Beam loss/collimation issues on 4GLS:
1. Parasitic loss should be localised deliberately rather than randomly - so we know where it's going. This appears to be the philosophy of the APS ERL proposal, and the Japanese ERL proposal. http://www.erl07.dl.ac.uk/Tuesday_Presentations/WG_2/Session_1/Xiao-ERL07-3.pdf
although admittedly they are using fewer collimators; but their beam power is about 10x greater (700 MW cf our 55 MW).
Also see:
http://www.aps.anl.gov/News/Conferences/2006/APS_Upgrade/apsmac/Borland_ERLPhysics.pdf
Japanese ERL proposal thinks about localisation too:
http://epaper.kek.jp/fls06/TALKS/WG221_TALK.PDF
but one of the things they are thinking about is trapping of very small losses (in the few Watt level) from gas scattering.
So I think that localisation is a sensible way to go. The collimator length does not depend much on the total power - the collimator lengths is just there to stop the shower coming out the end of the collimator with too high a mean energy. 200 mm Cu is the estimate for 550-750 MeV incident electrons.
2. We are worried about irradiating the undulators, and we have a lot of those. This is what XFEL worry about:
http://flash.desy.de/sites/site_vuvfel/content/e403/e1642/e1849/e1933/infoboxContent2164/TESLA-FEL-2007-051.pdf
(this is a very detailed paper that I have not digested yet!)
But I think at the moment that it is similar in principle to the BESSY-FEL collimation design:
http://web.elettra.trieste.it/fel2004/proceedings/papers/TUPOS02/TUPOS02.PDF
Remember that the beam power in the FELs is much lower than our 55 MW, so we should be as worried about collimation as they are!
3. At the moment, we don't know the best way to detect beam loss. Some issues:
a. Measurement of differential current monitors on the primary beam - I don't think this will be nearly good enough. The S/N isn't good enough - differential measurements are good enough for 'catastrophic' loss over ms, say, but not for ongoing losses.
b. Direct measurement of the loss, via scintillation or other techniques seems like the way to go. e.g. cable-based monitors like those used on ERLP. My favourite at the moment is a 'halo monitor':
c. Measurement of pressure rise sounds good (because it's cheap, and doesn't introduce extra components into the vacuum envelope that cause impedance), but I worry about the timescale of response of things like this, and what happens when they aren't working properly - we have lots of experience of that on the SRS!
d. Similarly, measurements of temperature rise are also not great, because you could for example get a mis-tuning of the beam optics which caused localised beam loss, and then not know about it until it was too late.
For all those reasons, I'm still in favour of collimators that will passively protect the other apertures as much as possible, hence have a fairly dense layout of them (regular spacing in phas advance), and to put monitors in as well.
Thursday, June 07, 2007
Friday, June 01, 2007
4GLS Phasing Options
Phase 1
Would be low current (good for focus on dynamics expts) and retain some of the combined sources vision. Included is the XUV FEL, the far IR FEL and the spontaneous loop (populated to some level).
There was consensus that completing the HAC loop is a good idea. There are good technical reasons for this relating to the RF. I think it also looks good from a general-project-stance consistency point of view.
Low current in Phase 1 means that fewer IOTs and a smaller cryoplant is needed. Feeling was that we'd buy the rating of IOT and PSUs needed for whole project and use one to power several cavities. When we move to higher power buy more of the units.
A move from three injectors to two injectors was proposed as the spec of the HACL system would cover the requirements of the IR FEL system.
The IR FEL beam would be taken off after Module 1 of the main linac. Need either to schedule operation of spontaneous loop and Far IR FEL at different times or use a kicker of some sort (not felt to be tricky at rep rates condidering).
A move from two upstream arcs to one upstream arc was proposed. There's then a requirement to split the XUV-FEL beam from the HACL beam. At the lower rep rates proposed in Phase 1 this is technically feasble with current technology.
Could phase the development of the XUV FEL, e.g. 6/8-40eV first and then 35-100eV. Might not need to do this.
For the discussion the current in the spontaneous loop was 0.3 mA (4.3MHz at 77pC). There was no real discussion of 13MHz at 200 pC (which of course gives 2.6 mA).
I've tried to capture some of the machine implications on the attached powerpoint file. I've not messed with the HACL though.
Phase 2
Would be high current in the spontaneous loop, the VUV FEL, further IR FEL development, further population of the spontaneous sources.
Need to mesh R&D roadmap with injector development.
More IOTs needed and cryo plant needs uprating.
R&D development on kickers permits a single upstream arc and kicker otherwise a second upstream arc needed. Again need to mesh with the R&D roadmap.
VUV FEL incorporated.
Need a set of milestones for eventual delivery of 100mA
Lia will be producing a document on the global view of R&D needed for ERLs with timeline as an output of ERL07.
Clearly the costing matrix would need to capture phasing of the RF and cryo, have the costs for XUV FELs separate, have the cost for the FAR IR photoinjector and FELs separate.
Hope this is clear, it's not that easy to summarise!
Would be low current (good for focus on dynamics expts) and retain some of the combined sources vision. Included is the XUV FEL, the far IR FEL and the spontaneous loop (populated to some level).
There was consensus that completing the HAC loop is a good idea. There are good technical reasons for this relating to the RF. I think it also looks good from a general-project-stance consistency point of view.
Low current in Phase 1 means that fewer IOTs and a smaller cryoplant is needed. Feeling was that we'd buy the rating of IOT and PSUs needed for whole project and use one to power several cavities. When we move to higher power buy more of the units.
A move from three injectors to two injectors was proposed as the spec of the HACL system would cover the requirements of the IR FEL system.
The IR FEL beam would be taken off after Module 1 of the main linac. Need either to schedule operation of spontaneous loop and Far IR FEL at different times or use a kicker of some sort (not felt to be tricky at rep rates condidering).
A move from two upstream arcs to one upstream arc was proposed. There's then a requirement to split the XUV-FEL beam from the HACL beam. At the lower rep rates proposed in Phase 1 this is technically feasble with current technology.
Could phase the development of the XUV FEL, e.g. 6/8-40eV first and then 35-100eV. Might not need to do this.
For the discussion the current in the spontaneous loop was 0.3 mA (4.3MHz at 77pC). There was no real discussion of 13MHz at 200 pC (which of course gives 2.6 mA).
I've tried to capture some of the machine implications on the attached powerpoint file. I've not messed with the HACL though.
Phase 2
Would be high current in the spontaneous loop, the VUV FEL, further IR FEL development, further population of the spontaneous sources.
Need to mesh R&D roadmap with injector development.
More IOTs needed and cryo plant needs uprating.
R&D development on kickers permits a single upstream arc and kicker otherwise a second upstream arc needed. Again need to mesh with the R&D roadmap.
VUV FEL incorporated.
Need a set of milestones for eventual delivery of 100mA
Lia will be producing a document on the global view of R&D needed for ERLs with timeline as an output of ERL07.
Clearly the costing matrix would need to capture phasing of the RF and cryo, have the costs for XUV FELs separate, have the cost for the FAR IR photoinjector and FELs separate.
Hope this is clear, it's not that easy to summarise!
Subscribe to:
Posts (Atom)
