundle of several units of identical switching PS. This allows us to increase its
electrical power with ease, simply by adding more switching PS units, in case a doubled DC power is
needed, when the high-current option is deployed. A common heater power supply feeds the
klystron filament power similarly. Another measure toward high reliability is the use of permanent
focusing magnets for MA klystrons. This eliminates the power supply for klystron solenoids and
associated sources of failures and electricity demand.
Over the past 20 years KEK has an operational experience of pulsed klystrons with the cathode
loading of 6.2A/cm2 and the maximum gun field of 22kV/mm with the gun voltage of 304 kV. While
the MA klystrons would operate with pulses longer (2ms) than typical pulsed klystrons at KEK, its
design parameters are chosen to be similar with those of a DC klystron: the cathode loading is less
than 2A/cm2 and the maximum gun field gradient is 6.54kV/mm with the gun voltage of 64kV (the
efficiency is 60%). On this basis we estimate the MTBF of the MA klystron to be 120khrs.
The Power Distribution System (PDS) is relatively simple; PDS components include: directional
couplers (DC), flexible wave guides, fixed phase shifter and three magic tee type power dividers. The
RF power from a MA klystron is divided into four feeds through two magic tees and brought to four
cavities. Circulators are not necessary, since the reflected powers from two cavities are not reflected
to the klystron, since they cancelled at the magic tees which offer a high degree of isolation. A
dummy load which is shunted to magic tees will absorb the reflected power from the cavities. An
average length of wave guide for each klystron is 2 m. The corresponding average loss is estimated to
be 0.8 kW. The specifications of these components are tabulated in Table 4.6.3.1.
In the current study the DRFS design is prepared for variation of cavity accelerating field in the range
of 25MV/m - 38MV/m, i.e. +/-205 around the nominal 31.5MV/m. This means the required feed
power in the range of 116 kW - 147 kW. The DRFS can accommodate suitable sorted cavity pairs, in
accordance with their field gradient performance, by introducing the semi-tunable power divider as
used at the STF at KEK [DRFS1]. For instance in an extreme case of a pair of 38MV/m cavities and a
pair of 25MV/m cavities to be driven by a single MA klystron, the required power would be 588kW.
By including a 20% margin for LLRF, the required RF power for this 4-cavity set would be 710kW,
which is well within the specifications for the MA klystron. The external Q's are matched by the
tuners, and the flattop and stable operation would be maintained by the LLRF control.
In case of a high-current (9mA) operation which is envisaged as a future upgrade possibility, the
population of the MA klystrons will be doubled, thereby each MA klystron feeds the RF power onto
two cavities, or conversely, 26 units of MA klystrons drive six cryomodules with 52 cavities.
52
Table 4.6.3.1: Pertinent specification parameters for the DRFS.
Klystron
Frequency
Peak Power
Average Power Output
RF pulse width
Repitition Rate
Efficiency
Saturated Gain
Cathode voltage
Cathode current
Perveance(Beam@64.1kV)
(Gun@56kV)
Life Time
# in 3 cryomodule
Focusing
Type of Klystron
DC Power supply per 6 cryomodules
# of klystron (6 cryomodule)
Max Voltage
Peak Pulse Current
Average Current
Output Power
Pulse width
Repitition Rate
Voltage Sag
Bouncer Circuit
Capacitor
Capacitance
Inductance
M. Anode Modulator
Anode Voltage
Anode Bias Voltage
1.3
750
7.5
2
5
60
GHz
kW
kW
ms
Hz
%
64.1 kV
19.5 A
1.2 Perv
1.56 Perv
110,000 hours
6.5
Permanent magnet focusing
Modulated Anode Type
13
71.5
253
2.53
181
2.2
5
<1
kV
A
A
kW
ms
Hz
%
26 F
260 F
4.9 mH
56 kV
-2 kV
A notably advantageous feature of the DRFS is with the LLRF control. The number of cavities to be
involved in one unit of vector sum control is only four (will be two in case of high-current option), as
opposed to 26 with RDR or much more with KCS. With relatively unsophisticated sorting of the
cavities in accordance with their performance, a highly efficient operation of these SC cavities is
expected to achieve a high average accelerating gradient. The DRFS also offers a robust system
platform where, in case of failures of cavities, couplers or tuners, the affected number of cavity units
is well contained (limited to four or two).
53
System Components
Figure 4.6.3.2 shows the configuration of rf units in the baseline case. Table 4.6.3.2 gives the
component counts. For comparison with the RDR design, the numbers are quoted for a system which
correspond to three units of cryomodules.
Figure 4.6.3.2 : System configuration of DRFS in the baseline case.
Table 4.6.3.2: The component count for the DRFS
in the baseline case. For comparison with the
RDR. The numbers are quoted for a group of
three cryomodules.
Figure 4.6.3.3 shows the reconfigured system in case of high current option. Table 4.6.3.3 gives the
component counted associated with each configuration.
Figure 4.6.3.3 : System configuration of DRFS in the baseline case.
54
Table 4.6.3.3: The component count for the DRFS
in the high-current case. For comparison with the
RDR, the numbers are quoted for a group of
three cryomodules.
Tunnel Layout
Studies of hardware installation of DRFS in the single-tunnel scheme so far have been made, and so
far it is known to be accommodated within a tunnel diameter of 5.2m. Figure 4.6.3.3 shows a cross
sectional view of such an installation layout.
Cryomodule is hung down from the ceiling. The support structure with suitable vibration
suppression needs to be worked out and this is an issue to pursue during TDP2.
RF sources are nearly uniformly distributed the tunnel: the modulators, DC power supplies, LLRF rack
units and other electrical devices are installed inside a shielded area for protection from the radiation
exposure. Thickness of the shield structure and materials are currently considered as per similar
implementation considered at EuroXFEL. They will be examined more closely during TDP2.
Underneath the floor the air ducts will be deployed with a total cross section exceeding 1 m2.
Figure 4.6.3.3: A cross section view of a possible DRFS installation in a main linac tunnel.
Figure 4.6.3.3 includes a working space for the traffic of transport girders to carry the cryomodules
and RF components. Figure 4.6.3.4 shows a three-dimensional isometric view of the DRFS installation
which is introduced by Figure 4.6.3.3. In addition to the scheme of Figure 4.6.3.3, a few other
55
installation layouts are under study: one having the cryomodules installed on the floor in a 5.75m
tunnel, and another having cryomodule placed on top of the shield structure within a 5.2m tunnel.
Figure 4.6.3.4: A 3D isometric view of the DRFS installation illustration the scheme given in Figure
4.6.3.3.
One of the potential issues in hardware implementation of the DRFS is the heat dissipation by the
components in the single main linac tunnels. Table 4.6.3.4 summarizes the present estimate of
pertinent parameters which are taken into consideration for the civil and tunnel engineering design
as discussed in Section 5 of this document.
Table 4.6.3.4: Summary of current estimate of heat water load and heat load due the DRFS
components .
56
Cost Considerations
The DRFS requires a large number of DC power supplies and MA klystrons. The reduction of the cost
is a key issue to validate its merit as a system solution. The following considerations are given.
MA Klystrons
The total of 4200 MA klystrons need to be manufactured during a construction period of 5 years
(850/year). During the subsequent operation period, approximately 200 MA klystrons need to be
produced per year. The perceived cost reduction measures are as follows:






For klystrons after the assembly, the vendors would only perform the baking of completed
klystrons, yet no HV processing. The HV/RF processing of the klystrons are to be done on the ILC
site. Consequently/RF, no modulators are to be maintained for HV/RF processing in the vendor’s
site,.
Deployment of hydro-forming techniques for manufacturing common parts of the klystrons.
Deployment of auto tuning machines for tuning of the cavities to use in the klystron body.
Elimination of ion pumps on the klystrons. Deployment of getter pumps.
Elimination of lead shield,.
Elimination of solenoid focusing magnets. Deployment of permanent magnet focusing.
DC Power Supplies and Modulators
As stated earlier, one common MA modulator and one common DC power supply, with one backup
unit each to ensure the system reliability, will drive 13 units of klystrons for cost saving. A bouncer
circuit is introduced in the DC power supply. This can be done with small capacitors in this application.
Since the MA modulator does not supply the pulsed power, its cost can be kept low.
Power Distribution
Very simple power distribution system, which results in cost reduction, is possible in this scheme; no
circulator, a power divider employing Magic-T with a high isolation for space saving, one phaseshifter with symmetric PDS between couplers or asymmetric PDS with a phase-fixed waveguide for
cost saving and 750 kW RF propagation in the dry air without any extra ceramic window.
Taking into account the above mentioned considerations, our present estimation of the construction
cost for the RF system for DRFS is approximately 2/3 of what has been evaluated for the RDR
HLRF/LLRF systems. The reason why it is not 1/2 is due to the conservative approach that we
currently take toward the costing of MA klystrons with permanent magnet focusing. The costing of
MA klystrons, and its reduction, is one of the issues that will be pursued in TDP2.
Operational and Availability Issues
One of the main availability issues for DRFS is the radiation damage of components. Almost all DRFS
components will be installed in the tunnel where we need to examine the potential radiation hazard
which is caused by acceleration of the dark current. The data concerning this issue is unfortunately
sparse. At this moment, we follow the considerations given in case of the EuroXFEL project, and
assume to introduce radiation shield of 10cm-thick concrete and 1cm-thick lead.
Another important availability issue is the maintainability of HLRF components, in particular the
klystrons. A large number of klystrons are use in the main linac tunnels, as one klystron feeds its
power to four cavities, as opposed to 26 cavities by one klystron in RDR. A longer MTBF for klystrons
57
is expected to be realized with the measures as described before. However, maintainability of
klystrons and other components remains to be a key issue. In the following, the maintenance
scenario for the DRFS and the associated workload are examined.
As discussed in Appendix 1, the operation schedule is assumed to consist of a repetition of 2-week
continuous operation (312 hrs) interrupted by one-day maintenance in between. The numbers of
failed components, which require replacement or repair, after each of the 2-week operation, are
estimated and are summarized in Table 4.6.3.2. The MTBF assumed in the estimates are also
indicated.
Table 4.6.3.2: Estimated number of HLRF components across the entire ILC requiring replacement or
repair, after completion of each of the 2-week operation period:
Component
DC power supply
MA modulator
MA klystron
# of units
requiring
MTBF assumed
replacement
or repair
2
50,000 hours
1.5
70,000 hours
12
110,000 hours
Total # of units
deployed at the ILC
325
325
4225
It should be noted that the scheme taken with redundant DC PS and MA modulator will provide
substantially longer "system" MTBFs for the DC power supplies and MA modulators than what are
quoted in Table 4.6.3.2. The MTBF quoted for the klystrons is based on the operational experience of
KEKB linac, where 4.7 million hours of cumulative operation time has been logged at 60 klystron
sockets. BI (barium-impregnated) cathode and lower cathode loading, which assure longer life time
of klystron, are adopted for this project.
In order to estimate the human resources that are required during each of the one-day maintenance,
the time that it takes to replace each of the components in question has been examined. Table
4.6.3.3 shows the summary of this study.
Table 4.6.3.3: Estimated times for the repair work of DRFS.
Action
Transportation of klystron
Time for unit piece of work
0.5 person-hours / tube
Removal of a failed klystron and
installation of a replacement
klystron
Time for personnel to move
from one point of repair to
another
Replacement of a MA
modulator
Replacement of a DC power
supply
4 person-hours / tube
Rationale
2 person in 2 hours could bring
8 tubes on one carrier.
2 hours with 2 persons
2/3 person-hours / tube
20 minutes with 2 persons
6.67 person-hours / modulator
27 person-hours / DC power
supply
58
As a consequence, the human resource that it takes to replace the 12 MA klystrons on each of the
one-day maintenance day would amount to 62 person-hours. Likewise, 1.5 MA modulators will
require 10 person-hours, and 2 DC power supplies will take 54 hours to repair.
In summary, the required human resource for HLRF during maintenance is calculated to be 126
person-hour, which corresponds to 16 person-days. This is likely to be manageable in the light of
other work that would be necessary during maintenance days.
R&D Plan and Schedule
Although the technologies of individual components for the DRFS are fairly conventional, it is still
important to demonstrate its system functionality by operating at least a few number of units.
Therefore, a task force has been commissioned at KEK in 2008 to pursue the R&D relevant to the
DRFS RF system.
Prototype RF components of a MA klystron together with parts of power supply and MA modulator
will be manufactured in JFY09 and further R&D will ensue. A prototype unit DRFS will be operated
and evaluated in the S1 global test at KEK in 2010, first with MA klystrons based on solenoid magnet
focusing. This system is expected to continue operation for the buncher section of the STF2 at KEK.
The goal is to accumulate operational experience in a semi-real-life working conditions through 20112012. The DRFS to be tested in this STF2 operation will be two units of MA klystrons driving eight
cavities, and/or four units of MA klystrons driving eight cavities. Development of permanent magnet
focusing for the MA klystrons will be pursued in JFY2010 and beyond.
As for the system implementation aspects, a more thorough cost study will be made together with
the CFS team during TDP2. Potential vibration issues of cryomodules, when they are hung off the
tunnel ceiling, will be studied, too.
59
4.6.4 Advanced Design Issues with Main Linacs in TDP2
Besides the new rf hardware configuration discussed in Section 4.6.2 and 4.6.3, and in addition to the
general issues with the high gradient cavity R&D as discussed in Section 2.1, we are intending to
address general main linac design and main linac operability issues during TDP2. While these issues
do not constitute a specific part of the SB2009 rebaselining proposal, they are presented here to put
the overall design and R&D subjects during TDP2 into perspective.
This subsection touches on three such issues:
1.
An analysis of alternative schemes for implementing the low current option. It is shown that by
fine tuning the average beam current, the RF pulse length can be controlled so as to be same as
the RDR specifications, at an expense of an increased requirement for the peak RF power.
2.
A comment on the 'zero-gradient' operation. It illustrates, with the positron target located at the
end (WA 3), how the main linac cavities can be adapted for low energy operation.
3.
A comments on the 'negative-gradient' operation which has been suggested in the RDR. It
compares the low energy operation in this Proposal against that in the Reference Design.
Alternative Current Option
In the SB2009 proposal, the beam power for the initial installation has been cut in half from that of
the RDR. This allows reduction of both the size of the damping rings and the rf power. A straight
forward approach, aimed at minimizing peak rf power, was to eliminate every other bunch, thus
halving the bunch train current. However, it has the effect, for the same acceleration gradient, of
doubling the cavity fill time, thereby increasing the required rf pulse width by 38% from 1.56ms to
2.16ms. This pushes the specification for the klystrons and modulators beyond what’s been originally
anticipated in RDR and assumed in the R&D effort so far. The longer fill time also increases the
cryogenic dynamic heat-load.
An alternative way to achieve the same reduction in average current or number of bunches per pulse
is to change both the bunch train current and the bunch train length. Examples are shown in Table
4.6.4.1. If the former is reduced to 69% of the RDR value and the latter to 72.5%, the required rf
pulse duration would remain unchanged. The increase in fill time is balanced by the shortened beam
pulse. In this scenario, the number of bunches would still be halved, with their spacing increased by a
factor of 1.45, rather than doubled.
Figure 4.6.4.1: Comparison of reduced beam power bunch train options and 9 mA power upgrade
with implications for rf power sources and cryogenics
Option A
32 (832)
1,335
4.5
969
19
2.16
170
368
1.406
1.071
# of rf units (cavities) per KCS
# of bunches per beam pulse
bunch train current (mA)
bunch train duration (s)
# of klystrons/modulators per system
rf pulse duration (ms)
peak rf power per system (MW)
rf pulse energy (kJ)
norm. cavity dynamic heatload
norm. total cryo dynamic heatload
60
Option B
32 (832)
1,335
6.21
702
26
1.56
240
375
1.019
0.932
upgrade
32 (832)
2,670
9
969
36
1.56
340
531
1
1
More rf sources are required for this option, but still significantly fewer than for the baseline current
and without extended pulses (which capacity would be unusable in the event of an upgrade).
Furthermore, the cryogenic dynamic heat-load can be actually reduced by a few percent.
This discussion illustrates one of the topics to pursue when further system optimization is sought
during TDP2.
Operation of Cavities with Zero Gradient
To run beam through all or part of the linac with zero gradient without detuning the cavities, or even
changing the QL’s, the rf should be set to arrive simultaneously with the beam at a power of ¼ the
nominal value. This pulse will effectively be fully reflected from the cavity into the waveguide load, as
the normal decay of the drive reflection is balanced by the rising beamloading emitted field. Because
the cavity is thus prevented from filling with rf, there is no post-beam discharge. The waveguide
loads would receive ¼0.45 = 0.1125 of the full gradient rf drive energy (where 0.45 is the fraction of
the nominal pulse width which covers the beam), as opposed to 55% of it (fill time portion) in
accelerating mode.
Operation of Cavities with Negative Gradient
With the undulator placed at the 150 GeV point in the electron main linac, operation below 300 GeV
cms energy requires deceleration of the beam. Specifically, to pick up where LEP left off, at 209 GeV
cms, the electron beam must be decelerated by 45.5 GeV (from 150 GeV to 104.5 GeV) over the
section of the linac designed to add 100 GeV (to 250 GeV). This calls for powering the cavities out of
phase at -45.5% of the design gradient, or -14.33 MV/m (average).
Reducing the QL’s, the fill time and the rf power by a factor of |V|/V0 will produce the desired
gradient at beam arrival. The rf pulse can then be terminated, and the retarding gradient would be
maintained constant as the beam loading buildup cancels the decay of the preloaded rf. During the
beam, the power extracted from the beam, equal to the preceding input rf, leaves the cavities to be
absorbed in the loads.
The nominal fill time is 55% of the rf pulse (for either low power bunch spacing option), which
fraction of the pulse energy is nominally deposited in the waveguide load during cavity fill and
discharge. Thus, for operation in a decelerating mode, the input rf energy is 0.55(|V|/V0)2 and the rf
heat load 0.55(|V|/V0)2 + 0.45(|V|/V0) times the nominal pulse energy. This is less than the
nominal heat load up to |V|/V0 = 0.671 and only 31.9% at the maximum value of 0.455 envisioned
above.
It would be simpler to achieve low energy by operating at full reverse acceleration for the
appropriate fraction of the linacs, leaving the power, timing and coupling unchanged. However,
unless the repetition rate were dropped, this would increase the rf heat load by 82%, in the local
waveguide loads and 29% in the couplers, requiring greater cooling capacity to be installed in the
tunnel.
61
4.7 Beam Delivery Systems
4.7.1 Overview
This section describes the design changes in the Beam Delivery Systems (BDS) of the ILC considered
as a response to the proposed changes of the design baseline. The elements in the baseline changes
which are of particular relevance to BDS are the following two:

Changes in the subsystem integration of the central region: As of the RDR, the BDS, the electron
source and the damping rings are clustered in the central region of the ILC accelerator complex.
The proposed changes in the baseline envisage relocation of the positron source system to the
downstream end of the electron main linac, so that they also join this central region. This
impacts the subsystem layout in ways that affect the implementation of electron side BDS.

Changes in the baseline parameter set: Proposed adoption of the low power beam parameter
set (same machine pulse repetition rate and the same bunch intensity, but a reduced number of
bunches per pulse) leads to a desire to push the beam-beam parameter, so that the same
luminosity as in RDR can be achieved. As a solution the so-called travelling focus scheme is being
considered.
However, in terms of beam energy handling, similar to the RDR design, the BDS design remains
compatible with 1TeV CM upgrade which is expected to be accomplished by installing additional
dipoles and replacing the final doublet.
In what follows, the revised design solutions to adopt in the BDS and their technical feasibilities are
presented.
4.7.2 System Layout
The central integration includes the sources in the same tunnel as the BDS. Relocation of the positron
production system to the downstream end of the electron linac means placing it just before the
beginning of the electron BDS. These changes need suitable design modifications to the layout of this
area. Figure 4.7.1 shows the proposed new layout of the electron BDS.
Figure 4.7.1: Proposed new layout of the electron BDS.
The most notable feature of the new electron BDS layout is the introduction of the dogleg so as to
create the required transverse offset between the electron beamline and the positron photon target.
62
Another important consideration is the protection of the undulator from miss-steered as well as an
electron beam with significant energy errors, which is now shares the same systems foreseen for the
BDS. These changes apply only to the electron side. The positron beam delivery system design
remains almost the same as in the RDR excepting few modifications such as separating the combined
functionalities of the machine protection energy chicane, the upstream polarisation measurements
and the laser wire detection, which was found to be problematic to both polarimeter and laser wire
detection purposes. These changes were agreed to be implemented on both the electron and
positron BDS designs in the TD Phase 2 phase before SB2009. The increase in the length caused due
to separating these functionalities will be absorbed by slightly shortening the BDS final focus designs
allowing slightly more emittance growth due to emission of synchrotron radiation at 1 TeV CM.
These and other features in the new electron BDS are summarized below:
1.
Sacrificial collimator now located at the linac end rather than in the BDS upstream end: The
RDR has sacrificial collimators in the beginning of e- and e+ BDS to protect the BDS from any
beam with error to enter from the large aperture of the main linac (r=70mm) into small aperture
(r=10mm) of the BDS. In the new layout, the small aperture undulator (~8mm full) is located
immediately after the linac and thus it needs to be protected against any error beam entering
the undulator. This is done by moving the sacrificial collimator section and an energy chicane to
detect the off energy beam in front of the undulator which reduced the electron BDS length to
2104m from 2226m as shown in Figure 4.7.1. Any beam entering this section with errors will be
detected and sent to the fast abort line just before entering the undulator. The fast abort line is
presently the same length as the RDR abort line, which was designed as a fast abort + tuning line
(the positron BDS side still has this combined functionality), however the fast abort beam dump
needs to be able to take only the number of bunches between abort signal and stopping the
beam at the extraction of the damping ring and does not need to be a full power beam dump.
The exact rating for this dump remains to be determined.
2.
Matching line after the fast abort detection energy chicane into the undulator and design
requirements for positron target location: The matching line to the undulator needs to allow
sufficient transverse separation for the abort line and then matches into the undulator FODO
cells. The photons generated in the undulator will pass through a drift length of 400m up to the
positron target (~1070m point in Figure 4.7.1). To implement the positron target and the remote
handling of the components in this area, a transverse offset of 1.5m is required between the
electron beamline and the photon target. The remote handing area needs a drift space of
approximately 40m in length. No BDS component are placed in this space. This is achieved by
using a matching section after the undulator to match into a dogleg, a dogleg itself giving a
transverse offset of 1.5m and a 40m long drift at the end.
3.
Dogleg lattice to create the required separation between the photon target and the electron
beamline: The dogleg lattice has been designed to be a TME (Theoretical Minimum Emittance)
lattice. This keeps the emittance growth due to synchrotron radiation at 1 TeV CM to be within
few percent. The dogleg provides an offset of 1.5m in 400m as required and the emittance
growth at 1 TeV CM is ~3.8%. The dipoles in the dogleg are presently not decimated but can be
decimated similar to the rest of the BDS so that only few dipoles are installed at 250 GeV. The
beam dynamics and tuning effects on the BDS due to the presence of the dogleg need to be
assessed.
4.
Matching section into the BDS diagnostics section: The 40m long drift is followed by a matching
section into the skew and coupling correction section, chicane for detection of the laser wire
photons and a slow tune-up (DC tuning) line leading to a full power beam dump. Since the fast
abort functionality is being taken care of by the fast abort line before the undulator, the energy
63
acceptance of the DC tuning line is much reduced and thus the DC tuning line can be shortened
using only DC magnets. This optimisation will be done during the TDP2 phase.
5.
Polarimeter chicane, collimation, energy spectrometer and final focus: The polarimeter chicane
will be located just after the take-off section for the tuning line, which is not shown in the
layout. This will need some additional length but will be accommodated by slightly reducing the
final focus length allowing some emittance growth at 1TeV CM. The polarimeter chicane will be
followed by the betatron and energy collimation, energy spectrometer and final focus sections
similar to the RDR.
6.
Post collision extraction line and main dump: The post collision extraction line remains similar
to the RDR and will be terminated in to a full power beam dump. The power of the main dumps
remains at 17 MW, as in the RDR, to allow future upgrades in the beam parameters.
The layouts to combine the full power tuning dump and the main extraction line full power beam
dump will be studied during the TDP for both the positron as well as electron BDS designs
As described in Section 4.3.1, the proposed solution to operate the ILC beam energy less than 125
GeV is to pass alternate 150 GeV and lower energy pulses from the linac. The 150 GeV pulses passing
through the undulator will generate the positrons and lower energy bunches will pass through the
BDS for collisions. As shown in Figure 4.3.1, the 2.5 Hz switching magnet located after the undulator
will send the 150 GeV pulses through an additional dogleg and a beam line which will merge with the
high power tuning beam dump. The design of this line will need modification to the layout shown in
the figure 4.7.1 by increasing the drift space between the end of the undulator and the start of the
dogleg to the BDS. A 2.5 Hz switching magnet will be installed in this space and an additional dogleg
and a beam line operating at 150 GeV will be designed to merge with the tuning dump in TDP2.
4.7.3 Reduced beam-power parameters
The proposed reduction in the beam power (number of bunches per pulse) requires us to squeeze
the beam-beam parameters to compensate the nominal factor-of-two reduction in luminosity.
SB2009 explores two possibilities:
1. Pushing the beam-beam parameters into a high-disruption regime close to the single-beam
kink-instability limits, at the expense of higher beamstrahlung and tighter collision tolerances.
The proposed parameters could in principle recover the nominal RDR luminosity to within
25% (1.5×1034 cm-2s-1).
2. Making use of the so-called Travelling Focus effect [BDS1], which can recover the remaining
25% luminosity without a further increase in the beamstrahlung. This approach comes at the
cost of a very high disruption parameter, and the need for additional hardware (see Section
4.7.4).
Both these approaches are made possibility in part by the existing Reference Design requirements
associated with the parameter plane. SB2009 effectively pushes the beam-beam parameters to the
limits of the parameter plane space, but does not in general exceed them. The BDS and in particular
the IR should in general be able to accommodate these modification without any further
modification (although clearly more studies are needed). One exception is the high disruption
parameter associated with the Travelling Focus proposal. This is a special case and is treated in more
64
detail in the following section. Table 4.7.1 shows the relevant beam-beam parameters compared to
the limits corresponding to the RDR parameter plane (reproduced from Section 2.4).
Table 4.7.1: RDR parameter plane ranges compared to SB2009 specifications
(TF refers to Travelling Focus). This table is reproduced from Section 2.4.
min
1
x 1010
RDR
nominal
2
max
2
SB2009
no TF
with TF
2
2
Bunch population
1260
2625
5340
1312
1312
ns
180
369
500
530
530
m
200
300
500
300
300
mm-mr
10
10
12
10
10
mm-mr
0.02
0.04
0.08
0.035
0.035
mm
10
20
20
11
11
mm
0.2
0.4
0.6
0.48
0.2
nm
474
640
640
470
470
nm
3.5
5.7
9.9
5.8
3.8
%
14
1.7
19.4
2.4
26.1
5.5
25
4
38
3.6
1.5
2
Number of bunches
Linac bunch interval
RM bunch length
Normalized horizontal emittance at IP
Normalized vertical emittance at IP
Horizontal beta function at IP
Vertical beta function at IP
RMS horizontal beam size at IP
RMS vertical beam size at IP
Vertical disruption parameter
Fractional RMS energy loss to
beamstrahlung
Luminosity
x 1034cm-2s-1
2
4.7.4 Travelling Focus Scheme
The travelling focus[BDS1] is a technique in which the focussing of opposing bunches is longitudinally
controlled so as to defeat the hourglass effect and to restore the luminosity. The matched focusing
condition is provided by a dynamic shift of the focal point to coincide with the head of the opposing
bunch. The longer bunch helps to reduce the beamstrahlung effect and improvement of background
conditions is expected. Similar to the nominal 500GeV CM case, the 250GeV CM parameters would
also benefit from application of travelling focus – the work on development of a corresponding
parameter set is ongoing.
The travelling focus can be created in two ways:
1.
Method 1 is to have small (uncompensated) chromaticity and coherent E-z energy shift E/z
along the bunch. The required energy shift in this case is a fraction of a percent.
2.
Method 2 is to use a transverse deflecting cavity giving a z-x correlation in one of the Final Focus
sextupoles and thus a z-correlated focusing. The needed strength of the travelling focus
transverse cavity was estimated to be about 20% of the nominal crab cavity.
At this moment, since the increased energy spread in method 1 is a potential issue for physics
research, method 2 is the preferred focus of studies. Detailed evaluation of the transverse cavity
design and dynamics of the beam is ongoing.
65
4.7.5 R&D and Design Work to Pursue in TDP2
The more demanding beam-beam parameters associated with SB2009 force us to be in a regime of
higher disruption. Although there appears to be no fundamental show stoppers, a comprehensive
study involving simulations is still required in an attempt to quantify the performance. Specifically:

The higher disruption results in a higher sensitivity to any beam-beam offset. Thus, operation
of the intra-train feedback and intra-train luminosity optimisation becomes more important
and more challenging than in the case of RDR. Early estimates suggest that in order to
contain the luminosity loss within 5%, a bunch-to-bunch jitter in the train needs to be less
than 0.2nm at the IP (~5% of a nominal beam sigma).

The parameter sets also have twice as small vertical betatron functions at the IP, which imply
either tighter collimation, with gaps 40% closer to the beam core. This has implications for
wakefields (emittance preservation) and fast feedback systems.

Enhanced beam-halo loss in the tighter collimation could potentially increase the number of
generated muons and hence the muon shielding requirements[BDS2]. (This is difficult to
quantify as it depends on the specifics of the models of beam halo used.)
In addition, the operability and ability to commission the entire system continue to be investigated
during TD Phase 2. Further design integration and careful planning of the orchestration of the
activities in the central area are envisaged.
References
[BDS1] V.E. Balakin, Travelling Focus' Regime for Linear Collider VLEPP Proc. 77th ICFA Workshop on
Beam Dynamics, May 13-16, Los Angeles, 1991.
[BDS2] Two alcoves, for 5m and 18m walls are foreseen in the RDR, while initial installation was
foreseen with just a single 5m set of walls.
66
4.8 Conventional Facilities
4.8.1 Overview:
This section discusses the revision of the conventional facilities design to take place as a result of
proposed changes of the design baseline (SB2009 Working Assumptions), as discussed in previous
sections of this document.
The SB 2009 Working Assumptions proposed in the Accelerator Design and Integration (AD&I) effort
affected all of the major cost drivers in the RDR Conventional Facilities design solution, and the cost
associated with it. Those cost drivers were the Underground and Surface Civil Construction,
Mechanical Support Systems and Electrical Distribution. By far, the most important single
contributor to the CFS cost is the underground volume required in the reference design, and thus,
the reduction of the underground volume and the corresponding reduction in cost is the primary
focus of this CFS AD&I effort. The net effect of the design changes proposed with SB2009 amounts
to reduction of the underground tunnels by approximately 26km. In addition, issues concerning Life
Safety and Egress requirements also required fundamental review, during the Conventional Facilities
AD&I effort.
It should be noted, however, that prior to the start of the AD&I work, the Conventional Facilities
Group conducted a comprehensive Value Engineering (VE) Review of the Process Water and HVAC
design contained in the Reference Design Report. A document that describes the VE process[CFS01]
was completed by the CFS Group. As a result of this review and subsequent analysis of the
alternatives indentified, a decision was made by the ILC Project Managers to accept a Process Water
design solution that was based on a larger system ΔT. This change to the RDR Conventional Facilities
Design became the baseline for the analysis of the SB 2009 Working Assumptions and the AD&I
process.
4.8.2 CFS Analysis and Review Process
Table 4.8.1 lists the proposed SB 2009 Working Assumptions (WA) as reference to the description of
the CFS approach.
Table 4.8.1: Labels for proposed baseline changes (SB2009 Working Assumptions),
as referenced in discussion of the conventional facility designs.
WA1. A Main Linac length consistent with an optimal choice of average accelerating gradient
(currently 31.5 MV/m, to be re-evaluated)
WA2. Single-tunnel solution for the Main Linacs and RTML, with two possible variants for the
HLRF
(a) Klystron Cluster scheme
(b) DRFS scheme
WA3. Undulator-based e+ source located at the end of the electron Main Linac (250 GeV)
WA4. Reduced parameter set (with respect to the RDR) with nb = 1312 and a 2ms RF pulse.
WA5. ~3.2 km circumference damping rings at 5 GeV, 6 mm bunch length.
WA6. Single-stage bunch compressor with a compression factor of 20.
WA7. Integration of the e+ and e- sources into a common “central region beam tunnel”,
together with the BDS.
The CFS approach to the analysis of these working assumptions was basically three-fold:
67
1.
WA 2a and 2b were reviewed by each of the three regions. The focus of this review was to
develop a single tunnel solution that would accommodate both the Klystron Cluster and DRFS
schemes for supplying RF power to the cryomodules in the Main Linac tunnels. It quickly
became apparent that geological conditions and life safety and egress requirements would not
only have a direct impact but would also lead to different design solutions in each of the three
regions (Asia, Americas and Europe). In addition, a regional preference for one or the other of
the proposed RF schemes was identified as a result of the analysis. Both proposed RF schemes
will be reviewed in each of the three regions, however the Asian Region prefers the Distributed
RF scheme, the Americas Region will evaluate both proposed RF schemes and the European
Region prefers the Klystron Cluster RF scheme.
2.
WA 1, 3, 5, 6 & 7 were combined into a single design solution based on the input from the
Technical Area Systems. This effort to reposition components and reduce the overall
circumference of the Damping Ring were combined into a comprehensive new Central Region
design and evaluated as a single CFS design solution.
3.
WA 4, which is the reduced parameter set, or “low power” option was reviewed as a single item.
In the subsequent sections the CFS analyses are given to each of these WA groups to outline the
corresponding technical solutions, considerations and cost estimates.
4.8.3 WA 2a & 2b CFS Analysis
Asian Region
The Asian Region has chosen the Distributed RF scheme, rather than the Klystron cluster scheme, as
the technical implementation of the accelerator system to focus, when developing a single tunnel
solution for the Main Linac tunnels. This is because of the characteristics of the Asian sample site
which assumes mountainous or hilly geography:

The hilly or mountainous geography favors the use of sloped or near horizontal path ways rather
than vertical shafts as access routes to the Main Linac tunnel complex.

While the horizontal access tunnels provided “grade level” access, they tend to be much longer
than the vertical shaft access adopted in the other two regions.

The long access path ways lead to a problematic implementation of the Klystron Cluster scheme
where large, long high-power waveguides have to share the same space as the traffic for
installation and maintenance.

In addition, all surface buildings which would house the Klystron Cluster would necessarily need
to be located at the entrances to the access path ways.
The Distributed RF system is free from these issues.
In the single-tunnel scheme of the Main Linacs with DRFS, a system was developed to
“compartmentalize” the Main Linac tunnel into ~500m sections, as shown in Figure 4.8.1. These
sections are separated by fire walls and doors. A ventilation system will be provided to isolate the
incident section and supply fresh air to the rest of the tunnel and a safe pathway is retained to the
next available access to the surface. These offer a reasonable solution to the requirements for life
safety and egress.
68
Figure 4.8.1: Example of Asian and European “Compartmentalized” Tunnel Layout
As shown by Figure 4.8.2, in the current DRFS solution by the Asian CFS team the cryomodules will be
suspended from the ceiling of the Main Linac tunnel. Supporting RF equipment is arranged below the
cryomodules. Additionally the floor space is allotted for movement of equipment and personnel
access. The whole set of hardware equipment and space is known to fit within a Main Linac tunnel
with a diameter of 5.2m. Optimization of the tunnel size, in particular its reduction, is an issue to be
examined more closely during TDP2.
Figure 4.8.2: Asian Region Main Linac Tunnel Cross Section
The Asian Region is also reviewing the implications and design with the Klystron Cluster scheme for
their sample site configurations. However, the solution in this case might result in a less than an
optimal design for the Asian sample site.
69
In consideration of specific characteristics of the sites to consider, the following additional issues
remain to be examined:

The specifics of the access methods to the single main tunnels have to be worked out. The RDR
design assumed the use of sloped access path ways every ~5km along the main tunnels. A design
optimization will be required in TDP2 by taking advantage of a mixture of vertical, sloped and
horizontal access paths.

The specific details in the facility design have to be developed. During the RDR period, the
facility system designs were worked out with an assumption that a world-common solution
would apply, except those for the civil construction and high-voltage power distribution. During
TDP2, specific design solutions have to be developed on the surface buildings, electricity, cooling
and air-conditioning.
It is the intention of the Asian Group to address these issues during TDP2.
Americas Region
The Americas Region initially focused on utilizing the Klystron Cluster RF scheme to develop a single
tunnel design for the Main Linac tunnel. Figure 4.8.3 shows a cross section view of the KCS
implementation in the main linac single-tunnel, as considered by the Americas regional team.
However, both the Klystron Cluster and DRFS schemes can be accommodated in a single tunnel
solution for the Americas sample site.
Figure 4.8.3: Americas Region Main Linac Tunnel Cross Section
70
The Americas reference design incorporates all vertical shafts of similar length for access to the Main
Linac tunnel. This arrangement is conducive to relatively uniform waveguide lengths using the
Klystron Cluster scheme for RF distribution. The increase in the surface building footprint at vertical
shafts to accommodate the new location of the Klystron Clusters can be problematic in denser urban
areas. However, there is no fundamental constraint that would preclude the Klystron Cluster or
Distributed RF scheme in the Americas Region. Specific characteristics of the Americas site include:

Relatively uniform surface elevation which allows for vertical shafts of similar length for
underground access.

This configuration provides the opportunity for fairly uniform waveguide lengths for the
Klystron Cluster RF option, but does not preclude use of the DRFS RF option.

The Americas sample site is in a relatively populated area. However, the central region and
related surface presence can be contained within the Fermilab site boundaries.

Shafts located beyond the Fermilab site will necessarily require acquisition of surface area.
However, positioning of the alignment can take advantage of existing open areas and
industrial sites and minimize the impact on residential communities.
With respect to life safety and egress, the Americas Region is subject to a more rigorous environment
of regulatory and code oversight. After identifying and then reviewing available code information, a
viable and compliant design solution was identified and applied to the single tunnel design for the
Main Linac.
The solution is based on the isolation of fire loads resulting from oil filled equipment primarily
located in the caverns at the base of the vertical access shafts. This isolation is accomplished through
the use of fire walls and doors which allows the majority of the underground space to be used as a
safe pathway to the next available vertical access to the surface. Use of the Klystron Cluster RF
scheme produced a solution with floor-supported cryomodules, floor space allotted for equipment
movement and personnel access. This tunnel cross section design results in a Main Linac tunnel with
a diameter of 4.5m, as shown in Figure 4.8.3. For the analysis of the DRFS scheme, the Americas
region utilized the Asian tunnel layout and will use the same 5.2m tunnel diameter for the Main
Linac.
European Region
The European Region also initially focused on the Klystron Cluster RF scheme to develop a single
tunnel design for the Main Linac tunnel. This decision was fundamentally driven by the geologic
constraints of the European sample site. The structural capacity of the rock in which the tunnel will
be constructed as well as the requirement for transverse ventilation ductwork at the tunnel ceiling
makes it difficult to support any substantial load, like a cryomodule, from the tunnel ceiling.
Based on this constraint, the Klystron Cluster RF system is the primary RF system currently being
considered for the European sample site. Although, transferring more equipment to the surface will
create additional environmental concerns, it is not possible to measure the impact of this change at
this stage. Figure 4.8.4 shows a cross section view of the KCS implementation in the main linac
single-tunnel.
71
With respect to the requirements for life safety and egress, the European Region uses the same
general approach (Figure 4.8.1) as the Asian Region. The concept is consistent with single tunnel
designs for CLIC and XFEL. Specific characteristics of the European sample site include:
Figure 4.8.4: European Region Main Linac Tunnel Cross Section

The current alignment for the European sample site follows a path under a low mountain
range.

Shafts are positioned along the alignment to take advantage of minimum shaft lengths
resulting in distances that can accommodate the Klystron Cluster RF option.

In general, the surface area above the alignment is basically rural in nature, but attention
must be provided to existing farms and vineyards.

Geologic conditions, as well as the ventilation ductwork located at the ceiling of the tunnel
and enclosures, preclude the mounting of cryomodules at the tunnel ceiling. For this reason
the European Region is focusing on the Klystron Cluster RF option at the present time.
The Main Linac tunnel is compartmentalized into 500m sections with fire walls and doors separating
adjacent area. A similar ventilation system to the Asian Region solution was provided to isolate the
incident section and provide fresh air to the rest of the tunnel and a safe pathway to the next
available access to the surface. Like the Americas Region, the European Region utilized the Klystron
Cluster RF scheme system which produced a solution with floor supported cryomodules, floor space
allotted for equipment movement and personnel access. Due to the need for supply and return
ductwork at the crest of the tunnel section, this tunnel cross section design results in a Main Linac
tunnel with a diameter of 5.2m. The European Region is also reviewing the implications and design
for using the DRFS scheme for their sample site configuration, although this solution is likely to result
in a less than optimal design for the European sample site.
WA2 Safety Issues
72
Table 4.8.2 gives a brief summary of safety considerations given by the three regional teams in
connection with the single-tunnel solutions for the main linac systems.
Table 4.8.2: Safety considerations given so far to the single-tunnel solutions for the main linac
systems.
Asia
Americas
Europe
Primarily DRFS.
Primarily KCS.
Primarily KCS
Isolation of incident
location
Firewalls / doors every
500 m along the
tunnels, and selective
closure of the airflow.
Utilize caverns at the
vertical shaft bases for
isolating oil-filled
equipment.
Firewalls / doors every
500 m along the
tunnels
Safe pathways
Rest of the tunnels.
Most of the tunnels.
Rest of the tunnels.
HLRF scheme
examined so far
The information summarized in the Table 4.8.2 has been taken from several Life Safety and Egress
Studies that have been developed in each of the three regional areas [CFS02-CFS06]. Based on the
information contained in these reports, each region developed solution to Life Safety and Egress
issues that is responsive to the regional guidance and local regulations. Each regional design solution
is described briefly below:
Asian and European Regions
The Life Safety and Egress approach for the Asian and European Regions are very similar in nature
and based on the fundamental idea to divide the underground single tunnel configuration in 500m
“compartments” separated by a fire rated wall and door with self-closing capability. A ducted supply
and return air supply is also part of the solution. Using automatically controlled dampers in the
supply and return air lines, smoke can be isolated within a fire incident compartment and allow fresh
air to remain flowing to unaffected compartments. Unaffected compartments can then be used for
personnel escape to surface areas and also for access by emergency response personnel.
Compartment walls will be installed with the initial construction and installation effort with doors
sized to accommodate movement of replacement components during the operational phase.
Americas Region
The Life Safety and Egress approach for the Americas Region is fundamentally different from that of
the Asian and European Regions. For the Americas sample site, regulation provides for the
underground space to be divided into “Building” space and “Common” space. Simply stated, the
“Building” space is considered the functional or operational portion of the underground space and
the “Common” space is developed underground space that is not considered “Building” and includes
such things as entrances. This interpretation of these requirements has resulted in a solution that
identifies those areas with the highest risk of fuel load and fire hazard as the “Building” space,
specifically the cavern areas at the base of each vertical shaft. These areas will be enclosed with fire
rated walls and doors to separate this area from the less hazardous “Common” space which is the
accelerator tunnel, which can be used for personnel escape to surface areas and also for access by
emergency response personnel. Enclosure wall will be installed with the initial construction and
installation effort with doors sized to accommodate movement of replacement components during
the operational phase.
73
As the component design becomes more defined, these requirements will be revisited. However, we
believe the current design provides ample space for all required beamline components and large
increases in size requirements are unlikely. In addition, space has been made available in the tunnel
cross sections and enclosure areas and volumes to accommodate process water and HVAC
equipment and electrical distribution components based on the criteria provided to date from the
various Area and Technical Systems. While the preliminary designs will be further developed in the
course of TDP II, like the technical components, we believe that ample space has been identified and
that large increases in requirements for component size are unlikely.
WA2 Summary
In summary, a single tunnel solution has been developed in each of the three regions for the
respective sample site. Each sample site has incorporated a preferred RF Distribution System, and
Life Safety and Egress solution that best accommodates local geologic constraints and regulatory
guidance and provides a defendable solution for a single tunnel design. All available information
regarding machine component size has been taken into consideration in the development of tunnel
cross sections and enclosure areas and volumes.
4.8.4 WA 1, 3, 5, 6 & 7 CFS Analysis
Major changes of the accelerator system layout in the SB2009, relative to the RDR design are as
follows:









Relocation of the Undulator based Positron Source to the end of the Electron Main Linac
Reduction of the Damping Ring Length to 3.2 km
Single stage Bunch Compressor in the RTML
Integration of the e+ and e- Sources into the central region Beam Delivery System area
Repositioning of the Damping Ring configuration completely to one side of the Interaction
Region
Positioning the Aborts and Dumps to point in the direction away from the Damping Ring side of
the Interaction Region
Relocating the BDS service tunnel to the Damping Ring side of the Interaction Region
Relocating the access aisle in the BDS areas to the service tunnel side of the BDS tunnel
Combing the existing shaft access to the BDS service tunnel with one of the existing Damping
Ring access shafts
With regards to the machine central area, as described above, all of these working assumptions were
combined into a single revised CFS solution. The primary focus for this effort, from the CFS point of
view, was to shorten the overall tunnel length and reduce the required underground volume. It
should be noted that although WA 1 was a part of this design review, the preferred cavity gradient
was maintained at 31.5 MV/m for the purpose of this analysis.
The input to the CFS Group came from the Technical Area Systems in the form of a beam layout
drawing that was compiled at the Daresbury Laboratory and based on available beam lattice files and
other technical input. In the course of developing a new design layout, various components from
different area systems were now found to be located in the same physical area which required a
cavern larger than the normal bored tunnel cross section. However, the final design result produced
an overall reduction in the underground volume required. In addition to the repositioning of area
system components, the location of the beam aborts and dumps required between the end of the
74
respective Main Linacs and the Interaction point was more clearly defined than in the RDR and
incorporated into the central region design.
Figure 4.8.5 shows the resultant topological diagram of the beamlines in the central area.
Figure 4.8.5: Topological diagram of the beamlines in the ILC central area.
To assist the members of both the area system groups and the CFS group in the SB2009 decision
making process, a set of 3D models have also been developed to assist in the SB2009 decision making
process. Figure 4.8.6 shows an example of such a model.
Figure 4.8.6: Example of 3D Enclosure Drawing
75
4.8.5 WA 4 CFS Analysis
The reduced parameter set or “low power” option for RF configuration was reviewed by the CFS
Group under the assumption that the election power capacity and distribution would be installed to
permit a relatively easy transition to a “full power” upgrade at some time in the future. As a result of
this assumption, the electrical distribution system did not differ in any great detail from the RDR
design. There was some reduction in the overall electrical distribution cost and some savings
identified in the process cooling water system, but overall the CFS contribution to cost savings for
this item was less than might have been expected. This approach put more emphasis and
importance on the transition to full power at a later date with minimum machine downtime to
accomplish the transition. Therefore, all transformer and conductor capacity are installed in the
initial construction. To design and install an electrical distribution system that only provided enough
capacity for the low power option would save on initial capital costs. However the upgrade to the
full power electrical capacity would require not only demolition of the lower power transformer and
conductor capacity and installation of upgraded equipment, but it would also entail a lengthier
machine downtime to accomplish the task.
4.8.6 Summary
The CFS Group has evaluated all of the SB 2009 Working Assumptions, and have produced a solutions
set to further pursue in TDP2. Figure 4.8.7 shows the overall schematic machine layout. The
directions of the beamlines are indicated with arrows. The information on the central region
provided in Figure 4.8.7 reflects only the Americas Region Design. Work is continuing in the Asian
and European Regions on the details of the central area of the machine and will be added as work is
completed.
76
Figure 4.8.7: Schematic beamline layout
77
Overall this present solution has brought a substantial amount of cost reduction, as will be
summarized in Section 5. By far, the largest single contribution was the elimination of the Main Linac
service tunnels and the establishment of a single tunnel Main Linac design solution. Although
geologic limitations have necessarily required local tunnel and enclosure design, each of the three
regions have developed a single tunnel solution based on the most appropriate RF distribution
scheme for each respective sample site. Each solution has been reviewed and provides a Life Safety
and Egress configuration that is compliant with local experience and known regulatory requirements.
The new central region layout has been established and will become the basis for the overall
machine design in all three regions. As with the Main Linac tunnel, the enclosures and caverns for
the central region layout will necessarily require adaptation to each region’s geologic requirements,
but the beamline layout will remain the same for all three regions.
The “low power” option for the electrical distribution did not generate a great deal of CFS cost
savings due to the assumption that the disruption to machine running during the eventual upgrade
to “full power” operation should be minimized. This is an issue that could be revisited in the future,
however not in the timeframe of this proposal. CFS cost savings were identified and will be added to
the technical system cost savings for an overall total savings for WA 4.
4.8.7 Technical Design Phase 2
The Accelerator Design and Integration effort has brought the Conventional Facilities design to a
more mature state and increased level of detail with respect to the previous RDR design. As a result,
the distinctions between the various regional design solutions have become more apparent and need
to be addressed. It is apparent that as we begin TDP 2, each region will necessarily have to develop a
single sample site design solution that is specific to regional conditions and technical system
preferences. As such, there will be the need to develop three distinct sets of 2D and 3D drawings as
well as three complete and distinct cost estimates. It is not anticipated that the development of
three distinct regional designs is likely to result in substantial changes to the existing CFS cost
estimates, however the distinctions and impact should be understood. The AD&I process has shown
the difficulty of developing a single 3D model, when different tunnel diameters and technical systems
are being used in the three regional designs. More such examples will become apparent over time
and the development of specific regional designs will help to accommodate these distinctions in the
TDP 2.
One final point needs to be addressed concerning the installation schedule during the construction
phase of the ILC. No further work has been initiated since the RDR effort by the CFS Group regarding
single tunnel installation issues and scheduling. This remains an open issue that will be studied as
part of TDP II.
References
[CFS01] “ILC Process Water and Air Treatment VE Cost Evaluation” - (ILC EDMS D*889835) http://ilcedmsdirect.desy.de/ilc-edmsdirect/document.jsp?edmsid=*889835 .
[CFS02] “Fire Safety for ILC Single Tunnel” - (ILC EDMS D*899575) http://ilc-edmsdirect.desy.de/ilcedmsdirect/document.jsp?edmsid=*899575 .
78
[CFS03] “Life Safety/Fire Protection Analysis for International Linear Collider – Single Deep Tunnel
and Near Surface Options” - (ILC EDMS D*899675) http://ilc-edmsdirect.desy.de/ilcedmsdirect/document.jsp?edmsid=*899675 .
[CFS04] Security and Workplace Safety Concepts for the Construction, Installation and Operation of
the XFEL Research Facility” - (ILC EDMS D*899615) http://ilc-edmsdirect.desy.de/ilcedmsdirect/document.jsp?edmsid=*899615 .
[CFS05] “Compact Linear Collider: Guidelines on Reduction of Fire Hazards” - (ILC EDMS D*899715)
http://ilc-edmsdirect.desy.de/ilc-edmsdirect/document.jsp?edmsid=*899715 .
[CFS06] “LHC Fire Safety Description” - (ILC EDMS D*899755) http://ilc-edmsdirect.desy.de/ilcedmsdirect/document.jsp?edmsid=*899755 .
79
5 Cost Studies
5.1 Purpose
This section describes the studies that have been made so far on the cost differentials that would
result from the changes of the design baseline that are presented in an earlier part of this proposal
document. The primary focus of this costing exercise, at this stage, is to understand and support the
cost changes in the context of overall evaluation of the design changes. Therefore, no re-evaluations
are considered for the costing numbers that were already presented in the RDR. The results
presented here, together with those from risk assessment, after due review process, would then
form the basis for the baseline assumptions in TDP2, where the full technical design and cost
estimating work will be built upon.
5.2 Methodology
Unit costs of individual components: The same unit cost numbers as RDR are used, whenever the
same components are used in the proposed new baseline scheme. Exceptions are the Klystron
Clusters and Distributed RF Systems (DRFS) for the main linacs where radically new components
which did not appear in RDR are introduced. For these, the unit cost numbers were evaluated for
each of the new components and are used to estimate the relevant system cost. Otherwise, no new
unit costs are generated, and relatively straightforward updates from the RDR estimate are made,
whenever possible. Parts of this study include some scaling or parameterization from RDR unit costs
or summary costs.
Technical systems and conventional facilities: The ILC RDR estimate had a single design and a single
cost estimate for the technical elements, such as power supplies, magnets, RF components, vacuum
systems, control systems and others. However, for the conventional facilities and civil construction,
the RDR studies were made on the three regional sample sites, separately, to account for effects of
differing geologies and geographies and differing conventional construction and facilities designs.
However, since they were more complete than for the Asian and European Conventional Facility
Estimates, only the Americas CFS estimates were used for these cost studies for SB2009.
The estimates for the SB2009 studies address two technical approaches and configurations for the RF
power source systems for the main linacs, i.e. the Klystron Clusters and DRFS. They are both
considered for each of the three regional sample sites. It is noted, however, that for these
considerations, there are some known inter-dependence between a particular site and a preferred
technical approach or configuration.
Evolution of the ILC Estimate:
Starting with the published Reference Design Report, the evolution through corrections and
modification of the RDR estimate is illustrated. This includes the Sendai 2008 decisions to change to
a 6.4 km racetrack Damping Ring and to a single-stage Bunch Compressor for the RTMLs and the
various combinations of options under consideration under SB2009 for the new baseline design. The
estimates quoted are intended to allow examination and comparison of various elements within the
proposed design changes in a more coherent fashion. In many cases, the design changes can be said
to have cost impacts in an "orthogonal manner" which can be factored. For instance, adopting a
single tunnel and Klystron Cluster/DRFS for the Main Linac does not affect the DR configuration or
associated cost differentials. However, exceptions also exist. For example, the choice of a 3.4 km
Damping Ring may make it difficult to adopt the Full Power (2625 bunches/train) option for the ML.
80
5.3 Scenarios Being Estimated
Due to cost confidentiality concerns, this section will not quote the estimates beyond the corrections
or adjustments to the RDR estimate, but will contain a short description on each. Members of the
ILC-GDE Cost Management Group may access the password protected version including cost
numbers at
https://www-ilcdcb.fnal.gov/estimates/SB2009/SB2009-Ch-5-Cost-Studies-with-Costs.doc
Note that 1 ILCU = $ 1 (2007)
6,618 M ILCUs – RDR estimate – average of 3 regional estimates
6,677 M ILCUs – RDR with Americas Regional CFS estimate
X,XXX M ILCUs (-66 M ILCUs) – RDR with Correction to Americas Regional CFS estimate,
corrected shaft base cavern volumes and outsourced civil engineering,
and added RTML invert floor.
X,XXX M ILCUs (-72 M ILCUs) – add Value Engineering for higher ΔT Cooling Water sys for ML only.
X,XXX M ILCUs (+X M ILCUs) – at Sendai, March 2008 –
changed DR from 6.7 km hexagonal  6.4 km racetrack (need updated magnet # and costs):
Wigglers 80  88 per ring, RF remained at 18 per ring, other DR magnets remained same,
RTML and e+ Source transfer lines are extended to mate with racetrack DR (long axis ⊥ ML).
X,XXX M ILCUs (-XX MILCUs) – at Sendai, March 2008, adopted Single-stage Bunch Compressor for
RTML. This change was enabled by the shorter bunch length from the racetrack DR.
X,XXX MILCUs (-XX M ILCUs) – removed 4*394 m of empty tunnel for expansion for energy margin,
including e- beam tunnel, e- service tunnel, e+ beam tunnel, and e- tunnel.
This was done to allow more direct comparison with the SB2009 estimates and drawings.
X,XXX M ILCUs – “modified RDR estimates without SB2009 considerations”: 250x250 GeV, 2 tunnels,
4.5 meter diameter, full # bunches, full complement of klystrons and modulators,
6.4 km hexagonal DR: 88 wigglers, 18 RF per ring, single-stage Bunch Compressor for RTML,
e+ flux concentrator, dE(undulator) = 3 GeV, compensated in e- ML,
no allowance for energy margin, no extra tunnel for drift for e+ timing.
X,XXX M ILCUs - SB2009 Updates to RDR (2 tunnels – 4.5 m dia.):
3.2 km DR, low Power, travelling focus, full complement of klystrons, Central Complex,
e- flux concentrator replaced by QWT, dE(undulator) = 4.1 GeV, compensated in e- ML,
no allowance for energy margin, no extra tunnel or drift for e+ timing.
X,XXX M ILCUs – SB2009 Updates to RDR (2 tunnels – 4.5 m dia.) with 3.5 % Energy Margin:
3.2 km DR, low Power, travelling focus, full complement of klystrons, Central Complex,
e- flux concentrator replaced by QWT, dE(undulator) = 4.1 GeV, compensated in e- ML,
no extra tunnel or drift for e+ timing.
X,XXX M ILCUs– RDR - Low Power: ½ klystrons and modulators (longer pulse), with travelling focus,
2 tunnels – 4.5 meter diameter,
6.4 km DR, e- flux concentrator, dE(undulator) = 3.0 GeV compensated in e- ML,
no allowance for energy margin, no extra tunnel or drift for e+ timing.
X,XXX M ILCUs – RDR – Low Power – with 3.5 % Energy Margin: 2 tunnels – 4.5 meter diameter,
with ½ klystrons and modulators (longer pulse), with travelling focus,
6.4 km DR, e- flux concentrator, dE (modulator) = 3.0 GeV compensated in e- ML,
no extra tunnel or drift for e+ timing.
81
X,XXX M ILCUs – RDR – Low Power with SB2009 Updates: 2 tunnels – 4.5 meter diameter,
with ½ klystrons and modulators (longer pulse), with travelling focus,
3.2 km DR: 88  32 wigglers, 18  8 RF per ring, Central Complex,
e- flux concentrator replaced by QWT, dE(undulator) = 4.1 GeV, compensated in e- ML,
no allowance for energy margin, no extra tunnel or drift for e+ timing.
X,XXX M ILCUs – RDR – Low Power with SB2009 Updates – with 3.5 % Energy Margin:
2 tunnels – 4.5 meter diameter,
remove ½ klystrons and modulators (longer pulse), with travelling focus,
3.2 km DR: 88  32 wigglers, 18  8 RF per ring, Central Complex,
e- flux concentrator replaced by QWT, dE(undulator) = 4.1 GeV, compensated in e- ML,
no extra tunnel or drift for e+ timing.
X,XXX M ILCUs – Klystron Cluster – 1 tunnel – 4.5 m dia. – full Power: 6.4 km DR,
no Central Complex, e+ flux concentrator, dE(undulator)=3 GeV, compensated in e- ML,
no allowance for energy margin, no extra tunnel or drift for e+ timing.
X,XXX M ILCUs – Klystron Cluster – 1 tunnel - 4.5 m dia. – full Power – with SB2009 Updates:
3.4 km DR: 88  32 wigglers, 18  8 RF per ring, travelling focus, with Central Complex,
e- flux concentrator replaced by QWT, dE(undulator) = 4.1 GeV, compensated in e- ML,
no allowance for energy margin, no extra tunnel or drift for e+ timing.
X,XXX M ILCUs – Klystron Cluster – 1 tunnel – 4.5 m dia.– full Power
– with SB2009 Updates and 3.5 % Energy Margin
3.4 km DR: 88  32 wigglers, 18  8 RF per ring, travelling focus,
with Central Complex, no extra tunnel or drift for e+ timing,
e- flux concentrator replaced by QWT, dE(undulator) = 4.1 GeV, compensated in e- ML.
X,XXX M ILCUs – Klystron Cluster – 1 tunnel - 4.5 m dia. – Low Power:
½ klystrons and modulators, travelling focus,
6.4 km DR, no Central Complex,
e+ flux concentrator, dE(undulator)=3 GeV, compensated in e- ML,
no allowance for energy margin, no extra tunnel or drift for e+ timing.
X,XXX M ILCUs – Klystron Cluster – 1 tunnel – 4.5 m dia. - Low Power – with SB2009 updates:
½ klystrons and modulators, travelling focus,
3.4 km DR: 88  32 wigglers, 18  8 RF per ring, travelling focus, with Central Complex,
e- flux concentrator replaced by QWT, dE(undulator) = 4.1 GeV, compensated in e- ML,
no allowance for energy margin, no extra tunnel or drift for e+ timing.
X,XXX M ILCUs – Klystron Cluster – 1 tunnel – 4.5 m dia. – Low Power
– with SB2009 updates and 3.5% Energy Margin:
½ klystrons and modulators, travelling focus,
3.4 km DR: 88  32 wigglers, 18  8 RF per ring, travelling focus, with Central Complex,
e- flux concentrator replaced by QWT, dE(undulator) = 4.1 GeV, compensated in e- ML,
no extra tunnel or drift for e+ timing.
X,XXX M ILCUs – DRFS – 1 tunnel – 5.2 m dia. – full Power: 6.4 km DR,
no Central Complex, e+ flux concentrator, dE(undulator)=3 GeV, compensated in e- ML,
no allowance for energy margin, no extra tunnel or drift for e+ timing.
X,XXX M ILCUs – DRFS – 1 tunnel – 5.2 m dia. - full Power – with SB2009 Updates:
3.4 km DR: 88  32 wigglers, 18  8 RF per ring travelling focus, with Central Complex,
e- flux concentrator replaced by QWT, dE(undulator) = 4.1 GeV, compensated in e- ML,
no allowance for energy margin, no extra tunnel or drift for e+ timing.
82
X,XXX M ILCUs – DRFS – 1 tunnel – 5.2 m dia. – full Power – w SB2009 Updates & 5 % Energy Margin
3.4 km DR: 88  32 wigglers, 18  8 RF per ring, travelling focus, with Central Complex,
e- flux concentrator replaced by QWT, dE(undulator) = 4.1 GeV, compensated in e- ML,
no extra tunnel or drift for e+ timing.
X,XXX M ILCUs – DRFS – 1 tunnel – 5.2 m dia. - Low Power: ½ klystrons and modulators,
travelling focus, 6.4 km DR, no Central Complex,
e+ flux concentrator, dE(undulator)=3 GeV, compensated in e- ML,
no allowance for energy margin, no extra tunnel or drift for e+ timing.
X,XXX M ILCUs – DRFS – 1 tunnel – 5.2 m dia. – Low Power – with SB2009 updates:
½ klystrons and modulators, travelling focus,
3.2 km DR: 88  32 wigglers, 18  8 RF, including Central Complex,
e- flux concentrator replaced by QWT, dE(undulator) = 4.1 GeV, compensated in e- ML,
no allowance for energy margin, no extra tunnel or drift for e+ timing.
X,XXX M ILCUs – DRFS – 1 tunnel – 5.2 m dia. - Low Power – w SB2009 updates & 3.5% Energy
Margin: ½ klystrons and modulators, travelling focus,
3.2 km DR: 88  32 wigglers, 18  8 RF, including Central Complex,
e- flux concentrator replaced by QWT, dE(undulator) = 4.1 GeV, compensated in e- ML,
no extra tunnel or drift for e+ timing.
5.4 Issues with the Cost Study
Some elements of this cost study have not yet been completed. Updated quantities, parameters,
and costs for conventional magnets for the hexagonal Damping Rings were not available – the RDR
estimates were used. There is still some uncertainty on the extent and cost of the pulse distribution
system for the DRFS. The positron timing drift for self-replacing positron bunches in the DR is quoted
as 416 meters for the Central Injector Complex with 3.2 km DR. However, subsequent calculation has
shown the drift length should be corrected to 462 meters. The cost impacts have not yet been
updated to account for this small correction. It is also noted that the positron timing drift of 462
meters (or even 416 meters) would cover the 401 meters needed for the 3.5 % energy margin for the
Klystron Cluster approach. However, the similar positron timing drifts for 5% energy margin for the
DRFS and for 3.5% energy margin for the RDR would be approximately 2.07 km and 1.35 km,
respectively. This is because the 5% energy margin for DRFS passes the boundary for another orbit in
the 3.2 km DR and the positron timing drift length is driven by quantized units of ½ of the
circumference of the larger 6.7 km DR. Rather than adding such long tunnels with mostly empty drift
space (FODO quads, trim magnets, instrumentation, and vacuum pipe, plus the cryomodules to
produce the energy margin), it is hoped that a 4.0 % energy margin, which fits within the 462 m
positron timing drift, will be within the uncertainty in machine performance to the recommended 5%
energy margin for the DRFS. Otherwise, we would expect that another more active solution could be
found to produce the self-replacing feature for the positron bunches in the DR, rather than just long
drifts. Although these evaluations were not completed in time for the deadline for this written
document, we will continue to study them and post updated versions when available.
New Electrical Power estimates have not being done completely. There has not been an electrical
engineer assigned to the ILC-GDE CFS team over the last two years. An electrical engineer has
recently been assigned at Fermilab. Only the impact on the electrical power equipment cost
estimate for the Low Power option for the Klystron Cluster has been evaluated at this time. Similarly,
due to lack of available personnel, the impacts on the Cryogenics estimates have only been evaluated
in a parametric scaling manner for these new scenarios. It is unlikely that the impact in these areas
will be fully evaluated for consideration by the SB2009 studies. However, it is necessary that
83
adequate support in all Area, Technical, and Global Systems, including electrical and cryogenic
engineering, will be available for the Technical Design Phase.
84
6 Risk Analysis
6.1 Risk Assessment – the process
Risk is defined as the probability of failure. The risk assessment process is intended to evaluate the
impact a given failure has on the project and the likelihood of that failure. A typical risk scoring
matrix approach (GAO-09-3SP, chapter 14) considers 6 kinds of failure: basic technology,
engineering, production yield, product reliability, existence of a viable backup, and schedule. The
project can respond to perceived risk at any time, and it is generally accepted that the penalty for
doing so increases with time. For the TDP-1 evaluation of risk we should consider only the first 2 out
of the list above: basic technology and engineering. (The remaining four kinds of failure are more
project-focused and are to be assessed as part of an implementation plan.) Commonly, ‘basic
technology’ refers to the scientific underpinning of the subsystem in question and ‘engineering’
refers to the deployment of a given technology to a specific application. Often, engineering related
efforts are not initiated until basic technical and scientific evaluations and tests are complete, and
the two sets of scores naturally represent a single sequence and can be merged.
The assessment of risk is derived from a series of simple questions which refer to status and plans at
the nominal ‘point in time’. The questions focus directly on the development life cycle of a given
subsystem or application of technology and it is assumed that risk is roughly proportional to the
scope of the needed R & D effort. The anticipated penalty is based on how the project would respond
and apply a mitigation strategy once failure is evident or the risk becomes too great. Both the risk
(probability) and penalty (cost of responding to failure) must be considered in order to gauge the
impact. It is the 'impact' which is recorded and summarized in the register.
During the development of the Reference Design a project schedule or timeline from the release of
the RDR to construction completion was made and RISK was to be evaluated at several intermediate
steps. Now it only makes sense to define and evaluate risk at two points in time, i.e. today (transition
between TDP1 and 2) and at the end of TDP2, (T1 in previous terminology of the RDR ‘Risk Register’:
http://ilc-edmsdirect.desy.de/ilc-edmsdirect/document.jsp?edmsid=D0000000*872285), because we
have some projection of level of support for R&D, including resource estimates and technical
milestones.
Adopting reference points at the end of TDP1 and TDP2 means the register contents are limited to
the perception of risk at or before that time and the impact to the project as assessed at that time.
The project-wide comprehensive process of estimating risk is a task for TDP2 and is beyond the scope
of what is described in this section. Here, we describe only the process which will be used to gauge
the risk of SB2009 compared to the Reference Design. First the concerns associated with each
SB2009 Working Assumption are identified. These are listed in table 1. Then each is evaluated
following the steps, 1) to 7), outlined below. The risk evaluation effort is underway; in this document
we will only show results from steps 1) and 2).
1)
score each component based on achievements to date and the expectation of effort to be
applied during TDP-2 using the following questions and scores. The expectation of effort to be
applied should be consistent with that listed in the ‘ILC R & D Plan for the Technical Design
Phase’.
a.
Basic Technology:
i. Is the scientific underpinning of the subsystem in question established?
ii. Specifically, “What remains to be done to establish it at this point in time?”
85
b.
Scoring question
Score
Within the state of the art? Score =
0
One year advancement with minimal resources required. (e.g.: no
beam test facility or high power experiments required at this point)
1
Two to three years advancement with moderate resources required.
(e.g. beam test facility or high power experiments may be required)
2
More than three years advancement with substantial resources. (e.g.
beam test facility or high power experiments are definitely required)
3
New technology required; development cycle unknown
5
Engineering
i. Is the engineering in support of the specific application complete?
ii. Specifically, “At what stage of the engineering life-cycle is the project at this point in
time?”
Scoring question
Score
Fully tested, completed production - units on hand?
0
Prototype exists and has been tested
1
Hardware and software development needed
2
Detailed design underway, development task effort not 'scoped'
3
Concept defined, engineering effort scope not understood
5
2)
Answer the scoring question with a one or two sentence summary of the planned effort.
Generate a score.
3)
Develop a practical mitigation strategy model. For example, what would the project do if
progress was deemed unsatisfactory until the end of TDP-2?
4)
Estimate the cost for the mitigation effort, using costing guidelines similar to those used for the
RDR.
5)
Roll the resulting scoring and associated mitigation costs up to create a summary 'risk
assessment' to be summarized and entered at the top level of the register.
6)
Review the most serious register elements in detail to ensure the scoring, mitigation strategy
and costing have been done consistently according to basic guidelines.
86
6.2 Risk Assessment of SB2009 working assumptions
Table 1 shows concerns that involve new or additional risk due to SB2009. The concerns are
identified by the associated WA number and a sub-category risk ID#.
Table 1: Top-level entries in the SB2009 ‘Risk Register’ showing the identification of concerns and
assignment of identifying numbers.
Evaluation of Incremental Risk due to SB2009 - Basic Technology and Engineering
SB2009 Working Assumption List
Risk or Concern: R and D
results / design work
indicate:
1
A Main Linac length consistent with an
average accelerating gradient of
31.5 MV/m and maximum operational
beam energy of 250 GeV, together with a
High-Level RF (HLRF) distribution scheme
which optimally supports a spread of
individual cavity gradients.
the choice of a reduced
gradient specification. (Note:
the operational gradient
needs to be re-evaluated as
part of the risk assessment
process)
2
A single-tunnel solution for the Main
Linacs and RTML, with two possible
variants for the HLRF
Infrastructure requirements
are not adequately met with
single tunnel
Klystron cluster scheme (KCS)
a smaller, lower power
system (full power system
performance inadequate)
WA
number
a
risk
ID #
Specifically:
1 lower gradient spread
2
1
2
3
4
1
increased power overhead
Radiation shielding
life safety/egress
availability
access to equipment
power handling capability
2 control flexibility
3 breakdown failure modes
1 component cost
2 redundancy schemes
3 maintenance
b
3
4
Distributed RF Source scheme (DRFS)
higher cost (expected cost
reductions not achieved)
Undulator-based positron source located at
the end of the electron Main Linac (250
GeV), in conjunction with a Quarter-wave
transformer as capture device
degraded beam quality
A lower beam-power parameter set with
the number of bunches per pulse reduced
by a factor of two (nb = 1312), as compared luminosity 'recovery' is only
to the nominal RDR parameter set.
partially successful
instabilities do not 'scale' per
prediction; threshold for
onset of instability is too
close to low power ring
parameters
6
Reduced circumference Damping Rings
(~3.2 km) at 5 GeV with a 6 mm bunch
length.
Single-stage bunch compressor with a
compression
of 20 and electron
Integration ofratio
the positron
7
sources into a common “central region
congestion of technical
beam tunnel”, together with the BDS,
components and
resulting in an overall consolidation of civil infrastructure
5
emittance dilution too high
4 Radiation shielding
1 energy spread
2 emittance dilution
3 polarization
1 backgrounds / BSM
2 traveling focus
1 instability thresholds
1 compression performance
installation predicted to take
1 longer
2 maintenance access degraded
These are the specific failures for each WA that have been identified and evaluated in order to show
the difference in risk with respect to the Reference Design. Thus, for example, the risk associated
with achieving the nominal accelerating gradient is not listed since it is basically the same as in the
Reference Design, (even though the impact on the project might be different following the adoption
of the above WA). But the risk associated with the new HLRF distribution schemes is listed.
87
Table 2 shows suggested answers indicating present (TDP1) status of the development process of the
subsystem. The left hand side of the table carries over the WA and risk ID# from table 1 and with the
suggested responses on the right hand side.
Table 2: Results of ‘step 2’; answers to the scoring questions for the concerns listed in Table 1.
WA
number
risk
ID #
Specifically:
1 lower gradient spread
1
a
4 Radiation shielding
1 energy spread
moderate effort simulation and analysis needed
2 emittance dilution
moderate effort simulation and analysis needed
3 polarization
minimal effort analysis needed; much already done
Very little can be done beyond simulation. Have to wait for testing to be done
upon completion of the project, during commissioning; present status calls for
some design work
Very little can be done beyond simulation. Have to wait for testing to be done
upon completion of the project, during commissioning; present status calls for
some design work
increased power overhead
Radiation shielding
life safety/egress
availability
access to equipment
power handling capability
2 control flexibility
3 breakdown failure modes
1 component cost
2 redundancy schemes
3 maintenance
b
3
1 backgrounds / BSM
4
2 traveling focus
5
1 instability thresholds
6
1 compression performance
installation predicted to take
1 longer
7
beam testing at FLASH to prove high gradient spread operation; controls
software development needed
beam testing at FLASH to prove specified overhead is functional;controls
software development needed
Prototype tests of radiation shielding needed
Design effort starting; concepts approved
Conceptual evaluation underway, not complete
detailed design needed, concepts satisfactory
full power tests are required to demonstrate power handling
work at FLASH is required to show that controls flexibility is adequate; software
development needed
work needed to catalog failure modes; hardware development must follow
Conceptual evaluation complete, detailed design of cost-effective HLRF
components to start
Conceptual evaluation complete, detailed design of cost-effective HLRF
components to start
concept exists, hardware layout to be designed
Radiation effects on high voltage/high power electronics largely unknown;
detailed design to start
2
1
2
3
4
1
2
Suggested answer to scoring question:
2 maintenance access degraded
moderate effort simulation and analysis needed to show low I system will work;
RD at CESR and other labs is underway
moderate effort simulation and analysis of low emittance transport tolerances
needed
installation concepts primitive; significant planning effort needed
detailed design of maintenance access needed
It is useful by way of explanation to elaborate on example entries from the table: 1) WA 2a.1, ‘the
risk that the power handling capability of the KCS does not meet expectations’. During TDP2, ‘full
power tests are required to demonstrate power handling’. As noted in section 4, these studies have
started and are expected to proceed during TDP2. 2) WA5.1, ‘the risk that collective instabilities do
not scale per prediction, namely, the threshold for onset of instability is too close to SB2009 ring
parameters’. During TDP2, we believe ‘moderate effort simulation and analysis is needed to show the
low I system will work’. R & D on this is underway at several labs, as described in section 4. For each
entry in the table, the general status of the development process is reflected in the appropriate
proposal text section.
88
Appendix A. Availability Task Force Report
A.1 Introduction and Goal
This appendix is the preliminary report of the availability task force. This task force was formed to
find a reasonable way to make sure that SB2009 met the availability requirements of the ILC. Those
requirements are unchanged from those used in the RDR:



There should be 9 months of scheduled running and 3 months of scheduled downtime for
maintenance and upgrades.
Total unscheduled downtime should be less than 25% of scheduled runtime.
In our simulation we allow only 15% unscheduled downtime. The other 10% is held as
contingency for things that were missed in the simulation or for major design or QA
problems.
So far the task force has concentrated on the use of a single tunnel for the linac using either the DRFS
or Klystron Cluster RF systems. This is one of the biggest cost savings in SB2009 and the biggest
availability concern. That study is relatively complete. Other changes included in SB2009, (the central
region integration and positron source changes), have had a more cursory examination and look OK,
but more work is needed.
We took a three pronged approach to the problem.



We refined the design of the RF systems to provide the goal availability. This included making
some components redundant, ensuring the DRFS klystrons could be replaced rapidly during a
scheduled maintenance day and changing the run schedule from a 3 month long downtime
to two one-month downtimes with a 24 hour preventive maintenance period scheduled
every two weeks during the run.
We gathered and updated information on reliability of components (mean time between
failures)
We integrated all this information into AVAILSIM, the program which simulates breakdowns
and repairs of the ILC to learn how much downtime there was and what the causes were. We
then adjusted some inputs (length of preventive maintenance period, energy overhead and
longer MTBFs) to achieve the required availability.
The availability task force has only studied the direct effects of the SB2009 changes on the
availability. Other effects of the change to a single tunnel such as fire safety; need for space to install
extra equipment in the accelerator tunnel; cost; installation logistics; radiation shielding of
electronics and the effect of residual single event upsets; and debugging of subtle electronics
problems without simultaneous access to the electronics and beam have been and is outside the
scope of this appendix.
Note that the important result of our work is not the availability we have achieved but rather the
design that was necessary to do so. The simulation acts as a guide that allows us to assess design and
component performance changes. The availability estimate obtained by iteration of the inputs
through this process was actually pre-determined as a goal before we started work. That said,
Section 2 gives the results of the simulations and summarizes the conclusions, Section 3 describes
the basic machine design we modelled with emphasis on availability, Section 4 describes many of the
ingredients that were needed to achieve the desired availability, Section 5 gives some details about
how the simulation was done, and section 6 describes some outstanding issues.
89
A.2 Results and Summary
Our results are preliminary for several reasons. Due to time constraints we have not updated the
inputs to AVAILSIM to the full SB2009 design. While the linac modelling is fairly accurate the 6 km
DRs and the injectors are not in the BDS tunnel for most of the simulation runs, (see Central Region
Integration section). There are also several details we are not pleased with and hope to improve
(such as the cryoplants and AC power disruptions being the largest causes of downtime or the long
recovery times needed after a scheduled preventive maintenance day). Nevertheless, we believe the
model adequately represents the differences in availability of the various machine configurations we
studied.
Our main result is shown in Figure A.1 which for four different RF schemes shows the simulated
unscheduled downtime as a function of the main linac energy overhead. This is a very important
figure, so it is worth making very clear what is meant by it.
Figure A.1: The unscheduled downtime as a function of the main linac energy overhead.

An example of the meaning of the horizontal axis is that for 250 GeV design energy and 10%
energy overhead, the beam energy, if everything were working perfectly, would be 275 GeV.
The plot is mainly relevant when we are trying to run at the full design energy. If for physics
reasons we are running at lower energies, then the effective energy overhead is much
greater and downtime due to inadequately working RF systems would be much less than
shown. The steep rise at the left side of the plot comes about because a small number of
linac accelerator components, (RF sources, klystron or modulator, or cryomodule internals);
fail quite quickly after starting up. Repair takes time, either because entry to the linac tunnel
is required (distributed RF system) or because it simply takes time to do the work (klystron
90







cluster / reference design). Also, a small number of cryogenic accelerator components, such
as cavity tuners and power couplers are predicted to fail and these require very long
shutdowns to warm up and fix. Both of these are indicated in the input table showing
assumed MTBF’s (Table A.1), below.
Four different accelerator configurations were simulated resulting in the 4 lines shown. The
general trend for each line is that as energy overhead is decreased, the unscheduled
downtime increases. This is expected as less acceleration related items need to fail before
the design beam energy cannot be reached and hence the accelerator must be shut down for
unscheduled repairs.
Note that the three one linac tunnel designs all have the same unscheduled downtime when
there is 20% energy overhead. This is such a large energy overhead that failures of
components whose failure reduce the energy cause no downtime. The residual 14.5%
downtime is caused by things like magnet power supplies, controls and AC power
disruptions.
For illustrative purposes, a horizontal line is drawn 1% above the points at 20% energy
overhead. This represents what the total unscheduled downtime would be if 1% downtime
were caused by broken components (e.g. klystrons or cavity tuners) making it so the design
beam energy cannot be reached. One can then see how much energy overhead each design
needs to avoid causing this much downtime.
At high energy overhead, the 1 tunnel designs have about 1% more downtime than the 2
tunnel design. This represents a doubling of the downtime caused by linac components and a
50% increase for the RTML components that share the linac tunnel. This increase is expected
as a typical mean time to repair for a component in a support tunnel is 1-2 hours. When that
component is in the accelerator tunnel (as is the case for the 1 tunnel designs) an extra hour
is added to allow radiation to cool down and a second hour is added to allow the accelerator
to be secured, turned back on and magnets standardized. Note that in studies done for the
RDR, the total downtime doubled when we went from two tunnels to one. What saves us
from that fate in this study is that only the linac was changed from two tunnels to one. In all
other areas devices like power supplies are modelled as accessible when the beam is on.
The one tunnel 10 MW design degrades fastest with decreasing energy overhead probably
due to the 40k and 50k hr MTBFs assumed for the klystron and modulator.
DRFS does better probably due to the redundant modulator and 120k hour klystron MTBF
assumed.
KCS does still better due to the ability to repair klystrons and modulators while running and
the internal redundancy built into a cluster. (A single klystron or modulator can die and the
cluster can still output its design RF power.)
The conclusion that can be drawn from Figure A.1 is that either approach (Klystron Cluster or the
Distributed RF System) would give adequate availability with the assumptions shown below in Table
A.1. The Distributed RF System requires about 1.5 percent more energy overhead than the Klystron
Cluster Scheme to give the same availability for all other assumptions the same. This small effect may
well be compensated by other non availability related issues. With the component failure rates and
operating models assumed today, the unscheduled lost time integrating luminosity with a single
main linac tunnel is only 1% more than the two-tunnel RDR design given reasonable energy
overheads. Note that all non-linac areas were modelled with support equipment accessible with
beam on.
Figures A.2 and A.3 show details about what area of the accelerator and what technical systems
caused the downtime in a typical simulation run (KCS with 4% energy overhead). Note that the
unscheduled downtime is distributed over many areas and technical systems. No one thing
dominates. In particular, the linacs and long transport lines of the RTML cause 20% of the total
91
downtime of 15% or a total of 0.20*0.15 = 3% of the scheduled running time. Note that the cryogenic
system plants and site power are the technical systems that cause the most downtime. Both of these
are modeled as single components in the simulation with availabilities based on aggressive estimates
from experts familiar with performance of such systems at existing labs. They are both worth further
study.
Total
RF power sources
4%
Cryo
2%
Magnets
5%
site power
8%
Vacuum
6%
PS + controllers
11%
AC power
3%
controls
11%
Global controls
4%
Diagnostic
0%
General
Recovery
13%
cryoplants
19%
Water
system
9%
RF structure
5%
Figure A.2: This figure shows the distribution of the downtime by area of the accelerator for a typical
simulation run (KCS with 4% energy overhead). The downtime fractions shown are percent of the
total downtime of about 15%. So the actual downtime caused by the cryoplants is 19% of 15% =
2.8%. ‘General Recovery’ is the excess, (beyond time nominally allotted), time spent recovering from
scheduled maintenance days and is lumped because it cannot be directly attributed to a particular
area.
92
e- turnaround
e+ extraction
Global
e+ Total
1% e+ compressor
1%
controls
turnaround
e4%
2%
1%
compressor
e- extraction
2%
1%
e+ transport
General
3%
Recovery
e- transport
13%
3%
e- DR
8%
e+ DR
5%
Cryo plants
19%
e- linac
8%
Site power
9%
e+
source
7%
e- BDS
2%
e+ BDS
1%
e+
linac
6%
e- source
4%
Figure A.3: This figure shows the distribution of the downtime by technical area for a typical
simulation run (KCS with 4% energy overhead). The downtime fractions shown are percent of the
total downtime of about 15%. As with Figure A.2, ‘General Recovery’ shows the unexpected
downtime incurred by recovery from failures that happened during scheduled maintenance periods.
As explained in the introduction, our goal was to find design ingredients that achieved the desired
availability goal. This section shows we have achieved that goal. The next two sections describe the
ingredients used to attain that goal. You will see it requires different construction and operations
methods than are typically employed for HEP accelerators. Methods used by light sources are
needed. Component mean time between failures must be comparable to the best achieved at any
accelerator and some are factors of more than 10 better than those achieved at major HEP labs.
Preventive maintenance is crucial as is built-in redundancy.
Preliminary conclusions of impact of single main linac tunnel on availability and Recommendation are
as follows:
•
The assumptions made to obtain the desired availabilities for all designs are quite
aggressive and considerable attention will have to be paid to availability issues during
design, construction and operation of the ILC to achieve the simulated availabilities.
•
The RF power system as described in the RDR is unsuitable for a single linac tunnel design
as there is a significant decrease in availability without further improvements in MTBF’s, an
increase in energy overhead and/or changes in maintenance schedules. Presumably this
could be the focus of additional study and R & D.
93
•
There are two alternate RF power system designs proposed for single tunnel linac
operation. (The Klystron Cluster and the Distributed RF System). Either approach would
give adequate availability with the present assumptions. The Distributed RF System
requires about 1.5 percent more energy overhead than the Klystron Cluster Scheme to give
the same availability for all other assumptions the same. This small effect may well be
compensated by other non availability related issues.
•
With the component failure rates and operating models assumed today, the unscheduled
lost time integrating luminosity with a single main linac tunnel is only 1% more than the
two tunnel RDR design given reasonable energy overheads. Note that all non-linac areas
were modelled with support equipment accessible with beam on.
•
The recommended RF overhead (see Figure A.1: ‘unscheduled downtime as a function of
the main linac energy overhead’) is 3.5% . This corresponds to the nominal availability of
about 15% for the Klystron Cluster HLRF scheme and slightly poorer availability for the
Distributed RF HLRF scheme. The cost impact of this recommendation is summarized in
section 5 of the Proposal.
A.3. Background - changes applied to the Reference Design
In this section we describe the machine model used to evaluate the availability with a focus on
differences with respect to the 2007 Reference Design.
A.3.1 Machine Model
Improvements proposed to the Reference Design are expected to affect the estimated accelerator
availability. In order to estimate availability in the new design, both the basic machine model and the
assumptions that support the simulation, (for example, operations and maintenance models),
needed review. Through this process, the availability of the complex was also re-examined and reoptimized.
The changes foreseen for the new baseline are described in the Overview, Section 2. The part of the
new baseline with the greatest impact on availability is the removal of the support tunnel for the
main linac and RTML. A single tunnel variant, in which no support tunnel space was allocated to any
ILC subsystem, was studied during the development of the Reference Design. The estimated
availability of that variant was shown to be poor unless there were very substantial improvements in
component performance. The configuration studied for the proposal described here allows access to
source, DR and BDS support equipment when the machine is operating. Only the main linac and
RTML support tunnel is removed.
A.3.2 Operations and Maintenance Model
The ILC is scheduled to operate 9 months of the calendar year, with the remaining 3 months
allocated for machine shutdown and preventative maintenance. In the RDR the assumption was for a
single 3—month shutdown. We have explored other operating models where the 3 months of total
scheduled downtime is distributed throughout the calendar year. It should be apparent that less RF
power overhead would be needed, (depending on how long it takes for replacement), if there is a
shorter time between maintenance periods when failed klystrons are replaced. For this report, we
have chosen a model with two 1-month shutdowns per year and one 24-hr preventative
maintenance period every two weeks. All failed klystrons and modulators in the main linac will be
replaced during the 24-hr maintenance days, so 100% of the installed RF power is available
immediately following a maintenance day.
94
Availability estimates made for the Reference Design were based on experience at labs with high
energy colliders. Experience with component performance and operations and maintenance practice
was included in the analysis. For further development of the ILC design, we intend to take advantage
of the knowledge gained in recent years at light sources and ‘factories’, such as the B-Factories. In
general, experience at these installations has been much better than older, more established
facilities. An important part of that improvement is due to changes in managing component lifetime
performance and maintenance. These typically involve a more rigid industry-like quality assurance
process, with a stronger reliance on scheduled downtime and less emphasis on ‘opportunistic
maintenance’, a practice which includes a substantial ad-hoc, rapid response maintenance and repair
effort. Additionally, the use of industrial standard-practice system designs and components (section
4) that have good reliability performance has been assumed in the analysis.
A.3.3 HLRF - Klystron Cluster System (KCS) availability design
The Klystron Cluster System is attractive because it allows relatively long distances, (compared to the
Reference Design), between the power sources and the accelerator. In the RDR, the waveguide
length between the closest cavity and its power source was around 7 m (25 m for the farthest),
whereas for the KCS the farthest distance is around 1200 m. This is made possible because of the
very low loss transmission characteristic of TE01 waveguide which has essentially no electric fields on
the waveguide wall. In the KCS, the total system waveguide power loss is roughly doubled, compared
to the Reference Design, but the total waveguide length is 50 times greater. Very high power TE01
mode systems function well in accelerator applications. The most well-known application is in the
commonly used ‘SLED’ system. Also, long waveguide applications have been developed and tested
for X-band transmission systems where losses in rectangular waveguide can be large. These included
tests of high-vacuum mode-convertor components, bends, pump-outs etc, as described in the system
description in Section 4.6.2.
One of the most attractive aspects of the KCS is the intrinsic simplicity of the accelerator tunnel. It
contains virtually no ‘active’ electrical components and reduced water cooling, compared to the
Reference Design. This makes the KCS ideal for relatively open site topography with rock-based
geology, as in two of the RDR sample sites. Although land use may be restricted at these two sites,
surface transport access is easy and the KCS long waveguide lengths may provide needed flexibility to
achieve ready access to complex equipment.
The impact of the KCS HLRF system on machine availability is considered negligible. Redundancy will
be built in to the low level RF, and the high power systems contain spare sources so most failures do
not cause downtime. The proposed Klystron Cluster System allows continuous access to all active RF
source equipment. Further, the system includes provisions for maintenance of the high power
equipment with little or no impact on the operation of the linac.
The ‘common-mode components’, whose failure would cause extensive downtime, are the
distribution waveguide and the associated bends, vacuum pump-outs, couplers and expansion-joint
hardware and are assumed to be robust against breakdown. The peak surface electric field in the
Circular Tap – Off (CTO) with 10 MW through its WR650 ports (5 MW each) is 1.57 MV/m. This is
lower than in the VTO (variable tap-off) at 5 MW (1.86 MV/m) which has been tested with room
pressure nitrogen with no RF breakdowns. The CTO will most likely be under vacuum which should
make it even more robust against surface breakdown. The L-band vacuum window design, (also used
in the Reference Design), has had extensive running at the ~5MW, 1ms level, where it has a similar
peak surface electric field (1.51 MV/m), although the surface is stainless steel rather than aluminum.
If we scale the observed maximum surface field of 40 MV/m that we found to be safe in X-band
component designs (copper), using the apparent 1/6 power pulse width dependence, we get a rough
95
breakdown threshold of 10 MV/m. Furthermore, for the low power configuration, only the
combining CTO's will see 10 MW, while the CTO's in the tunnel extract only 5-7 MW.
These components are also designed to be robust against surface pulsed-heating. The peak surface
magnetic field for the maximum 340 MW (full current upgrade) flowing through a CTO is ~14.7 kA/m,
on the edge of the tube defining the gap. This should give a pulsed heating temperature rise of ~2.7
degrees C, well below the tens of degrees level known to cause surface damage and be related to RF
breakdown in copper X-band structures. In the tapers and main 0.48m waveguide, the surface
magnetic fields are much lower, and surface electric fields are non-existent (except to the extent that
parasitic modes might be excited).
SLAC and other labs have decades of experience running high power in overmoded TE01 mode
waveguide and cavities, e.g. SLED, BPC, SLED-II, multi-mode SLED-II and X-band transmission. The
mode can only break down by gas ionization, so good vacuum is required.
The two bends required in each main waveguide may be the most risky part of the system, as they
carry the maximum power and must depart from the pure TE01 field pattern. We believe, however,
that suitable bends can be designed, based on past experience with X-band.
It is important to emphasize that any fault (e.g. breakdown or vacuum leak) in the ‘common mode’
distribution system is a single point of failure and will cause downtime. The availability simulation
does not include these faults.
A.3.4 HLRF – Distributed RF System (DRFS) availability design
Availability Issues for DRFS (2-week Operation + Half Current)
There are two main availability issues for DRFS: Radiation damage and replacement of failed HLRF
components.
A large number of klystrons will be needed since one klystron feeds its power to four cavities in DRFS
scheme compared to 26 cavities in RDR. Maintainability of klystrons is a key issue for DRFS. A longer
MTBF for klystrons is required for DRFS availability and a redundant scheme will be adopted for their
power supplies.
One of the main availability issues for DRFS is the radiation damage of components. Almost all DRFS
components will be installed in the tunnel where we need to examine the potential radiation hazard
which is caused by acceleration of the dark current. The data concerning this issue is unfortunately
scarce. At this moment, we follow the considerations given in case of the European XFEL project, and
assume the use of a radiation shield of 10cm-thick concrete and 1cm-thick lead will help reduce
problems caused by radiation.
Another important availability issue is the maintenance of HLRF components, in particular the
klystrons. A large number of klystrons are used in the main linac tunnels, as one klystron feeds its
power to four cavities, as opposed to 26 cavities by one klystron in the RDR. A longer MTBF for
klystrons is expected to be realized with the measures as described in the DRFS section of this
proposal. However, maintainability of klystrons and other components remains to be a key issue. In
the following, the maintenance scenario for the DRFS and the associated workload are examined.
As discussed above, the operation schedule is assumed to consist of a cycle of 2-weeks of continuous
operation (312 hrs) interrupted by one-day maintenance periods. The numbers of failed
components, which require replacement or repair, after each of the 2-week operational period, are
estimated and are summarized in Table A.2 (reproduced from Section 4.6). The MTBF assumed in the
estimates are also indicated.
96
It should be noted that the scheme taken with redundant DC PS and MA modulator will provide
substantially longer "system" MTBFs for the DC power supplies and MA modulators than what are
quoted in Table A.3 (reproduced from Section 4.6). The MTBF quoted for the klystrons is based on
the operational experience of KEKB linac, where 4.7 million hours of cumulative operation time has
been logged at 60 klystron sockets. BI (barium-impregnated) cathode and lower cathode loading,
which assure longer life time of klystron, are adopted for this project.
Table A.2: Estimated number of HLRF components across the entire ILC requiring replacement or
repair after completion of each of the 2-week operational periods, assuming the listed MTBF and
component count.
Component
DC power supply
MA modulator
MA klystron
# of units
requiring
MTBF assumed
replacement
or repair
2
50,000 hours
1.5
70,000 hours
12
110,000 hours
Total # of units
deployed at the ILC
325
325
4225
In order to estimate the human resources that are required during each of the one-day maintenance
periods, the time that it takes to replace each of the components in question has been examined.
Table A.3 shows the summary of this study.
Table A.3: Estimated times for the repair work of DRFS.
Action
Transportation of klystron
Time for unit piece of work
0.5 person-hours / tube
Removal of a failed klystron and
installation of a replacement
klystron
Time for personnel to move
from one point of repair to
another
Replacement of a MA
modulator
Replacement of a DC power
supply
4 person-hours / tube
Rationale
2 persons in 2 hours could bring
8 tubes on one carrier.
2 hours with 2 persons
2/3 person-hours / tube
20 minutes with 2 persons
6.67 person-hours / modulator
27 person-hours / DC power
supply
As a consequence, the human resource that it takes to replace the 12 MA klystrons on each of the
one-day maintenance day would amount to 62 person-hours. Likewise, 1.5 MA modulators will
require 10 person-hours, and 2 DC power supplies will take 54 person-hours to repair.
In summary, the required human resource for HLRF during maintenance is calculated to be 126
person-hours, which corresponds to 16 person-days. With 9 hours to perform the work this will
require 43 people on an average day and perhaps double that on some days due to upward
97
fluctuations in the number of failures. This is likely to be manageable in the light of other work that
would be necessary during maintenance days.
A.4 Key ingredients and their impact - Estimating MTBFs
A.4.1 Estimating MTBF’s
Any simulation is only as good as the raw input data. Actual MTBFs realized in the field for a
specific installation are dependent on many factors, some that can be predicted or
estimated and many that cannot. Consequently, any MTBF estimate projections have
uncertainty.
Availability and Reliability prediction and estimation has been (and continues to be) the
subject of much study by industry and research institutions, and consequently there are
several generally accepted methods for estimating MTBFs, including:
1. Projections based on the technical design using standardized data on individual parts,
design margins, and service usage.
2. Projections based on accelerated life testing, where units are subject to extreme
operating conditions that cause early failures.
3. Comparisons with similar equipment where MTBFs are already known.
4. Actual data on failures from equipment that is installed and operating in the field.
Each method has advantages and disadvantages. For example, projections based on the
technical design do not take into account the various ‘soft’ issues such as quality assurance
processes, which can have a large impact on the in-service reliability. On the other hand,
they are the only option available for new equipment where there is no experience base.
Methods based on testing or actual in-the-field failure rates can provide the most useful
information, especially when estimating MTBFs for equipment with similar operating
conditions and environments that apply to the equipment whose MTBF is being estimated.
However, statistically significant MTBF estimates require enough cumulative testing or inservice operating hours such that multiple equipment failures have occurred, i.e.
significantly more operating hours than the equipment MTBFs.
A.4.2 Process for developing MTBFs for the final models
The simulation program, ‘Availsim’ receives as input a set of component MTBFs that were derived
from operational and in-the-field experience. Since one goal of the simulations was to determine a
set of MTBFs that met the availability criteria for the machine design, these MTBFs were treated as
the ‘Starting MTBFs. The initial Availsim simulations using these MTBFs gave an estimate of the
overall machine availability given these starting MTBFs, the overall machine design, and the various
other input assumptions that have been described previously in this appendix.
The results of the simulations were evaluated against the machine availability criteria. A second set
of MTBFs was generated through ‘educated-guesswork’ based on the gap between the machine
availability criteria and that achieved in the simulation. Typically parts that were contributing more
downtime than most had their MTBFs improved. A second round of Availsim simulations were
performed using the updated set of MTBFs. Additional iterations were performed until the machine
98
availability criteria were met. The Improvement Factors coming from the Availsim simulations (the
ratios between final MTBFs and starting MTBFs) have triggered availability-focused R&D programs
for several technical systems, including HLRF modulators, magnet systems, and Control Systems. As
noted above in the Introduction and Goal section, this is the real output of the process.
It should be noted, however, that there is no unique set of ‘final’ MTBFs, since during the
optimization process the relative MTBFs can be adjusted to apportion total downtime differently
across the various technical systems.
A.4.3 Source of the Starting MTBFs
As noted earlier, the MTBFs comprising the Starting MTBFs for the Availsim simulations were as
much as possible derived from in-the-field experience.
The set of Starting MTBFs use for the RDR simulations was largely drawn from the years of operating
experience at the SLC and PEP-II at SLAC and the Main Injector and Tevatron accelerators at
Fermilab. As such, the set of MTBFs were representative of the availability that had already been
achieved at operating high energy physics accelerator facilities.
For SB2009 simulations, the Availability Working Group proposed reviewing and revising ‘up’ the
Starting MTBFs for individual components so as to use best-in-class achieved MTBFs unless there
would be a clear and significant cost penalty over comparable systems. In particular, the intention
was to take into account recent experiences with the Light Sources and the B-Factories, both of
which had achieved better overall availability than had been achieved at both SLC and the Tevatron.
Similarly, it was desired to take into account industry and commercial MTBF data for Commercial Off
the Shelf (COTS) components. By and large, each of the components fits into one of the following five
categories.

Utility or industrial equipment where there is published MTBF data based on operating
experience of a very large installed base. For example, we have used the IEEE ‘Gold Book’ (IEEE
Std 493™2007) for electrical distribution MTBFs.

Commercial off-the-shelf commodity equipment such as computing systems and networks
where there are high availability solutions routinely deployed.

Commercial low-complexity components.

Specialized or custom-designed equipment that is accelerator specific and where there is
already operating experience. Our MTBF estimates are based on experience from accelerator
laboratories.

Most difficult to estimate are MTBFs for new specialized equipment where as yet there is little
or no operating experience. In these cases, we can only apply good engineering judgment and
where appropriate make comparisons with other existing equipment.
Largely due to time constraints, there have been only limited updates of the Starting MTBFs in the
RDR, and to date only ten of the Starting MTBFs have been revised, leaving others as potential
candidates for revision.
99
A.4.4 MTBF Data from the RDR and SB2009 Simulations
Table A.1 summarizes the MTBF data from both the RDR and the SB2009 simulations. Each row
corresponds to an individual device in the Availsim simulations. The information in the various
columns is as follows:
New starting MTBFs are those used in the initial step of the simulations prior to iteration and
optimization. As much as possible, they correspond to the best actual MTBFs achieved in the
field. It should be noted that some of these New starting MTBFs have been revised upwards
from those used as starting MTBFs in the RDR simulations. The specifics will be described
later in this section.
New ending MTBFs are those obtained from the final iteration of the optimization process.
Collectively they are one set of MTBFs that achieve the desired machine availability in
simulations of the SB2009 machine configurations under the specified criteria. It is important
to note that the same identical New ending MTBF numbers were used for all four of the
SB2009 configurations.
MTBF factors are the ratios of the New ending MTBF to the New Starting MTBF. It indicates
the necessary degree of improvement in MTBF over present field experience in order to
meet the availability goals of the SB2009 machine.
The five right-most columns are actual operational MTBFs achieved at four accelerator
facilities or elsewhere (either accelerators or industry). The highest MTBF in each row was
used as the New starting MTBF for the corresponding component. The colors indicate the
relative improvement factor necessary to achieve the New final MTBFs from the SB2009
simulations. Empty cells indicate that no data was available for the given device from the
corresponding source. Note that several rows have only empty cells, and these cases, the
New starting MTBFs were the result of ‘educated guesswork.’
Changes made to the Starting MTBFs from RDR to SB2009 are as follows (Device #'s refer to Table
A.1):
Magnet Power Supplies: PS_controller, PS_corrector, and PS (Device # 2, 7, 8) were revised based
on operations availability data from the Advanced Photon Source at Argonne, where an MTBF of
500,000 hrs has been achieved over 30,000 operating hours of the ~1600 power converters in the
APS storage ring. As an aside, it should be noted that there is no built-in redundancy in the APS
power converters.
Water Instrumentation: water_instr (Device # 26) was revised based on updated operations
availability data from the Fermilab.
Electrical Utilities: elec_small and elec_big (Device # 27, 28) were revised using data in the IEEE Gold
Book, that includes surveys in-the-field failure rates for electrical utility equipment.
Machine Protection System: MPS_region (Device # 41) was revised based on operations availability
data from the Advanced Photon Source at Argonne.
Components for the Klystron Cluster and Distributed RF Systems: mb_klystron (Device # 13) was
used for the two RDR configurations and for the klystron cluster. DRFS_klystron (Device # 14) was
used for the DRFS.
100
Device #
Table A.1: Starting and Final MTBFs for RDR and SB2009 Availsim simulations. For each item, the
table gives the MTBFs used in the SB2009 simulations along with MTBFs from operating experience
on four accelerators. The color codes indicate the extent of any gap between operating experience
and the MTBFs used in the simulations. The color key is as follows:
Green: MTBF has already been achieved.
Yellow: MTBF is up to a factor 3 lower than used in the simulations.
Orange: MTBF is up to a factor 10 lower than used in the simulations.
Red:
MTBF is more than a factor 10 lower than used in the simulation.
White: No reference data available. Subject for further studies in TDP2.
New
New
starting MTBF
Device
MTBF
factor
1 mttf_electronic_module
1.00E+05
2 mttf_PS_controller
1.10E+06
3 mttf_controls_local_backbone 1.00E+05
4 mttf_magnet
2.00E+06
5 mttf_sc_magnet
3.00E+07
6 mttf_small_magnet
3.40E+07
7 mttf_PS_corrector
1.10E+06
8 mttf_PS
1.10E+06
9 mttf_kicker
1.00E+05
10 mttf_kickpulser
7.00E+03
11 mttf_modulator
5.00E+04
12 mttf_dr_klystron
3.00E+04
13 mttf_mb_klystron
4.00E+04
14 mttf_DRFS_klystron
1.20E+05
15 mttf_cavity
1.00E+08
16 mttf_coupler_intlk
1.00E+06
17 mttf_coupler_intlk_electronics 1.00E+06
18 mttf_mover
5.00E+05
19 mttf_VacP
1.00E+07
20 mttf_VacP_power_supply
1.00E+05
21 mttf_valve
1.00E+06
22 mttf_vac_valve_controller
1.90E+05
23 mttf_fs
2.50E+05
24 mttf_xfrmr
2.00E+05
25 mttf_waterpump
1.20E+05
26 mttf_water_instr
1.30E+05
27 mttf_elec_small
1.60E+06
28 mttf_elec_big
1.60E+06
29 mttf_vac_mech_device
1.00E+05
30 mttf_laser_wire
2.00E+04
31 mttf_wire_scanner
1.00E+05
32 mttf_klys_preamp
1.00E+05
33 mttf_vacG_controller
4.70E+05
34 mttf_cavity_tuner
1.00E+06
35 mttf_cavity_piezo_tuner
5.00E+05
36 mttf_power_coupler
1.00E+07
37 mttf_cryo_leak
1.00E+05
38 mttf_JT_valve
3.00E+05
39 mttf_cryo_big_prob
1.00E+07
40 mttf_target
4.4E+04
41 mttf_MPS_region
3.00E+04
3
3
10
10
1
1
1
3
1
5
1
1
1
1
1
5
1
1
1
1
5
5
30
1
1
3
1
1
5
1
1
1
1
1
1
1
10
1
1
1
1
New ending
MTBF
3.0E+05
3.3E+06
1.0E+06
2.0E+07
3.0E+07
3.4E+07
1.1E+06
3.3E+06
1.0E+05
3.5E+04
5.0E+04
3.0E+04
4.0E+04
1.2E+05
1.0E+08
5.0E+06
1.0E+06
5.0E+05
1.0E+07
1.0E+05
5.0E+06
9.5E+05
7.5E+06
2.0E+05
1.2E+05
3.9E+05
1.6E+06
1.6E+06
5.0E+05
2.0E+04
1.0E+05
1.0E+05
4.7E+05
1.0E+06
5.0E+05
1.0E+07
1.0E+06
3.0E+05
1.0E+07
4.4E+04
3.0E+04
SLC
MTBF
1.0E+05
8.0E+04
FNAL
FNAL
Main
Tevatron Injector
MTBF
MTBF
1.8E+05
5.0E+05
1.1E+05
APS
MTBF
other
MTBF
1.1E+06
2.0E+06
1.6E+06
3.4E+07
4.3E+05
4.3E+05
1.0E+05
6.6E+03
6.4E+04
1.8E+05
1.8E+05
1.1E+05
1.1E+05
1.1E+06
1.1E+06
4.0E+04
5.0E+04
1.7E+05
9.6E+04
9.6E+04
5.1E+05
3.8E+06
1.0E+06
1.9E+05
2.2E+05
1.2E+05
3.0E+04
3.6E+05
3.6E+05
1.3E+05
1.3E+05
6.7E+05
1.6E+06
1.6E+06
4.7E+05
5.1E+05
5.0E+03
3.0E+04
Key: MTBF relative to New final MTBF
Meets or exceeds the New Final MBTF
Less than a factor 5 away
Factor of 5 -10 away
Factor of more than 10 away
101
It is outside the scope of this document to present detailed plans for meeting each of the
improvement factors shown in Table A1. Nevertheless, it is appropriate to briefly describe some
illustrative examples:
Magnet power supplies: It has been shown at APS that significant improvements in power supply
availability can be made through ‘practical’ quality control processes, and rigorous/proactive
preventative maintenance. It is anticipated that additional improvements could be achieved by
incorporating operational lessons into the entire design and implementation life cycle.
R&D is ongoing at SLAC to explore the use of redundant regulators in addition to the more
conventional redundancy of the power modules. The RDR costing already takes redundancy
into account for the magnet power supplies.
Control system front ends and other electronic modules: Already incorporated into the RDR Control
System design concept was a proposal to use the ATCA standard as a means of benefiting from
telecommunications-like approaches to achieving high availability, An ATCA-based solution was
assumed in the RDR costing.
Cooling water flow switches: Table A1 shows a required improvement factor of 30 for flow switches
(Device # 23). By no means should this be interpreted as needing flow switches that are 30-times
less likely to fail. This is representative of cases where availability improvements would be made
through engineering design and integration. For example, flow switches might be avoided completely
on individual magnet cooling circuits. Instead, over-temperature interlock on the magnet windings
could be used to detect loss of cooling.
A.4.5 Ingredients for meeting the required MTBFs
Irrespective of the exact optimization of Final MTBFs or the exact details of the model, it is clear
there will be many challenges to building the ILC accelerator that can operate with the desired
availability. There are many ingredients that must be considered that go beyond availability
considerations in the design of the overall machine and of individual components. Two such
ingredients are the role of Quality Assurance and Preventative Maintenance.
Quality Assurance / Quality Control
There are numerous examples where good reliable designs have been compromised in application by
poor attention to detail in the implementation. Similarly, there are many examples of the benefits of
effective quality control. Since almost every component in the ILC accelerator is duplicated many
times, consistent and well controlled processes will be essential.
Preventative Maintenance
The fundamental principle of Preventative Maintenance is to take advantage of the scheduled down
times in order to increase the availability during the scheduled operating periods by performing work
that preemptively reduces the likelihood of equipment failures. Preventative Maintenance activities
are planned in advance and carefully controlled in their implementation. The types of tasks
considered Preventative Maintenance take a wide range, including servicing of electrical switchgear
and mechanical pumps, to replacement of water hoses when they reach 80% of their nominal useful
life, to preemptive ‘stress testing’ of equipment during a shutdown in order to induce a failure in the
weakest units so they can be removed from service.
102
A.5 Simulations
The simulation results[Avail1] described above were done with the program AVAILSIM developed by
Tom Himel over the past several years. It is a Monte Carlo of the ILC that randomly breaks
components, checks whether the set of presently broken components prevent the ILC from
generating luminosity and if so, schedules repairs. A description of this program was given at the
2007 particle accelerator conference[Avail1] and some of its features are described below.
Each component fails at a random time with an exponential distribution determined by its MTBF.
When a component fails, the accelerator is degraded in some fashion. A klystron failure in the main
linac simply reduces the energy overhead. The accelerator keeps running until this overhead is
reduced to zero. Similarly there are 21 DR kickers where only 20 are needed so only the second
failure causes downtime. Some components such as most magnet power supplies cause an
immediate downtime for their repair.
Each component can be specified as hot swappable (meaning it can be replaced without further
degrading the accelerator); repairable without accessing the accelerator tunnel, or repairable with an
access to the accelerator tunnel. A klystron that is not in the accelerator tunnel is an example of a
hot swappable device.
Without doubt the downtime planning is the most complicated part of the simulation. This should
come as no surprise to anyone who has participated in the planning of a repair day. It is even harder
in the simulation because computers don’t get a gestalt of the situation like humans do. Briefly, the
simulation determines which parameter (e.g. e- linac energy overhead or e+ DR extraction kicker
strength or luminosity) was degraded too much, and plans to fix things that degrade that parameter.
Based on the required repairs, it calculates how long the downtime must be to repair the necessary
items. It then schedules other items for repair, allowing the downtime to be extended by as much as
50 to 100%. Some other issues must also be taken into account:

If an access to the accelerator tunnel is required, one hour is allowed for prompt radiation to
decay before entry. One hour is also allowed for locking up, turning on and standardizing power
supplies.

The number of people in the accelerator tunnel can be limited to minimize the chaos of tracking
who is in the tunnel. For the present work, we have not limited the number of people. We have
estimated that number and it is not unreasonably large, but we still need to use the simulation
to limit that number and see how it affects the results.

The accelerator runs a total of 9 months a year. There are two one month shutdowns for major
maintenance and upgrades each year. In addition, every two weeks there is a 24 hour scheduled
downtime to perform preventive maintenance. This includes work which is explicitly simulated
like replacing DRFS klystrons which have died in the previous two weeks and work which is not
explicitly simulated but which may be need to attain the long MTBFs like replacing pumps which
are vibrating because their bearings are going bad.

Recovery from a downtime is not instantaneous. Things break when you turn them off for the
downtime and the problems are only discovered when they are turned back on. People make
mistakes in hardware and software improvements and these must be found and corrected.
Temperatures change and the ground moves so the beam must be re-tuned. Rather than trying
to model downtime recovery procedures in detail, AVAILSIM simply assumes that the time it
takes to get good beam out of a region of the accelerator is on average proportional to the time
that region was without beam. The constants of proportionality used for each region typically
were 10%, except for the DRs and interaction region, for which 20% was used and simple
103
transport lines for which 5% was used. Altogether, this recovery time on average slightly more
than doubles the downtime due to a repair.
The accelerator we simulated is not exactly the SB2009 design but we expect it is close enough that
the simulation results are meaningful. We will continue our studies and add the remaining SB2009
changes in the near future. The simulated ILC had the following features:









The linac could be in 1 or 2 tunnels
Low power (half number of RDR bunches and RF power)
Three RF systems were simulated for the main linac: RDR, KCS, and DRFS. Short linacs (injectors
etc.) were always simulated with the RDR RF scheme.
Two 6 km DRs in same tunnel near the IR. (Going to 3 km would decrease the component count
and slightly improve the availability for all simulation runs.)
RTML transport lines in linac tunnels
Injectors in their own separate tunnels. (Putting them in the BDS will probably slightly decrease
the availability for all simulation runs.)
The e+ source uses an undulator at end of linac
There is an e+ Keep Alive Source of adequate intensity to do machine development and to tune
up during a downtime recovery.
Injectors, RTML turn-around, DRs and BDS have all power supplies and controls accessible with
beam on. (Pre-RDR 1 vs. 2 tunnel studies had these inaccessible for 1 tunnel.)
A.6 Outstanding issues and concerns
Availability studies, generally, require input on technology, system layout and operations and
maintenance models. It is critically important, therefore, to make sure that the results of these
studies are well documented and accessible to provide guidance for ongoing and future R & D
activities.
The following outstanding availability issues and concerns have been identified. These are important
for both the Reference Design and the proposed baseline.

Integration with sub-system designers: Component MTBF and MTTR performance
requirements have been assigned and tabulated as a by-product of the availability analysis. In
some cases, the performance criteria needed represent an advance beyond that routinely
achieved in large colliders. In this case, industrial best-practice guidelines or light-source
performance has been assumed. For both of these, added design effort, compared to large
colliders, has to be applied.

Estimating impact on cost: Improved MTBF performance has a cost increment. In simple terms,
higher quality off-the-shelf components, with adequate reliability performance, will cost more.

Improving recovery times: The simulation includes a ‘luminosity recovery process’ model,
whereby the time to recover full-spec performance in each subsystem is proportional to the
time without beam, (due, for example, to an upstream component failure). Completed
simulation results indicated this to have a substantial impact, amounting to about half of the
total simulated downtime. The proportionality constant used is based on observations done at
several facilities, and some degree of variability is seen. A consistent approach to improving
system availability includes effort on both component MTBF and the associated recovery
process.
104

Safety – impact on Operations and Maintenance model: Each intervention for maintenance,
including preventive activities and incident response, must be subject to a fully integrated safety
analysis. The simulation includes several aspects, such as radiation cool-down wait-times and an
allowance for proper preparation and entry procedures. It is reasonable to expect additional
constraints.

Accessibility – equipment transport: For repairs within the beamline enclosure, the availability
model includes an estimate of the time to move equipment from a storage area to the accessway. With more detailed underground equipment layout, the model will be updated to provide
a more realistic estimate of transit time.

Radiation effects: Radiation effects may reduce electronic component MTBF substantially. This
is important because the new baseline layout does not provide nearby support equipment
enclosure space for the main linac and RTML. Several accelerator labs have experience with
electronics located near the beamline, notably the PEPII B-Factory and (soon) the LHC. The
availability simulation assumes that electronics is protected from radiation and does not take
into account possible reduced performance.

Initial commissioning: None of the studies include an estimate of initial commissioning effects.

Staffing levels: The model shows, correctly, the balance between component failure rate and
total time to replace or repair. An important ingredient of that balance is the estimate of the
staffing level so that maintenance activities can proceed in parallel. Due to the statistical nature
of component failure, the simulation can provide a good estimate of the required staffing levels.

MTBF data: The basis for the estimate is the reliability data collected from labs and industrial
literature. In many cases, representative estimates are difficult to obtain, so the analysis has
been based on data from only one institute, or is simply an estimate. There are several reasons
why valid MTBF estimates are difficult to obtain, notably the tendency for organizations to
group failures according to who in their organization is responsible for repairs associated with a
given subsystem.
References
[Avail1] "Availability and Reliability Issues for ILC", T. Himel, J. Nelson, N. Phinney, (SLAC) , M. Ross,
(Fermilab) . PAC07-WEYAB02, SLAC-PUB-12606, Jun 2007. 4pp. In the Proceedings of Particle
Accelerator Conference (PAC 07), Albuquerque, New Mexico, 25-29 Jun 2007, pp 1966. Also
in *Albuquerque 2007, Particle accelerator* 1966-1969
105