Present: Hayano, Padamsee, Himel, Nagaitsev, Weise, Toge
We went over the proposed changes to our S2 report.
Decided to clarify that the phase 1.3 system test should be done before 1% of the final industrial production. This is to make it clear preproduction cryomodules do not count.
Himel will make the above change and also add an executive summary bullet that summarizes why we need beam
People liked Hasan's writeup on why we need beam and decided the full text would be included in some appropriate location in the S2 document. (To be inserted by Himel.) There was concern that “ILC-like beam” was not well defined and that insisting on full ILC specifications would in fact make the beam source too difficult to build and expensive.
Nagaitsev agreed to add the specificaton of ILC like beam into Hasan's writeup.
We hope to finish our work by email, so no meeting was scheduled.
Present: Hayano, Padamsee, Himel, Nagaitsev
This meeting was held to discuss how to address the comments made by the executive committee during their preliminary discussion of our report at the Beijing meeting. These comments (notes made by Himel and Willis) were emailed to the S2 group recently.
Padamsee will write a paragraph clearly explaining why the system test requires ILC like beam. This will concentrate on HOM production and absorbtion. This is to address the EC concerns that beam is expensive and time consuming and might not really be needed.
Nagaitsev and Padamsee will both check the spacing of HOM absorbers planned for the ILC and email the answer to everyone. We were unsure if it was 1 per cyromodule or 1 per RF unit. It effects the difficulty of doing numerical simulations of HOM production and absorbtion.
Nagaitsev and Padamsee will both check with the relevant computing experts to find out if full modeling of HOM production and absorbtion has been done; can be done; and whether the results can be trusted.
Himel will add a paragraph describing tests to be done at TTF during phase 0 to the section defining the phases. This is to address the comment that perhaps we should add a phase zero. We already have a phase zero (the TTF) and decided it would be good to describe some of the new tests that should be done there. This gives some goals that can reasonably be required for the EDR. Padamsee extract four tests for this paragraph from our big table listing tests.
Test effects of a dirty vent
Test HOM absorber (this is included with cryomodule 6)
Continue work on RF fault recognition and reaction to the faults
Continue work on reliability evaluation
Himel will write a paragraph and modify the executive summary to describe how S2 results should be linked to the overall ILC project schedule. S2 had explicility NOT made such a linkage because we couldn't reach a consensus. The EC was concerned that if the GDE didn't make some kind of linkage, some review committee or funding agency might do it for us. The logic that led to our final choice of linkage included:
Phases 1.1 and 1.2 are not significantly better tests than the TTF. They are simply part of the path towards phase 1.3.
Linking phase 1.3 to completion of the EDR would delay the EDR by several years.
The EDR is not well defined yet, but it is definitely NOT a full engineering design. People are thinking more in terms of 30% of the design completed.
The TTF is a major system test that reduces the risk that the phase 1.3 system test will uncover a major problem that requires major redesign.
Nevertheless, there are many changes in design since the TTF which is why we need a new system test. What we want to avoid is producing a zillion cryomodules that have to be rebuilt due to an undiscovered design error.
We decided that the project schedule should show that a successful S2 test must be completed before 1% of total number of ILC cryomodules are produced in industry.
The EC was concerned that the S2 report showed costs which will clearly be different than the costs either Fermilab or KEK would tell their funding agencies they need to do a system test. This could cause confusion. Also, the funding agencies will want to see a plan with actual costs. To address this we decided:
Himel will add more words saying how rough our cost estimate is and how it is likely to include some items that already exist in a laboratory.
Hayano and Nagaitsev will prepare budgets (they already have versions) showing what it would cost to do the sytem tests at their lab. These will not be put in the S2 report yet. They can be made availble for later planning by the regions or by the GDE. If this doesn't satisfy the EC, we can consider putting modified versions of the budgets in our report.
All of the above homework is due by close of business Thursday March 8 so people have a chance to read it before our next meeting.
Our next meeting will be Monday March 12 at our normal time (6:00 AM at SLAC) and phone number.
Present: Toge, Weise, Himel, Nagaitsev
We went over some proposed changes to the draft S2 report that had been circulated by email.
We agreed on wording in the executive summary about how our system test phases should be correlated with major ILC milestones. This is the matter that Toge and Nagaitseve sent email about. Here is the wording that was put in bold: There are number of phases to the system tests we propose (starting with 1 cryomodule and ending with several RF units). We are not making a recommendation on which phase should be achieved before major milestones of the ILC project such as approval of funding or start of construction.
The rest of Toge's comments were intrusted to Himel to incorporate in the draft.
We discussed Adolphsen's email that said measuring heating from HOMs would require more than one RF unit. While we admitted this is a difficult thing to measure as one must take the diffence between two large numbers, we didn't think it would be that much more accurate to have 2 or 3 RF units instead of 1. Hence we left the text unchanged.
Adolphsen's comment about $50M for stuff needed for beam being too much was right on. We found we had made an error in interpreting the numbers Weise had supplied us. The corrected costs are now in the draft and are $15M for non-beam infrastructure and $35M for adding beam. Nagaitsev indicated that this was close to the Fermilab estimates with labor left out.
Nagaitsev will make the Fermilab NML cost estimate available to us to put on our Wiki. This will just be for background material. He will give us the very detailed information they have and also try to summarize it for us.
Weise cannot easily make his cost information available as the split between FEL and linac costs was tricky(?) and is confidential at this point.
We agreed to change the expections for FP7 to be less than we had. The wording from Weise and Elsen will be used.
We agreed that Himel would make the necessary modifications and send it directly to the R&D board for approval. [Himel did so a few hours after the meeting.]
We went quickly over the talk Himel had prepared for Beijing. A typo was fixed and the costs on the last slide were updated to match the infrastructure costs mentioned above.
Present: Padamsee, Adolphsen, Toge, Weise, Hayano, Kephart, Himel, Nagaitsev
Nagaitsev presented his talk on how one could test the phase stability of the RF with respect to the beam. His conclusion is that to do it properly, one needs at least two klystrons and two LLRF systems. It could be done with two full RF units, but it would also work with two cryomodules connected to separate klystrons. His talk is on our wiki here.
Nagaitsev will modify his talk to clarify that the proposed test includes spectrometers both before and after the RF unit so that the change in energy can be measured. He will then get the talk reposted on the wiki.
We went over Adolphsen's email of Jan 3 that sketched a way to decide on the size of the phase 2 test. People liked the method although there was some discussion about whether the numbers he used were correct. In particular, there was concern that a goal of having 10% of the cryomodules fail during the 20 year lifetime of the ILC might be too big. If the failure involves only minor repairs (e.g. adding a turbopump) then 10% is probably OK. If it involves replacing the cryomodule then 10% is too big.
Himel will add text similar to Adolphsen's to the report. He will say it is a rough, crude calculation.
Himel will add text to the report saying that special accelerated tests need to be done separate from the full system test. Examples are repeated warm-up cool-down cycles and cycling tuners often.
We went over section 7, Cost of the S2 tests, and settled some open questions.
Himel will send summary to Toge who will then incorporate it in the section [Both of these have been done as of 1/16/07]
Present: Kephart, Himel, Nagaitsev
Only 3 people were present. Neither Adolphsen nor Nagaitsev had done their homework, so there was nothing to discuss. We ended the meeting after 5 minutes.
Present: Kephart, Toge, Himel, Hayano, Nagaitsev
The whole meeting was devoted to going over the risk assignments that Kephart had added to our spreadsheet. We made only a few changes. Kephart has now posted the revised version on the wiki page.
The next phone meeting will be Monday November 27 at the normal time.
Present: Schulte, Weise, Kephart, Toge, Adolphsen, Stanek, Himel, Hayano, Padamsee, Nagaitsev (by phone for part of meeting)
We agreed that the minimum size for a phase 1 test was 1 RF unit of 3 cryomodules. We worked on the detailed wording of our conclusions. We did not converge at this meeting. The discussion continued by email and then in a meeting we had on Nov 8 to go over the talk Himel would give in the plenary session. The final version is near the end of that talk which is on our Wiki here.
Nagaitsev went over his analysis of whether diagnostics resolutions are adequate to measure the energy and phase of the beam well enough to check the LLRF performance. We found an error in his assumption on how the phase measurement would be done as he took TTF as an example where they actually measure the bunch lenght after the bunch compressor to measure the phase. The proposed test facilities do not have this bunch compressor.
Nagaitsev will look at other ways to measure the phase stability, in particular running the beam at the zero crossing of the RF and measuring the energy. He will make a spectrometer design as part of this work.
We discussed testing cryomodules in a different region than they are built. There are two problems here. One is people think a design change is needed to make them transportable. Two is that each country must show that the cryomodule meets its pressure vessel codes and is not dangerous. Note that these problems MUST be solved if multiple regions are to contribute cryomodules to the ILC.
Weise told us that cryomodules 4, 5, and 6 will be directly connected (welded) to each other and be run from a single klystron. Hence they are an RF unit. Also, there will be a HOM absorber after cryomodule 6. This will be the first test of such an absorber.
We went over Kephart's table of tests needed if various modifications are made to a cryomodule. The updated version is on our Wiki here.
We agreed that out plan was flexible enough to handle both large and small budgets.
We decided drafts of our report sections would be due on Nov 30. The outline and writing assignments are here
Kephart agreed to take our table of tests that are needed and make 2 risk columns: 1 that judges the probability it will be a problem and 1 that evaluates how bad it will be if we don't find and fix the problem before the ILC is built.
Adolphsen agreed to figure out how big phase 2 should be based on risk minimization techniques.
Present: Padamsee, Himel, Weise, Lilje
Spent most of meeting going over the list of tests that could be done at TTF that Marc Ross had generated. The revised version of this list is on the wiki here.
Hasan presented the reasons why beam tests are needed to check HOMs. This involved both a powerpoint presentation and a spreadsheet. He convinced us beam tests were necessary.
Hasan will revise his powerpoint talk to explain and include the spreadsheet so the arguments will be clear to someone looking at the talk. He will post the revised talk.
Both Weise and Nagaitsev developed schedules for how long it takes to replace a cryomodule that is in a beamline. They are both in the spreadsheet on our Wiki here. Basically it takes about one month.
We went over the outline that Hasan produced and modified it. It is posted on the Wiki here
Present: Padamsee, Himel, Kephart, Toge, Nagaitsev
1) More Progress Items on Homework since KEK S2 meeting
2) Continue discussion of S2 milestone list started (see Wiki page)
3) How to prioritize work between now and Valencia
1) Bob reported more detail news about LHC evolution: R&D, string tests, pre-construction and construction. There were several string tests. He will put the information on the wiki page.
2) Tom pointed out that the list of milestones so far reflects more closely the planned activities at test facilities, rather than an ideal set of major milestones to carry out the S2 R&D obectives. Hasan pointed out that this list should only be treated as a first step in an iterative process, whereby we collect the information of on-going plans, come up with ideal plans, then try to match.
We continued going through the list of milestones.
Hasan will put the latest version on the wiki page.
Tom agreed to make a pass at milestones using the info in the starting milestone list, our list of tests and KEK thoughts on size of a test.
We agreed to skip the meeting on October 23. Next meeting will be October 30.
Tom urged all to get more homework done and posted before the next meeting.
Supplied by H. Padamsee
Present: Bob, Sergei, Hans and Hasan
1) Tom has put minutes of KEK S2 meetings on the web. Please read.
Assignments are in bold.
Comments on Minutes if any
2) Hasan put a summary talk of S2 activities on the web (comments welcome). He gave this talk at the TTC meeting plenary session.
Comments on Summary Talk if any
3) Comments on list of questions at bottom of Minutes of KEK meeting?
4) Progress on Homework since KEK S2 meeting. Comments on first attempt at milestones list.
5) EC report on S2 tentatively schedule for Oct19..
Do we need specific input from EC on any items.
6) How to prioritize work between now and Valencia
Everyone should email their comments on Tom’s list of questions at the end of the minutes of the KEK S2 meeting.
Item to add to the spreadsheet of RF unit tests.
What gradient spread can be handled by LLRF system. This test should be done with and without beam loading.
Sergei and Simrock have done some modeling work on this question, and their results will be made available soon as reports.
Hans reported that there has been serious discussion of carrying out tests on a dirty vent on module 3* next spring. A plan is under preparation. Both safety issues and cavity performance degradation issues will be addressed. CERN carried out a similar test for the LEP cavities. A report is available.
Pagani reported on behalf of Zanon that they have the capacity to fabricate as many as 100 cavities per year if the demand is there.
ACCEL reported a similar number.
XFEL plans to order several hundred starting in 08
JLab plans to order 100 cavities in 09 for the upgrade.
The rest of the time was spent going through the first two slides of the milestone list started by Hasan. Some items were added and wording clarified.
Present: Willis, Kephart, Ross, Stanek, Weise, Toge, Padamsee, Himel, Hayano, Adolphsen, Nagaitsev
This was an 8 hour meeting and a lot was covered. The presentation materials are on our documents page and just a few notes are given here.
We went over the spreadsheet of reasons for doing a system test. Each one was discussed and we decided whether it was best tested in phase 1 (loosely defined as a 1-2 RF unit test done before project approval) or in phase 2 (loosely defined as a larger test that would take place after project approval. Some ended up being labeled 1.5. Also we labeled some with 0 meaning we didn't think the test was necessary enough to justify a system test and with 3 meaning while the test would be nice, it required way too many RF units to be practical. Comments were also added. We also estimated how many RF units were needed for each test. The resulting spreadsheet is here.
There were some remaining open questions where we assigned homework to people:
Adolphsen and Weise will look into the testing of the effects of a dirty vent. The XFEL is planning such a test. They will see if the design of transitions from warm to cold and protection valves can be made similar enough in the XFEL and ILC so one test will suffice for both.
Kephart will investigate what tests are needed on the final production of industrial cryomodules and also whether special tests are needed when the cryomodule design gets modified to make it cheaper to manufacture. These tests may or may not be system tests our task force is responsible for planning.
Ross will determine if a system test (involving multiple RF units) is needed to understand things like phase reference lines, or if standalone tests without cryomodules are sufficient
Nagaitsev will check the beam diagnostic resolution necessary to check the energy and phase stability of the system to determine if this is a practical test.
Himel showed the LHC dipole production rate plot and pointed out how they increased the rate by a factor of 7 in a year.
Padamsee went over industrialization of LEP cavities and CEBAF cavities and LHC dipoles. His slides are here.
He noted that LEP cavity production started 6 months after they were ordered. LEP found new problems with the cavities when they were exposed to beam: multipactoring occurred in couplers due to different standing wave patterns.
He reported that CEBAF cavities got to full production rate in 2 years. The cavities were built in industry and the assembly was done at CEBAF.
Weise reported that the XFEL needs 100 cryomodules. They intend to split the contracts between two companies. A pre-series will be done by the companies with DESY help (starting now). All cold tests will be done at DESY.
We started a discussion on how my RF units were needed for a phase one system test. We then went around the room asking each person what they thought. (just to get a first impression of where we stood) Answers were 2 in a string, 1 built in each region then put together, 1 per region, 1 or 2 total, and 3-6 total.
Himel showed an optimistic schedule for how the project gets from where we are to the beginning of digging a tunnel. The intention was to show that there is considerable time between completion of the EDR and the need for cryomodules in the tunnel. Himel was told he was overly optimistic in having only 21 months from DOE deciding they want to build the ILC to approval by congress. Two years would be better. The schedule shown here
has that adjustment made but is still considered to be optimistic as it assumes the politics goes as fast as possible with no hiccups. There are many possible causes for delays in funding such as negociating among countries or a desire to wait for the XFEL to have tested most of its cryomodules (scheduled for mid 2013).
Kephart mentioned that the time early for completion of 1 RF unit in the US is 2010 because funding was not as high as requested.
Toge gave his talk How much is enough?
We found some arithmetic errors in his slides which have been corrected in the version referenced above. Some very crude numbers from Nobu's talk:
Each region would have to spend $21-30M for a phase 0 test of 1 to 2 RF units. This includes $12M of infrastructure.
Stage one would cost $7-21M per region for 1 to 3 RF units per region.
Stage 2 would cost $5.5-15.5M per region for 1 to 3 RF units per region.
For the whole world, the 3 stages would cost $36M for infrastructure (including e-beam welders but not cryogenics) and $65~194M for the RF units themselves. Note that he assumed the costs would gradually go down as we produced more units.
In the discussion it was stated that we might not need to buy e-beam welders as the cavities would be built in industry and they might own them already. Also, LLRF, beam diagnostics, and injectors were not included in Toge's numbers and they would add about $20M to the costs. (I think this $20M was per region that has beam, but not positive of that.)
As a sanity check, Weise pointed out that the TTF with 6 cryomodules (roughly 2 RF units) would cost about 75M euro to replace. Hence $100M for the total cost of S2 might be reasonable.
Padamsee plus everyone should consider whether any of the lessons learned on previous projects indicate new lines we should add to our spreadsheet of reasons for doing a system test.
Ross volunteered to write a section for the final report on the maturity of cryomodule technology relative to SC magnets.
all the work that has been done on whether beam is needed in the test linac. Basically too big a linac would be needed for any practical beam dynamics test.
It appears there are two driving reasons left to have beam. One is to check for trapped higher order modes. This has been done at FLASH, but we are changing the design of the endgroups and are considering different cavity shapes. The second reason is to check if at least 90% of the HOMs are extracted to room temperature as assumed in the design. It was believed (and Ross later confirmed) that this has never been tested at FLASH but could be.
Adolphsen will check to see what we know about beam putting power into Helium.
Padamsee showed a small spreadsheet
that showed the number of cavities we could expect to be available for our use from the S0/S1 production. These were enough to produce 1 RF unit in 2009 and 3 in 2010 for the whole world.
Weise pointed out that DESY presently makes about 25 cavities per year. Padamsee showed that increasing to 50 from each of the three regions in 2007. This seems overly optimistic.
There is a cryomodule (at Fermilab I think) that will be done in 2008 that Hasan forgot to include.
Padamsee and Lilje will get better estimates of the rate industry can give us cavities.
Toge-san commented that Japan doesn't require a system test before funding. However, they don't allow budget changes after project approval.
Weise commented that TTF was accepted as a system test for ILC. He thinks that ILC is making too many changes (new tuners, new RF system, maybe new couplers, new quad location) for TTF to be a valid system test.
At the R&D board meeting that took place just after the S2 meeting the above point was discussed. People wondered if the changes which required the new system test were really needed enough to warrant the added time and expense of the new system test.
We continued our discussion of what size a system test should be. As it looks like even phase one of a system test might not be done by the time the EDR is done and we marked some of the tests in our spreadsheet as phase 1.5, we decided we should simply make a model with many milestones and not just two distinct phases.
It was mentioned that Europe's contribution would be the XFEL and there could be no extra test linac before 2014. A possibility was left open that extra XFEL cryomodules could be built for us.
Japan has plans to build two RF units in 2010 and 2011.
Padamsee and Himel will make a list of milestones and a rough budget and schedule.
If things go very very well, our group will discuss this in phone meetings and settle it before Valencia.
More likely we will discuss it at Valencia.
We decided to have an in person meeting in Valencia. This will be the afternoon of November 5 and the morning of November 6.
We will have a phone meeting at out regular time (6:00 AM at SLAC) and number on Monday, Oct 9. Padamsee wil distribute an agenda. [This was decided well after the meeting by Himel and Padamsee. Himel will be on a flight to Caltech and hence will miss this meeting.]
Everyone should think about some remaining issues: How are our tests tied to industrialization? Do industrial cryomodules need to be tested with beam? Do we need independent tests in the 3 regions or should we just say what the world needs and if a region wants to do more than their contribution to the world program (e.g. to stimulate industry) the are free to do so? How close to final design must the cryomodules (and other systems) be? Future year budgets are unknown. How do we make a plan that is flexible enough to handle both large and small budgets? Should all cryomodules be identical, or just plug compatible?
Present: Weise, Toge, Padamsee, Himel, Hayano
Most of the meeting was devoted to a talk by Weise titled: “S2 Input from TTF / FLASH and XFEL”. This is posted on our Wiki at this location. A few highlights and some of the discussion are captured in these minutes, but please see the slides to get the whole content.
The gun and last 2 RF modules can run at up to 10 Hz.
Need 0.01% amplitude and 0.01 degree phase stability due to bunch compressor. At present get 0.03 and 0.03. This is determined by the LLRF.
No cryomodules have been warmed up for two years.
There is no sign of aging.
There are extra attenuators or different couplers to adjust gradients on some cavities. This is not planned for the XFEL.
They typically have 1 shift per week scheduled maintenance. This doesn’t show in the downtime statistics as it is scheduled.
71% of the downtime is due to klystron or modulator problems. Note that due to the bunch compressors which separate different cryomodules, all of the klystrons and modulators must work at once for the TTF to be up. This will be different for the XFEL and ILC which will have built in redundancies.
Dark current from the RF gun has been a major problem. It has stopped long pulse running. The addition of a kicker shortly after the gun has reduced this problem. They have run with 0.5 ms bunch trains.
The XFEL electronics will be in the tunnel. They will not be put near quads where the radiation is expected to be higher as the quads over-focus the dark current and make it exit the cavities. There will be 0.5 meters of longitudinal shielding in front of the electronics. The amount of transverse shielding has not been determined yet.
With TTF as a “test” facility there are always excuses for downtimes (making improvements etc.) Only get real availability numbers if operators run the accelerator (probably for users) and no improvements are made.
One community should build both XFEL and ILC. If XFEL has problems with cavity/cryomodule production, this will affect the ILC.
XFEL success is very important for the ILC
Better than looking at analyses like this one is to have operators come and run TTF. That is the best way for people to learn what it is like to run an accelerator like this.
Tests (probably to destruction) will be done on a cryomodule in late 2007. This will include multiple thermal cycles and then a catastrophic vent.
It costs about a factor of 2 more to make a test facility than to just build the cryomodules and RF system needed for such a test facility.
We then discussed the draft agenda for our meeting at KEK.
Suggested additions were:
How to incorporate lessons learned into our plans. That is, does it tell us how much we need to build or run. Nobu will lead this discussion.
Summarize our needs for beam and the emails which have been flying around about the need or lack thereof for DFS steering tests. Hayano-san will lead this discussion
Plan our schedule for later conclusions and report writing of our task force. Himel will lead this discussion.
The KEK meeting is the afternoon (1:30-5:30) of Friday Sept 22 and the morning (8:00-12:00) of Saturday Sept 23 at KEK. Note the MAC meeting ends at noon on Friday and the RDB meeting begins at (the GDE calendar says noon, but I hope it is actually 1:00) on Saturday. We may choose to continue our meeting late on Friday or go to dinner together with a vow not to discuss business. The meeting will be held at KEK in building 3 on the third floor in room 325.
Present: Hasan, Tom, Nobu, Bob
Update on EC meeting on S2
At EC’s request, Hasan presented the S2 charge and described our work plan and status. The presentation was close to what Tom presented at Vancouver for S2 summary with some updates.
Barry asked (1) how we intend to get funding for the scope of work that might emerge and (2) how we intend to deal with duplication
Hasan’s response to (1) was essentially that getting funding is not part of our charge, but that our planning could have a positive feedback on agencies who are planning funding amounts for up coming years, as is the case for DOE.
His response to (2) was that some amount of duplication is good for positive interaction between groups, cross-checks on procedures and results, esp in the learning mode. There are also some strategic reasons for regions to have core capabilities in SRF for ILC and other projects based on SRF. Ultimately it is up to the RDB to determine whether there is duplication.
Brian Foster emphasized that even though it is not within the declared scope of TTF-II (FLASH) to assemble an RF unit at 31.5 MV/m, there is now discussion within the FP7 framework to propose an SRF installation for ILC activities at CERN. Cavities, cryomodules and RF units could be part of the scope for the European region.
Nobu commented that the scope of S2 is very important and that ideas from all arenas and regions need to be heard. Europe esp needs to be heard in S2.
For the next meeting we should try and get Hans to attend and comment on Nobu’s analysis of lessons learnt from TTF and what we might learn for ILC. Hasan agreed to email Hans.
Update on interface between S0, S1 and S2.
At the last meeting we discussed how the plans emerging from the S0/S1 task force will interact with plans that emerge from S2. For example, is the planning for the number of cavities > 35 MV/m consistent between the desires of the two groups. This is difficult now because we (S2) have not decided on the number of RF units needed. At the previous meeting Hasan showed what he presented to the US Regional Interest Panel, estimating the number of RF units based on the capacity to prepare and test cavities that exists and is under planning in the US labs, as well as a comparison to the capacity which exists globally. (see last week’s minutes and slides) To start the interface Hasan made a similar presentation at the one-day S0/S1 task force meeting. At this meeting it was agreed to divide S0/S1 work plan into two stages: (1) tight-loop and (2) production batch. Very briefly, the first stage will address the yield of the final preparation steps by repeating these many times in a tight loop to determine the initial yield. Subsequently improvements will be put into the process (from a parallel/coupled R&D program) and the cycles repeated to measure a new (and hopefully) improved yield. In the tight loop stage about 10 cavities will be processed, and none of these will be available for cryomodule and RF units until the tight loop studies are complete.
The goal of stage 2 is to improve the yield of production cavities. There will be several large production batches of cavities ordered in series (some to qualify new vendors). These will go through the processes (or improved processes) If they reach 35 MV/m they will move to cryomodule population. If they are limited by field emission, they will be reprocessed until they clear 35 MV/m. If they are limited by manufacturing or material defects they will be further analyzed. The size of the production batches is under discussion. One provisional example to judge the scale : for 3 RF units (one per region) at ILC spec, we will need 9 modules at 31.5 MV/m , or 72 good cavities at 35 MV/m. The number of starting cavities and the number of processing cycles to reach this goal will depend on yield assumptions, under discussion. Each region may also need to assemble a “learning RF unit” before the final one, so the number of cavities and cycles will be much larger.
Discussion of Fast Track Scenario
Nobu outlined in more detail an evolutionary scenario, which (by the time of these minutes being finished) is now on our Wiki Page. It addresses learning issues as well as hardware compatibility between region issues, and the possibility of assembling regional RF units into a larger string.
What can we do by KEK
Bob reminded us of the need to look backwards from ILC production to where we need to be in the intervening years. His model is on our wiki page. He reminded us that there are likely to be significant ILC R&D funds in the US HEP community in 08-09 timescales due to turning off the B-factory and the Tevatron.
Tom is also following the looking backward from production approach and will think about other models for the intervening years, possibly based on production experiences at LEP and LHC.
Hasan agreed to work with Nobu to look more at the evolutionary scenario.
Our next meeting will be on Monday Sept 11 at the usual time.
- update by Nobu on operations information
- update by Hasan on overlap of S1/S2 and Regional Panel activities
Present: Bob, Nobu and Hasan
Nobu went through his slides which analyze the operations experience of TTF phase 1 and TTF phase 2 and XFEL prep. This was a very nice summary of the lessons learned from the many papers supplied by Hans Weise and now on the Wiki page. But it would be best if Hans would supply some top level summaries of his own to make sure that everything is interpreted properly.
Here are some of the comments made during the presentation of the slides.
- Now that we have gone through a few cases (TRISTAN and TTF) our task force needs to make a general list of items that we need to apply to S2 goals from these operations lessons. This will help us focus our analysis of the operating experiences for other cases to follow, CEBAF, LEP, SNS…For example, how can our RF units do better than TTF-II, which systems need special attention for improvements.
- In figuring out the scale for S2, we have talked about two approaches: look backward from where we need to be to get ready for production, look forward from present test facility plans. A third approach may be appropriate: what scale of system tests are necessary to address how we can improve system performance towards the reliable operation we desire for ILC. i.e how do we get from here to there on reliability?
- How much improvement has there been in availability over time from TTF-I, TTF-II,
- we need to get more data on the heat leaks measured for the cryomodules ( I think these are available)
- we need a definition of “operation up time” – is that cold time or RF on time?
- we need a distribution of down time between RF, couplers, tuners, cavity, etc, if available
- the XFEL has factored in some lessons learned from TTF – I and II. Do these lessons apply to ILC? E.g test every cryomodule
nobu’s last slide (remarks) deserve special attention
Next, Hasan gave a summary of material that has been presented and discussed over the last two months in the Regional Panel Meetings. particularly on S0, S1, S2 related topics.
He emphasized that the main reason for visiting this material is to shoot for consistency between the final outcomes for S0, S1 and S2. i.e. will we have enough cavities at 35 MV/m and CM at 31.5 MV/m coming out of S0/S1?
It was clear that we need to visit the same issues in the S0/S1 task force, and Hasan said he will bring this up.
Bob mentioned that at the request of the DOE he is preparing a schedule for the ILC goals for S0, S1, S2 milestones. It will be a technically limited schedule.
Nobu suggested that as a possible next step, our task force should perhaps discuss a “fast- track’ scenario for a minimal number of RF units at 31.5 MV/m.
We agreed to meet on Monday Aug. 28 at the usual time, since we have to skip Monday Sep 4 due to US holiday.
Present: Himel, Hayano, Toge, Padamsee
Padamsee went over the spreadsheet on regional plans for building and testing RF units that was originally scheduled for the Vancouver meeting. The PDF of his spreadsheet is here with the Vancouver presentations. Some comments included that the plans may be optimistic (they certainly depend on funding) and that we should ask for more detailed schedules (Hasan agreed to do this).
Hasan commented that we may need to justify having 2 test linacs (one at KEK and one at Fermilab). Toge commented that we should set the scale of the tests needed for technical reasons and let someone else above our pay grade decide where the test(s) are done. Himel expressed agreement with Toge.
Toge gave a talk about Tristan's experience with SCRF cryomodules. Both it and the papers it is based on are on S2's main Wiki page. A few notes from the talk:
If ILC cryomodules work as poorly as those at TRISTAN ILC will be in serious trouble.
We may be able to learn specific mistakes to avoid from the TRISTAN experience
Components that can be tried out before design freeze and production launch should be tried out in realistic stress testing conditions thoroughly.
Pay particular attention to power, mechanical, thermal cycles, and radiation
Some failure modes, before showing up, require a certain operation time on each individual component, rather than just a big total accumulated operation time.
Seeing a failure mode is one thing. Understanding it and developing a cure is another (i.e. takes time).
We should repeat similar analysis of colleagues’ experiences from KEKB, CEBAF, LEP, SNS. Toge will do this. [After the meeting Himel suggested that TTF should be added to this list.]
There was some discussion that it might take repeated tests to make sure we have a good cryomodule. This could happen because we want to test improvements to the design that we make to correct problems found in an earlier test, or because the first test is not with the final industrialized design.
[A comment from Himel not given at the meeting: It is becoming clear that we can come up with multiple reasons to have multiple large tests. Ultimately we are going to have to weigh these reasons against their cost and schedule implications (do we really want to spend 15 years and a billion dollars building 3 or 4 generations of test strings? How are we going to make the tradeoff of risk vs. schedule and budget? We all need to give that some thought. End of sermon.]
The next meeting will be Monday, August 21 at our normal time (6:00 AM SLAC time) and phone number. Padamsee will chair the meeting. Himel will be absent.
Present: Himel, Hayano, Toge, Kephart
We decided our in-person meeting would be the afternoon of Friday Sept 22 and the morning of Saturday Sept 23 at KEK. Toge will arrange for a room and telephone. This meeting time exactly coincides with a planned meeting of the RDR management and cost engineers. Some of us may get called out to that meeting for short periods of time. We will adjust our meeting schedule appropriately in real time. We will try to have dinner together on Friday evening and may meet more after dinner if we feel it is necessary.
Our next phone meeting will be Monday August 7 at the same time (6:00 AM SLAC) and phone number. This is off our nominal every-other-week schedule because Himel and Toge will be away the following week.
Toge agreed to get Kubo-san to organize the writeup of the fact that DFS steering has been well enough modeled and tested in real accelerators that we do not need a further test.
We spent the second half of the meeting going over the spreadsheet of reasons for doing a system test. Edits were made and viewed in real time. The philosophy at present is to put down any reason we can think of for doing a system test. Later we may decide that some of them are not worthwhile but for the moment they stay on the list. We added everything we could think of and then went back to the top and discussed each item on the list through line 7. In some cases we assigned homework to a person indicated in column F. Himel did a bit more editing after the meeting (mainly filling in whether the test required a RF units in a string or beam and it is now up on our Wiki page as Version 4.
Present: Padamsee, Himel, Hayano, Toge, Kephart, Weise?, Nagaitsev, Kubo, Walker, Lilje? The ”?” indicate people that my fallible memory says were there but I didn't write down. Could others please confirm this attendance list?
Kubo: length needed for beam test
Nagaitsev: Fermilab ideas for ILC test linac
Toge: Reliability evaluation and comments
Himel: previous industrialization efforts
Kephart: An ILC industrialization plan
Padamsee: Update on timelines fom TTF, STF and SMTF
General discussion of our task force plans and informal progress reports. A piece of this discussion will be a brain storming session to list reasons we might want a string test facility. That is, what things do we need to test that require such a facility.
We ran out of time before getting to items 6 and 7. All the talks are posted here on our wiki.
Below are some notes on the talks and the discussion.
1. Kubo: It would take about 4800 cavities to measure emittance growth due to cavity misalignment if beam from an RF gun were injected. This is impractical.
In the real ILC, if we want the jitter to be less than 0.2 sigma at the end of the linac, the quads would need to vibrate by less than 17 nm. In discussion, it was thought better numbers to use would be 0.5 sigma and 30 nm. With the original 17 nm assumption and test linac with 30 cryomodules each having a quad, the beam would jitter 0.4% of the 83 micron beam size. Walker commented that we don't need a string test to measure quad vibration. An accelerometer is a more appropriate tool. Nagaitsev commented that we may need 3 RF units (a cryostring) to measure vibrations as caused by fluid flows. Walker responded that this can be done with a single cryomodule where the fluid flows are adjusted. DESY is doing this measurement. [After the meeting, Himel found there are actually 4 RF units in a cryostring, not 3.] Kephart pointed out that vibrations caused by fluid flow might depend on the cryo controls. We need to talk to some knowledgable cryo people about this. Himel will contact Peterson. Walker commented that the actual quad vibration tolerance would be 50 or 100 nm, not 17. After all, there can be an end of linac intratrain feedback.
2. Nagaitsev: He showed a list of things that might be tested in a string test but did not go into most of them in detail. We should include his list in our brainstorming session.
In detail he covered tests of beam based alignment / dispersion free (or matched) steering). He started with a 0.4 mm-mRad vertical emittance which has been seen in the A0 photoinjector. (This compares to the 1.0 that Kubo assumed in his talk.) He calculated numbers for an 18 cryomodule string which produces 5 GeV. This size was chosen for no particular reason. Each cryomodule was assumed to have a quad to maximize the effects of quad offset problems. Full acceleration was assumed. Nagaitsev will repeat the calculation assuming small acceleration which should increase the sensitivity. His present result showed that the emittace growth before DFS steering was about twice the minimum amount that could be measured. After DFS steering it was about one fifth the size that could be measured. If a damping ring were used the effect would be easier to see.
Walker commented that there have been positive results from LEP on DFS so we don't need to test it here. If we had a linac we would choose to do the test, but this not an argument to build a test linac. It is icing on the cake. Nagaitsev agreed. Padamsee suggested that we should document the fact that we are confident and why. This documetion can be done by Kubo, PT, Walker, Schulte. I need a volunteer from the committee to organize this.
3. Toge spoke about reliability. He has been reading textbooks and gave us an introductory course. His rough conclusion is that it is difficult to run enough items for long enough to demonstrate an MTBF > 100,000 hours (a typical required MTBF). A system test would more serve the purpose of detecting major screw-ups. As designs are likely to change with time, testing of MTBFs may need to be done more than once as it is often the details of the design that mess up the MTBF.
4. Himel related what he had learned about the industrialization of superconducting magnets for SSC and RHIC. Both started their industrialization after project approval. A detailed calendar was shown for RHIC. Note that the manufacture of the SC cable started well before project approval, it was the magnet fabrication technology that was transferred to industry after approval. Himel will research the industrialization experience for LHC magnets and LEP cavities. Toge will do so for Tristan SC cavities.
5. Kephart described the cavity and cryomodule industrialization plan he has worked out. As part of this he has a spreadsheet (on the web along with the talk) that we can play with to try alternate plans.
We ran out of time so did not cover agenda items 6-7.
We will all think of reasons why one might need a system test and email them to our mailing list. We will discuss them and continue brainstorming at our next meeting.
Our next meeting will be by telephone on Monday July 31 at 6:00 SLAC time, 2:00 PM European time, 10:00 PM KEK time. We will use the phone numbers normally used for the RDB meeting. If anyone is missing these, please ask me or Maxine.
In general, people with information to present should put it on our wiki a couple of days in advance of the meeting and email the group about its existance.
Future phone meetings will occur every two weeks unless changed for some reason.
We decided to have an in-person meeting associated with the TTC meeting at KEK. This TTC meeting runs from Sept 25-28. Himel will contact people to determine a precise date.
Present: Hasan Padamsee, Tom Himel, Hitoshi Hayano
2) Suggestions for Work Plan revisions and Work Leader Assignments
3) Progress on work plan activities
4) Interim goals for work to complete by Vancouver
Tom suggested he could set up a public Wiki page for our meetings so anyone can keep track of our discussions, esp consultants.
Tom also suggested he would set up a discussion forum for emails.
The organizer will prepare the minutes for the S2 Task Force meetings and give summary report to RDB. The minutes will be posted on the Wiki page.
The organizer role will swap monthly between Hasan and Tom (co-chairs). For July, Tom will be the organizer.
We discussed the options for the regular phone meeting time. Every other Monday at 9 am EST seemed the preferred choice on the table. Any comments from others?
We need a better (toll-free) phone number for Japan where it is 10 pm. Nobu tried hard to join our phone meeting from home.
The next phone meeting is proposed for July 10 (July 3 is close to the July 4th holiday in the US).
Email exchange is encouraged in between.
2)Members present accepted their Work Leader assignments.
3) Progress on work plan activities (and discussion)
Hayano reported work by Kubo on a rough estimation of the required length of the test linac for testing cavity alignment, wakefields and quad vibrations. A report will be forthcoming. There was a presentation of this work at KEK, and comments of this presentation will also be provided.
The summary of Kubo’s work is that a test linac with about 200 cavities (24 modules, 8 RF units) can be useful. The beam could be generated from an RF gun with flat beam (like at A0). There are several assumptions which should be discussed by the Task Force.
Tom agreed to find out the nature and extent of the work done by Tor on the necessary length of a test linac for emittance growth measurements.
Hasan agreed to find out about the nature of similar studies carried out by Nagaitsev.
Tom reported he is collecting information on pre-production and production scenarios for other projects such as SSC, RHIC, LEP and LHC, XFEL. This will help determine where we need to be in 2010 to be “prepared” for ILC construction.
There was a short discussion of considerations that will bear on the question of how many cryomodules will be needed globally by 2010.
We should develop a strong rationale for the number needed.
Tom expressed doubt if one can address any reliability questions regarding trip rate etc. with any reasonable (fundable) number of modules.
Why do we need more than one RF unit for a system test ?
Hasan suggested that the first RF unit would be a “training ground” and is not likely to reach the ILC performance for components, for e.g the average gradient would be less than 31.5 MV/m, the cryomodule would not be the ILC module (e.g. generation IV), the LLRF/tuner would be evolving .
A second RF unit would come closer to using the “building blocks” of the ILC.
And a third (or fourth) RF unit would need to be assembled from industrially built cryomodules. We need to discuss the relationship between these numbers and regional interests.
Hayano mentioned that KEK addresses such questions at STF meetings from a domestic point of view, what is needed , what can be funded.
Hasan pointed out that similar discussions are going on during US Regional Interest Panel meetings (formerly called Bid to Host Panel), held once a month.
The S2 task force needs to hear about the considerations discussed during these regional interest activities.
Hasan mentioned that he would like to gather input from TTF, STF and SMTF about which items of R2…they plan to accomplish so as to fill out the table for Stage 1. Nobu has diligently gone through the TRC and made a list of items classified as R2, R3 and other. It would also be useful to attach a year by which such an activity will be addressed. Hasan will send out email to task force members on this subject.
4) Interim goals for work to complete by Vancouver
Hasan proposed that our aim for the Vancouver RDB presentation should be to state our charge, describe our overall work plan and present some details about progress on one or two of the specific items.