Re: Question 4 - Straight or curved tunnel?

[Philip Bambade <bambade@xxxxxxxxxxxx>]

Dear colleagues,

I was notified by Peter Tenenbaum that my question below was not 
distributed to the WG1 mailing list. Now it should reach you.

Sincerely yours,
Philip Bambade.

---------------------------------------------------
Philip Bambade       E-mail: bambade@xxxxxxxxxxxx
http://www.lal.in2p3.fr/recherche/delphi/bambade/
Home Institute:                At CERN:
LAL, Campus Universitaire      Delphi Collaboration 
pièce 129,Bât.208,B.P. 34      EP Division 
F-91898 ORSAY CEDEX France     F-01631 CERN CEDEX
Tel. 0033-(0)1-64468370        Tel. 0041-22-7676012
Fax. 0033-(0)1-69079404        Fax. 0041-22-7823084
---------------------------------------------------

On Sun, 11 Sep 2005, Philip Bambade wrote:

> Dear colleagues,
>
> Concerning evaluating straight/curved linac tunnels for the ILC, is it 
> established that there is no significant impact on vertical emittance growth 
> up to multi-TeV CMS energies (in case one would like to make the ILC site 
> compatible with such an upgrade in the future) from:
>
> - synchrotron radiation emitted in the required bends,
>
> - chromatic filamentation of the phase-space in case of mismatched betatron 
> or dispersion functions at the input or along the linac ?
>
> Is there a reference which can be consulted describing the results of 
> calculations or simulations, taking into account, for the latter effect, the 
> larger correlated energy spread needed e.g. in CLIC to mitigate 
> wake-field effects ?
>
> I think it is an important question and we should have quantitative answers 
> to be able to feed any political debate on the articulation of ILC and CLIC 
> with proper technical arguments on the compatibility of the sites.
>
> I have asked these questions to our CLIC colleagues who attended Snowmass, 
> and they promised to make a list of all aspects relevant to this and 
> invest some work in this direction. But maybe there are papers with 
> computations of this already available in the literature ?
>
> Yours sincerely,
> Philip Bambade.
>
> ---------------------------------------------------
> Philip Bambade       E-mail: bambade@xxxxxxxxxxxx
> http://www.lal.in2p3.fr/recherche/delphi/bambade/
> Home Institute:                At CERN:
> LAL, Campus Universitaire      Delphi Collaboration pièce 129,Bât.208,B.P. 34 
> EP Division F-91898 ORSAY CEDEX France     F-01631 CERN CEDEX
> Tel. 0033-(0)1-64468370        Tel. 0041-22-7676012
> Fax. 0033-(0)1-69079404        Fax. 0041-22-7823084
> ---------------------------------------------------
>
> On Sat, 10 Sep 2005, Peter Tenenbaum wrote:
>
> > It's not as clear cut as this because the magnitude of the dipoles used as 
> > guide fields in the continually curved case is comparable to the magnitude 
> > of the dipoles used as steering correctors in simulation studies of the 
> > laser-straight case.  This means that:
> >
> > 1.  We may well find that the MPS implications of dipole failure in the 
> > stright tunnel are just as dire as the implications in the curved tunnel,
> >
> > or
> >
> > 2.  That since the guide field in the curved case is equivalent to a few 
> > hundred um of systematic misalignments of the quads, the placement of the 
> > quads along the curve of the tunnel produces enough guide field to prevent 
> > disaster even in the absence of the guide field from the dipoles.
> >
> > I realize that this is not the way a storage ring works -- there, if the 
> > guide fields fail, the beam leaves the vacuum chamber.  The difference here 
> > is that we are talking, essentially, of a "storage ring" with a radius of 
> > curvature which is orders of magnitude larger than any existing facility; 
> > the rules are different.
> >
> > Having said that:  I agree with Marc's judgement that it is desirable to 
> > have places in the linac where a runaway beam can be safely dumped.  The 
> > dumps can probably be relatively simple (at least compared to the main 
> > dumps) because of the low average power:  assuming that a dump may have to 
> > take a few hundred bunches, if we dump that many bunches at a given 
> > location once per hour the average power is still only 65 watts.  We would 
> > want to be sure that the abort kicker cannot accidentally put the beam into 
> > an aperture in the linac other than the runaway beam dump. One possibility 
> > would be to put a short warm section into the linac, and use a circular 
> > collimator with an aperture much smaller than the cavity iris as the dump. 
> > The strength of the kicker could be such that it could put the beam on the 
> > dump aperture but not onto a cavity iris.  It would probably be necessary 
> > to sweep the beam on the collimator as well (say with a vertical kicker to 
> > get the beam pointed at the collimator, and a swept horizontal kicker to 
> > spread out the train on the face of the collimator).
> >
> > The reason I want this is the following:  there are two proposals for the 
> > MPS interlock in ILC's LET:  the pilot bunch and a more conventional system 
> > which looks at magnet currents and fault statuses, RF fields, etc, and 
> > which also limits the amount that devices can be changed by operators 
> > during luminosity running (the former system catches faults but not cockpit 
> > errors; the latter system prevents cockpit errors).  In the case of BC or 
> > BDS, both pilot bunch and hardware permits can provide equal levels of 
> > protection, and so there is an added level of safety. In the linac, if 
> > there are not a series of runaway beam ramps, then the pilot bunch provides 
> > no protection and only hardware permits can prevent a disaster.  This means 
> > that each device's MPS thresholds have to be set correctly because if they 
> > are set wrong there is no additional protection from a pilot bunch.  I am 
> > worried that this would be an error-prone and failure-intolerant protection 
> > system.
> >
> > -PT
> >
> >
> > > 1) Does a kinked-tunnel linac require additional protection
> > > collimators? Does a non-straight linac require abort kicker / dump
> > > systems. In totalling the failure modes, effects, time-scales and
> > > likelihood, we find that a laser straight tunnel is relatively easy
> > > to protect. Because of this, the proposed pilot bunch scheme does not
> > > contribute to the linac MPS, and is really intended only for the
> > > bunch compressor / beam delivery MPS. With a curved or kinked tunnel,
> > > the potential failure of the linac dipoles becomes one of the main
> > > failure modes which the MPS design must be effective against. It may
> > > become necessary to implement several abort kicker/dump stubs along
> > > the length of the linac. The number required should be roughly equal
> > > to the number of kinks or one per kilometer (the spacing appropriate
> > > for a 10-15 us pilot to main train gap). In the kinked scheme
> > > protection collimators may be effective. Both of these have
> > > significant cost impact of perhaps a few 10's of M$.