Oct. 27, 2006

The ARCS project has moved ahead steadily over the past years. The project is nearly 80% complete. All major components are designed, and either delivered or in procurement. Most of the remaining tasks involve installation. You can watch installation in near real-time on the ARCS webcam: http://it.sns.ornl.gov/images/webcam/ARCS-Live.html The big ribbed item in the foreground is the large side of the scattering tank, which holds the detector array and the sample chamber.

The ARCS project will deliver an instrument with all promised technical features, including the detectors in vacuum and a windowless region around the sample for the forward and single- scattered beams, owing to an innovative gate valve for sample isolation. The detector coverage is expected to be close to pi steradians when the instrument is ready for commissioning, and an elliptical guide will extend through the shutter to the sample. Unfortunately, the T_0 chopper will not be delivered until the end of 2007 -- spinning a 70 kG load at 180 Hz with phase control has proved an engineering challenge, but the procurement is moving well now. Incident neutron energies exceeding 1 eV should be possible with this T_0 chopper.

The ARCS project is meeting its quality and cost goals, but the schedule has slipped. There was a thorough review of the status of the ARCS project last month. It is now expected that ARCS will be complete in the summer of 2007, and ready for experiments before the end of 2007.

It is not too early to think about experiments that will use ARCS. The first users of any inelastic instrument are likely to experience a higher background countrate than later users because the optimization of shielding for background minimization is a major part of commissioning. Nevertheless, the shielding of the ARCS instrument should be largely complete in 2007. A single crystal goniometer with a displex refrigerator will be available, capable of temperatures of 15-20 K, a large angle of rotation about the vertical axis, and two smaller ranges of specimen tilt.

Conservatively, we suggest first experiments with incident neutron energies less than 200 meV (owing to possible absence of the T_0 chopper), and samples that are strong scatterers (owing to possible background issues), preferably filling the 5x5 cm incident beam. The SNS source power is projected to be well above 100 kW in the fall of 2007. With these parameters, the performance of ARCS will be world class even for the first users.

Procedures for proposal submission and evaluation are being developed now, as are plans and procedures for commissioning. Our informal estimate is that the instrument will be useful for early experiments in late 2007. Please consider ARCS in your long-term plans for inelastic neutron scattering experiments.

The ARCS data analysis software is in its first release, and you are welcome to try it. At present, it produces S(Q,E) from data files from LRMECS and Pharos. It saves files in a format compatible with Mslice. With special installation it runs on multiprocessor computer systems, but a version with a handy graphical user interface is distributed as precompiled binaries for Windows and MacOS. You can download this version, or explore others, at:

http://www.cacr.caltech.edu/projects/ARCS/arcs-1.0/web/Download.html

The construction period for ARCS has been a long one, but it has generally gone well without surprises that changed the technical capabilities of the instrument. We can now see a bright light at the end of the tunnel.

Sept. 27, 2005

Dear Members of the Executive Committee of the ARCS IDT:

We are having an ARCS breakout session at the SNS on Oct. 12 during the SHUG meeting. Besides reporting on the progress on the instrument hardware, software and the neutron source (all are still largely on schedule), we need to discuss a few other topics. These include:

1) Procedures for allocating beamtime to IDT members. It should be possible to do some science with ARCS in 2007. We need to develop a procedure for allocating beamtime to the members of the ARCS IDT. Doug and I offer the suggestion that this role could be played by the Executive Committee, at least at first.
- Would you be willing to help evaluate IDT proposals?
- Is there a better idea that we should propose for a review committee?

2) Update on magnetic field experiments with ARCS and SEQUOIA.

3) Your suggestions for other business (?)

Regards,
Brent

April 15, 2003

The ARCS Project has made some exciting progress in both hardware and software. Here is the status as of mid-April 2003.

Hardware:

As you know, the working plan for the ARCS spectrometer is to have the detector tubes mounted inside the main vacuum tank of the instrument, whereas all other chopper spectrometers have their detectors mounted in atmosphere outside aluminum windows. Better detector coverage, improved signal-to-noise, safety of the vacuum tank, and electrical isolation of the detectors are expected advantages of the detector-in-vacuum design. The technical risks were perceived as high voltage breakdown at intermediate gas pressures, overheating of electronics, and aging of components with dielectrics. A few months ago, tests of the detector tubes in vacuum showed that issues of breakdown and overheating are not problems.

There remain concerns that prolonged vacuum operation, or perhaps air/vacuum cycling, will lead to damage of the detectors or perhaps their electronics. The best way to test this is with prototype detector modules and electronics. Doug AbernathyÕs design team did the mechanical design of the module, and Rick Riedel and Ron Cooper of the SNS built couple of iterations of the detector modules. As of a couple of days ago, a Òpre-productionÓ unit was built and tested for spatial resolution, using the QUIP beamline at IPNS. Using absorbers to mask the beam, Ron reports that the spatial resolution is of order 1.5 cm at 1850 V for a broad range of neutron energies. The resolution was seen to improve with increased operating voltage, although higher voltages are expected to limit the counting rate capability. Resolution as a function of energy may be done later, but the basic performance looks good for now. Over the next month, the module will be subjected to a number of vacuum tests at Argonne, including air/vacuum cycling, and operation at a slightly elevated temperature. Assuming these tests go well, we will commit to the detector in vacuum design in early May. This will allow us to proceed with design of the vacuum tank, a long-lead item.

The core vessel insert and the shutter procurement are moving forward in synchronization with such purchases for the other instruments at the SNS.

The original design of the T0 chopper was developed by the shared engineering effort of the SNS. It uses a horizontal axis for a rotating blade. The beam is blocked only at the time of the prompt pulse. Owing to the prospect of the SEQUOIA instrument residing next to ARCS, however, we have been thinking that it may be better to use a coarse Fermi chopper for the T0 service. This unit will open only for the neutron energies of interest, keeping most of the neutron pulse out of the primary flight path. This seems like a good idea for reducing the background in ARCS, and reducing the effect of ARCS on the background of SEQUOIA (and vice-versa). Doug Abernathy and Garrett Granroth are working together closely on this new design, which involves a vertical axis of rotation, and a fairly fast rotation frequency to allow a beam of higher energy neutrons. With this vertical axis design, the ARCS instrument would not need a bandwidth chopper, although we may leave a space for it in the guide design. Settling the T0 chopper design is the last step required before starting the procurement of the neutron guide for the ARCS instrument. A consulting firm is analyzing the stresses in both designs to help determine the maximum rotation frequency of the T0 operation.

Likewise, a small company, Thermionics, has been selected to develop a conceptual design for the isolation gate valve for the sample region (so that samples can be changed without venting the ARCS vacuum tank). Robert Chave Applied Physics is moving ahead with the 3-axis cryogenic goniometer, and has a growing number of parts now at Caltech.

In summary, we are nearly ready to start procurement of the detectors, choppers, and guide. These are big-ticket items, and we want to be sure that the ARCS IDT is comfortable with these decisions. We have a tentative plan to organize an ARCS IDT meeting at the end of May. Perhaps this will be the last ARCS meeting at Argonne, since Doug is in the process of moving to Oak Ridge.

Software:

We have had two national meetings on the concept for Distributed Data Analysis Architecture for Neutron Scattering Experiments (DANSE). The meeting on Dec. 13, 2002 was described in the previous news message, archived at:
http://arcs.cacr.caltech.edu/arcs/ARCS/News.html
On March 13, 2003, we had another meeting at Caltech, attended by two representatives from each neutron facility in the U.S. and one from Chalk River. The full list of attendees can be found on the web site for the meeting:
http://arcs.cacr.caltech.edu/arcs/ARCS/DANSE_OrgMtg_Mar03.html
Although the different neutron facilities have their own development activities for data analysis software, the DANSE architecture can often incorporate these different packages. ISAW was modified slightly to run under DANSE, for example.

DANSE is an architecture for software, analogous to the technology for rough plumbing in house construction (PVC vs. cast iron vs. copper). Modularized software components (vaguely analogous to bathtubs or dishwashers) are embedded in Python, and communicate with other components through standard data streams. Because the streams can connect across networks, the components can reside on different computers and can operate in a distributed network.

We have demonstrated the essential functionality with the reduction of a data set from the Pharos chopper spectrometer at Los Alamos, converting it into an approximate phonon density-of-states. An independent calculation of the lattice dynamics runs under this framework too, and the experimental and calculated data can be displayed together on an equal footing, in ISAW for example. This architecture was also demonstrated and explained at the Experimental Facilities Advisory Committee Meeting at the SNS on March 27, 2003. (So far, the DANSE plumbing has held water.)

The plan today is to establish the DANSE architecture as the framework for the SNS data analysis software. Development of the DANSE system can be undertaken now within the tools available already. These are not so easy to use today, and we do not have a good hardware platform for development and testing by many users. (We need quick-connects for new appliances, and robust protocols so that nobody connects the toilet into the sink.) Nevertheless, it is hoped that a few brave souls will come to Caltech and learn how to set up code as cores to software components. Please let us know if you are interested. (You will learn skills vaguely analogous to sweat-soldering of copper plumbing.)

Extension of the DANSE architecture beyond the data analysis needs for inelastic scattering experiments will require two developments:
First, the hardware and software tools need to be refined so they are as easy to use as programming on your own laptop.
Second, we need to extend the scope of the science beyond inelastic scattering to all of neutron scattering.
Some ideas on how to do this are described in the white paper at the web site above.

December 27, 2002

The concept for a single crystal goniometer has evolved into the a triple-axis cryo-goniometer, based on commercial components from Kohzu. The goniometer will be capable of more than 180 degree rotations about the vertical axis without occlusion of the forward beam, and tilts of about +-20 and +-13 in the two other orthogonal axes. There is concern about secondary scatter from the goniometer components, and neutron shielding will be an important component of the testing in early service. The plan is to operate this unit on Pharos at LANL on a loan basis before ARCS is ready for it. Authorization for detailed design and construction has been given to Robert Chave Applied Physics. Solid-model CAD pictures are posted on the web site:
http://www.chave.net/client/arcs/
The triple-axis cryo-goniometer may be ready for the next run cycle on Pharos. Please think about how you might use it on Pharos. We will need the benefit of your experience.

Radiation shielding for ARCS continues to be an area of concern. The plans for the other SNS instruments are being reviewed to see how we can best benefit from the work that has already been done on basic materials and strategies for installation. The access to choppers and the interaction with SEQUOIA will be of special interest to ARCS. A large amount of work has already been done to evaluate by Monte Carlo methods and others the expected dose rates and backgrounds at detectors for our current layout. A new procedure is emerging for synthesizing this type of work, where an individual with experience in these calculations at the SNS takes charge of interacting with the instrument team. We have identified for ARCS the highest priority tasks to be evaluated, and we hope that the new group at SNS will prove efficient in getting the answers we need to finalize key instrument parameters.

These neutronic calculations need to be checked against real-world experience. We are starting to evaluate existing Pharos data to learn more about shielding configurations, especially about issues related to the performance of ARCS at the SNS. For example, does the amount of material in a T0 chopper need to be scaled up for the higher SNS flux, or can the current designs be used? Similar questions about beamstop design also need to be answered.

It has become clear that in cases when the sample region cannot be covered with thick shielding, as when a cryostat is in use for example, it will be necessary to have a hutch of some sort to suppress the radiation emitted into the target building. Please think about this. We will probably query the ARCS IDT about access and vertical clearance needed in the hutch. Hutch design may be an issue for other SNS instruments too.

The task of testing of detector modules in vacuum is now moving along well. Prototype preamplifier boards for groups of 8 PSDs have now been built, and are undergoing evaluation by Ron Cooper and Rick Riedel at the SNS. The readout and control board is not far behind. The extra time spent waiting for these boards was put to use in the detailed design of the packaging for the detector modules. We hope that near-final versions of the 8-tube module package will be ready for testing in January. The vacuum test system at Argonne is ready to go, and in the meantime it has been used for miscellaneous measurements of the outgassing of materials such as B_4C "crispy mix" so we can better estimate pumping requirements for the main vacuum tank.

Administratively, the ARCS project has now completed a baseline project plan that is consistent with what we have accomplished and spent at the end of calendar year 2002. A number of tasks were rearranged, some added, and some deleted, but fortunately the overall cost and schedule look pretty much the same as we had expected earlier in 2002. To explain how we will follow this plan, a Project Execution Plan has been written and submitted to DOE BES. It includes lists of milestones, change control thresholds, handling of contingency funds, and institutional responsibilities. The operational goal of the ARCS project remains the construction of a working inelastic chopper by Sept. 2006. The only uncertainties at this point are the memorandum of agreement between the SNS and Caltech, and the possibility of changing the budget authority from DOE to accommodate more hardware purchases earlier in the project. When last located, the MOA was in the hands of DOE administrators at ORNL, who are trying to help us with it. Some of these documents are posted on the ARCS web site (note new URL)
http://arcs.cacr.caltech.edu/arcs/ARCS.html
under the category: "Additional Reference Material in Adobe Acrobat .pdf format:"

The ARCS web site serves as a source of documentation on the project, but a more dynamic site is the ARCS wiki, located at:
http://viz.cacr.caltech.edu:8000/arcs/1
Please take a look at the wiki. It is a web site that allows on-line editing by anyone, including yourself.

On Dec. 13 we had a one-day software workshop at Caltech to discuss interactions between Caltech and the SNS on the development of software for data analysis. The agenda and some presentations are at:
http://arcs.cacr.caltech.edu/arcs/ARCS/SNS_CIT_12_02.html
Much of the meeting was an open discussion of high-level issues in the software architecture and how it may be pertinent to other instruments at the SNS. The architecture for the ARCS data analysis software is a: Distributed Architecture for the Analysis of Neutron Scattering Experiments (DANSE). This means that the user can select parts of the analysis or graphics procedures to be resident on his or her local computer, and direct a central server to arrange for the other computations and remote handling of large data sets. Our plan for access to the server through a web site is an approach that can be used beyond the needs of ARCS; possibly for the other instruments at the SNS. We are thinking about how to proceed with this broader direction.

October 14, 2002

The one-day ARCS IDT meeting at Los Alamos on 30 Sept. had a focus in the morning on goniometer designs for single-crystal manipulation, and a focus in the afternoon on management, hardware, and software issues.

Two goniometer concepts were presented in the morning. The insertion device, designed around a displex refrigerator, is described in detail at:
http://www.chave.net/client/arcs/
login: ARCS password: firefly
The present concept is a two-axis design, to be donated to the Pharos spectrometer at Los Alamos. With its pixelated detector, Pharos has many similarities to ARCS, and would offer experience with both hardware and software for single crystal experiments. Questions about this insertion device concept included how to add a third axis of rotation, which would be eventually required, and the ultimate base temperature that could be achieved with the displex unit. The cost for engineering and components (including displex and computer control) were detailed accurately by Robert Chave Applied Physics, and came to about k$ 100. This project may be pursued after the questions are answered.

Robb Williams presented some information about a sample stage based on a hexapod actuator:
http://www.caltech.edu/~matsci/btf/ARCS/hexapod.html
This unit would offer a general-purpose flange upon which cryostats, displexes, furnaces, or magnets could be mounted and rotated about three axes. Detailed engineering has not yet been performed, but the hexapod and vacuum bellows are relatively expensive. An 80 cm flange is prohibitively expensive with engineering and fabrication it may cost a good fraction of M$ 1. A system with a smaller 40 cm flange is probably affordable, and the intent now is to keep this option open as the designs for the sample region move forward.

The present role for the ARCS IDT was discussed. One possible activity would be working with the single crystal goniometer on Pharos to ensure a smooth transition to single crystal work on ARCS. A second role for the ARCS IDT would be to write a proposal for a high-field magnet. This proposal could include a large hexapod system for tilting and rotating the entire magnet system when installed on the ARCS spectrometer. Heinz Nakotte, Bruce Gaulin, and Ward Beyermann expressed some interest in pursuing this magnet proposal.

Two other hardware issues are of concern. The shielding calculations indicate that the ARCS beamstop will be big. Really big -- 4 m in length and 2 m in diameter. This long, wide beamstop could restrict the range of the detector angle bank on a high-resolution MAPS-like instrument on the adjacent flight path 17 (named Sequoia in the proposal by Steve Nagler). A complementary problem may be the effect on the ARCS low-angle detectors from the Fermi chopper of Sequoia. Another result from shielding calculations is the high radiation level, especially under white-beam conditions, around the sample well. There is a possibility that a hutch will have to be constructed to provide shielding when a cryostat is in place, for example.

The second hardware issue of concern is the prospect for operating the 900 odd position-sensitive detectors and their electronics within the ARCS vacuum tank. Although preliminary tests showed no serious problems with arcing or overheating, longer-term experience is lacking. We are ready with a vacuum testing station at Argonne, but prototype detector electronics are not yet available from the SNS. Concerns are the outgassing of plastics that are used in capacitors and high-voltage cables, and whether such outgassing may affect the properties of these components. It may be possible to test some components individually within vacuum, but any serious testing plan should include the full detector electronics at some point. More knowledge about vacuum operation is required before we can commit to the in-vacuum design, but this decision probably must be made in early 2003.

An important issue in software was resolved at the IDT meeting. One of the big efforts in software development is supporting multiple platforms. Do we need to release software for Windows NT, Linux, UNIX, Mac OS, etc., and maintain all these platforms with each subsequent software upgrade? This adds a large multiplier to the software effort. A solution was proposed, and a prototype system demonstrated, where the data analysis is performed as a web service. The user logs onto a web site that sends a java 1.4 applet to the users brower. This applet offers a graphical user interface, or command line interface, where the user specifies the data analysis procedure and required database information. At runtime, XML-RPC commands are issued from the browser to the server, which configures Python scripts and performs the analysis on the server. Output is sent back to the users browser for display or further analysis. This scheme separates the user specifications for data analysis from the actual implementation, allowing us software developers at Caltech to focus on supporting Beowulf clusters as the main computation engines, for example. The IDT members were willing to relinquish local computing, if the interface supports customization of data analysis procedures and good graphics. At present, the graphics requirements may be a bit demanding of internet bandwidth, but the customization is certainly not a problem. We are tentatively planning the software effort around this concept of data analysis as a web service.

Finally, we seek your help in finding a postdoctoral fellow for the ARCS software effort at Caltech. This is a scientific opportunity -- the work does not involve much more programming than expected in experimental inelastic neutron scattering work. Experimental measurements on chopper spectrometers are part of the effort in obtaining data sets for analysis. Alternatively, a theoretician with interests matched to inelastic neutron scattering could also be a strong candidate. Do you know of anyone who you could recommend?

July 18, 2002

It has been a while since the last ARCS news report, but I have been waiting to see the report from the Baseline Review last March. I am not sure that I have seen the final draft of this report, but I think I know what will be in it. I will forward it to you once I have received a final copy.

During the ARCS review contact was made with the group at ISIS building MERLIN, a new chopper spectrometer with very similar goals to ARCS. An agenda is being drafted for a workshop to explore common design issues and concerns between the MERLIN group and ARCS. This will most likely take place at ISIS in September. Anyone interested in more information should contact Doug Abernathy.

High Resolution Chopper Spectrometer

At the American Conference on Neutron Scattering held this June in Knoxville, a group got together to discuss a proposal for another chopper spectrometer at SNS to compliment ARCS. Many of you are aware of this, but it was certainly good news to hear from SNS management that DOE is looking to fund 4 new IDTs based at national laboratories. Steve Nagler at ORNL is the principal investigator for the proposal. He and Garrett Granroth of SNS are putting together the scientific case, conceptual design and cost estimate. Assuming this goes forward, ARCS will need to make some accommodations, such as changing the executive committee membership. Areas of common interest and concern are already under discussion, such as the development of T0 choppers and the coordination of shielding and beamstops.

Radiation Shielding

Estimating radiation backgrounds in and around neutron instruments is challenging because these backgrounds are often dominated by the high energy neutrons that undergo many possible weak scattering events. Radiation backgrounds have been troublesome for many instruments in the recent past, and we have therefore started on this early in the project. A number of calculations are being contracted with the ORNL shielding group, and there are some cultural issues getting the instrument issues translated into practical calculations. Nevertheless, we can already identify some important shielding issues for the ARCS instrument: 1) the background of fast neutrons in the forward direction will be quite substantial when there is nothing in front of the moderator. This corresponds to the case of no choppers in the beam and 1 MW of operation. Personnel safety issues will require a beamstop of 4 meters in length. Such a length will impact the positioning of the high resolution magnetism chopper spectrometer on flight path 17. 2) under the same white beam conditions, the radiation flux through a 1 meter hole above the sample (e.g., through a cryostat) is 1400 mR/hour. Although it is wonderful to see a calculation of so much scattering from the sample, this radiation level is far beyond what we can tolerate in operations. The T_0 chopper will reduce this background considerably, but it is not possible to plan for sample access with the assumption that the T_0 chopper is phased properly. I suspect that the sample region of the spectrometer will have some sort of room like a hutch at a synchrotron beam line, interlocked to the main beam shutter. 3) ARCS is located next to the wall of the proton beamline to the target, the RTBT. The shielding of the RTBT is good enough for the personnel background requirements, but there is a significant background of fast neutrons that could impair the signal-to-noise ratio of the instrument. There is a need for serious shielding between the detectors and the RTBT wall that may impact the decision to position the detectors inside or outside the vacuum tank, in addition to the concerns about detector access versus detector coverage.

Detectors

A few tests have been performed to check if the detector tubes and their local electronics can operate inside the vacuum tank of the instrument. Overheating does not seem to be a problem so far, but longer-term tests have just started. These tests included some purchases of detector tubes, their electronics, and a vacuum test chamber with pumping station. These are all located in a new laboratory space at Argonne. This laboratory will be used also for testing the outgassing of components such as borated shielding for the internal walls of the detector tank.

Single Crystal Goniometer

The Baseline Review suggested an earlier effort on the sample goniometer for single crystal manipulations. This is now an active subproject centered at Caltech. A subcontract has been set up with a former designer of cryogenic instruments at JPL, Robert Chave, who is pursuing a couple of goniometer concepts. The first design will be for a system with two-axes of rotation (one vertical, the other horizontal). Our plan now is to design and build it at Caltech, and present it to Rob McQueeney for service on Pharos. User experience will be important before deciding to duplicate the system, add a third axis, or alter the design. We will have a review of the goniometer sometime this fall.

** We need input on basic issues such as range and accuracy of angle control (we hope these will be -45 to +45 deg +- 0.1 deg), base temperature (perhaps 12 K). Your input now would be helpful, if you could please reply by e-mail. Also, let us know if you are interested in participating in the review of the sample goniometers at Caltech in the fall.

This effort is in addition to a current project to come up with possible concepts for sample manipulation that are permanently installed on the instrument. When these are more mature, with reasonable cost estimates, then a decision will have to be made to see what is built into the spectrometer and what will be done by sample environment based manipulation.

Software

The software effort is now moving quickly. We still do not have a detailed project plan with detailed work-breakdown structure, and perhaps we never will, but we are completing an assessment of tasks to help in more detailed planning. 1) The user surveys of software opinions, preferences, and comments were extremely helpful. Sincere thanks to all of you who returned them, and especially for your detailed comments. We have used this input as a key piece of our planning effort. 2) The plan is to organize the software in modules accessed by a high-level scripting language, Python. For the modules themselves, all code that we write at Caltech will be in the language C++. This requires a level of discipline and care that we are learning from Michael Aivazis, who will enforce the standards on our coding. We have weekly meetings now to discuss these issues in detail, and in the bigger picture of what qualifies as a module, for example. 3) The specific issues at hand are: a. rewrite/adapt the NeXus application program interface in C++, so it can be Pythonized. b. organize the successor to Open Genie. This might be tested on Pharos. c. pick a scientific application to develop as a project. This may be the analysis of inelastic phonon scattering data on pixelated detectors. d. Prepare a test version of the software structure, in part to test the user interface.

Management/Budget

On the financial management side, we are completing a subcontract that allows us to transfer funds from Caltech to Argonne. I am optimistic some legal obstacles will be surmounted, and the subcontract will be in place sometime in the next couple of weeks. The immediate need for this subcontract is to smooth out the Argonne funding so that Doug will be able to keep his present staff in place in spite of 1) a budget shortfall (that should be fixed on Oct. 1) and 2) the need for extra technical staff to assess vacuum outgassing and detector performance in the vacuum.

March 19, 2002

March 14 was the DOE Baseline Review of the ARCS Project at Caltech in Pasadena. It was done in one full day. The committee of 8 was chaired by Jim Richardson of Argonne. With the ARCS IDT members and SNS staff, the talks were attended by up to 24 persons.

The final report from the Richardson committee is not yet available, but the draft report and the discussion of it have already proved helpful to the ARCS project. We expect that the project will move forward along its present trajectory, likely with tighter cost management. The instrument itself will have essentially the same parameters as presented at the Baseline Review, although some new ideas on detector configurations and incident beam collimation were suggested and will be examined in detail.

The materials for the Review, and the Powerpoint Presentations by the ARCS team and others are at the web site:
http://www.caltech.edu/~matsci/btf/ARCS/Review/ARCSrev.html
This site is presently password protected with the login name:
RevARCS
and password:
CIT_ANL
Please take a look at the presentations. Most are in place today.

Three issues are emerging as priorities for the user community over the next few months:

1) We need to assess the present state of software packages for inelastic neutron scattering. A questionnaire is being developed, and will be sent to you in a few days. In addition to your responses, we need to determine how well the packages can be adapted into the Python language framework around which the ARCS software project will be built.

2) A second task of growing importance is that of sample environments, especially for the manipulation of single crystals. This issue has no obvious solution. A central question is whether the tilt mechanism is built into the spectrometer itself, or into the sample environment equipment. At present we are leaning in two directions. Eventually, we may need to do both:
2a) The first is to construct a tilt system with limited range out of the horizontal plane (perhaps +-5 degrees) and a (possibly large) angle of rotation in the horizontal plane. Experience at ISIS is that this is important for studies on single crystals, even for 1D excitations, since orientations can change during a change in temperature, and resolution is distorted when the sample is off-axis. Perhaps this capability could be built into the spectrometer itself.
2b) The second is to construct a gimbal movement allowing large rotations both in and out of the horizontal plane. It is seems most likely that this large tilt unit would be a component of a special cryostat, for example. It may come with some limitations on base temperature of the sample.
We thank those of you who have already given input on the sample goniometer(s). Please forward any additional thoughts to us. If you are interested in following either one of these directions, it might be reasonable for the ARCS project to support such work. We expect that the development of single crystal capabilities will be driven at least in part by somebody's scientific agenda.

3) Finally, as we are beginning to assess the design of the neutron guide, we are getting closer to defining the size of the beam at the location of the sample. A tentative plan is for the beam to be 5 x 5 cm. In your opinion, is this too small or too large? There are some limitations on beam size imposed by characteristics of the Fermi chopper, but we have some flexibility here too.

February 8, 2002

The entertainment industry of Southern California has a famous expression, "There is no such thing as bad publicity." The implication is that all publicity is good. Perhaps this is even true for the "entertaining" article about ARCS written for the Caltech alumni magazine, Engineering and Science. You can read about "Hot and Cold Running Neutrons" and how much better Superman would be if only he had neutron vision, rather than x-ray vision. Obviously I didn't write this myself, and some of it makes me cringe:

http://www.its.caltech.edu/~matsci/btf/ARCS/ARCS_ES.pdf

But seriously, folks, there are a couple of items of business for your information and for your attention.

1) You need to know that the baseline design for ARCS differs from the one in the original proposal. The new thinking of the majority of the Executive Committee is that it will have one 3 meter secondary flightpath rather than a 3 m high-angle flightpath plus a 5.5 m forward flightpath. The two flightpath concept was always a controversial issue with the ARCS design, as you recall. There are two reasons why we have to take this new approach:

i. It looks increasingly likely that a second "high-energy" chopper spectrometer will be built at the SNS. It is probable that Steve Nagler will be heading a proposal from ORNL to DOE to build an instrument for magnetic scattering, perhaps like the MAPS spectrometer at ISIS. This proposal will probably be submitted within a year. Even if there are bumps in the road to a second instrument, it will happen because this is the right thing to do at the world's premier neutron facility. ARCS will have much better energy resolution than was expected in the original concept for two nested instruments (2% versus 5%), and should serve the magnetism community well for several years.

ii. The two-flightpath design is too expensive to construct if we are to allow a project contingency of 25-30%. I don't like the possibility of using detector coverage as a contingency because there is a real risk that it may take too long to finish the instrument if it starts operation with only partial detector coverage. I really want ARCS to be complete, both in hardware and software, when it is commissioned.

2) DOE has become concerned about the management of IDT's, especially their interactions with the SNS. I have been asked to put the ARCS management plan in writing. A draft is available for you to see at:

http://www.its.caltech.edu/~matsci/btf/ARCS/ARCS_Manage.pdf

I hope you can read this plan, which was derived from the ARCS proposal. Let me know your comments.

3) Doug Abernathy has put together a work breakdown structure for the ARCS project, and is double-checking some of the cost estimates. This is the real thing, with contingencies on all the individual tasks. This means that it is about time for a baseline review of the ARCS project. The purpose of this review will be to assess the neutronic design of the instrument, and assess the cost and schedule for the project. This is not an IDT meeting in the usual sense, and attendance is restricted. Nevertheless, if you haven't heard from me in about 2 weeks, but you would like to participate, please give me a call and perhaps we can work something out.

Regards,
Brent

Date: Thu, 4 Oct 2001 13:22:21 -0700
To: ARCS Spectrometer IDT:;ARCS.Spectrometer.IDT:
From: Brent Fultz
Subject: ARCS News and Requests

Dear ARCS IDT and Interested Others:

The ARCS IDT meeting on Sept. 24 included up to 20 persons in the conference room at Argonne, and videolinks to up to 5 persons at the SNS site in Oak Ridge, and links to individuals at Caltech and McMaster Univ.
*Would those of you who prepared your presentations electronically please e-mail them to Doug Abernathy?

The meeting began with an overview of hardware issues. It was assumed that the secondary flight path issues will be resolved later once the Canadian involvement is clarified. This may be Oct. 31 if there is a quick No, but a Yes may require until the summer of 2002. (This is a concern that we will revisit after Oct. 31.) In the meantime, it was decided to proceed on the design of the sample region, including the vacuum gate, oscillating radial collimators, and the single- crystal goniometer. Requirements for the goniometer need to be settled soon. A vertical axis of rotation is relatively straightforward, but there is a clear need for a second perpendicular axis to provide at least +-5 degrees of tilt. Is this enough? For example, would a large angle of tilt, perhaps +-90 degrees, enable new science? An alternative to a goniometer mounted on the instrument would be to include sample movements in the sample environment equipment. Presumably this would be a cheaper option (at least for a few types of environments) but would have the disadvantage of losing the calibration with the detector positions when dismounted. The large-angle tilts may have to be done this way, however.
* Let us know your requirements for goniometer capability.

Software was the main focus of the meeting because it was necessary to settle on a software project plan. We are confident that Rick Reidel of the SNS will settle the issues of data acquisition, formats and storage. I am also pleased to tell you that Michael Aivazis of Caltech has agreed to join the ARCS project. In earlier days, Michael was a high energy physicist, but more recently at Caltech he has coordinated the software engineering and computing activities of a large software project under the Accelerated Strategic Computing Initiative.

The ARCS software development effort will concentrate on data analysis and visualization. Many different types of experiments will be performed with ARCS, and this demands a flexibility in data analysis. A first draft of a software roadmap is provided for you to see at:
http://www.caltech.edu/~matsci/btf/ARCS/Software.html
*Please look over this roadmap and suggest extensions or concerns. (We appreciate that some of the tools on the sides of S(Q,E) could move to either side.)

This broad software plan requires the software be arranged in modular units that can be connected in flexible ways. We propose to do this with Python, an object-oriented scripting language. Python bindings can be written for code in FORTRAN, C, Java and most other languages. (Matlab can be incorporated to some degree.) With these Python bindings, pre-existing code can be treated as Python function calls. Since Python is an interpreter, different combinations of data analysis tools can be connected in flexible ways, even automatically or dynamically as the program runs. An obvious first step is to "Pythonize" an existing piece of code and test its functionality. Michael Aivazis will look into this soon.

An important next step involving all of us is to identify all software packages that fit into the roadmap at the above URL. This requires an inventory of what already exists. Some of the packages are listed on the roadmap.
*Please look over this roadmap and suggest additional software packages. Even if these packages are not to your liking, we should know about them.

The question of what science will be performed under the ARCS project was not settled, although several ideas were discussed in the mid-afternoon.

Regards,
Brent, Doug, and Michael

Date: Fri, 20 Jul 2001 21:38:19 -0500
Subject: Re: ARCS is approved

Dear ARCS Team:

Bill Oosterhuis informed me that he will recommend ARCS be initiated on September 15, 2001 at the full level of funding we had requested, and on a schedule close to what we had requested.

I am on vacation next week, but I expect Caltech can prepare the final budget while I am out. There is more to discuss, but I wanted to alert you to the good news before I left.

Thanks for all your help in getting us this far.

The adventure has begun.

Regards,

Brent

Date: Wed, 23 May 2001 18:43:22 -0700

The Cold Neutron Chopper Spectrometer (CNCS) IDT met at the NIST Center for Neutron Research on May 21 and 22 to discuss specifications of the CNCS instrument. I attended as a CNCS IDT member, and also to preview upcoming issues for the ARCS project.

CNCS Issues:

Much of the CNCS discussion centered around the issues of energy range and Q resolution. The cold moderator for CNCS is capable of providing useful flux to perhaps 80 meV. There is hope that the instrument will be useful to 40 or 50 meV, although some technical problems conspire against this. It would be nice if the energy ranges of ARCS and CNCS had good overlap for more than 10 - 20 meV.

The Q range of CNCS seems likely to go down to 0.03 A-1 for long wavelength neutrons. There did not seem to be any objection to this as a lower limit, but if you are interested in critical phenomena you might want to think about this limit. The Q-resolution at the low end of this range will be quite poor, perhaps 100%. The general consensus was that there is no need to seek a Q-resolution below that set by the incident divergence (which is on the order of 0.5 degree).

In other developments, the CNCS detector configuration now emphasizes solid angle rather than resolution, and it looks likely that the detector bank will be -130 to +130 degrees at 3 meters, although only one side is likely to be populated initially. A ballistic guide with large aperture will be investigated.

Project Size: The disk chopper spectrometer at NIST has been completed quite recently (I think they have had two rounds of experiments so far). The NIST DCS instrument is similar to CNCS and ARCS in size and scope. After listening to John Copley and Christoph Brucker describe the details of the project, it seems that a 5 year construction schedule is about right, but the pace will not be leisurely.

ARCS Proposal:

The CNCS proposal was submitted a bit over 1 month ago, and the reviewing is not yet finished. If the ARCS proposal is submitted in 2 weeks, Bill Oosterhuis of DOE was optimistic that the reviews of the ARCS proposal could be completed by July, and we could receive some of the funds this year. His preference is to give 3.5 M$ to CNCS and 0.5 M$ for ARCS in the FY 2001 budget. This would be okay for the ARCS project if we get additional funds in early FY 2002 (that's only 3 months later). I suppose Bill thinks that CNCS has its act together a bit more than ARCS, although I think the ARCS design is probably firmer at this point.

The action item for the ARCS project is to complete the proposal and submit it. If you are thinking of making any changes to the science case, now would be an excellent time to let me know. I gave a draft of it to Bill Oosterhuis so that he knows what we are thinking about. The good news is that the present budget number of 14.8 M$ for model D, is not out of line. Bill was thinking that there would be 5 M$/year for both ARCS and CNCS over a 5 year period. I do not yet see the need to reduce the scope of the project below that of model D.

Question:

Is anyone attending the workshop on VITESS 2 and other packages for simulating neutron scattering at the Hahn-Meitner-Institut on 25-27 June?

Date: Fri, 4 May 2001 17:59:01 -0700

Here is where we are in the effort to build a direct-geometry chopper spectrometer of medium to high incident energies. Recall that previous efforts included the HELIOS and VERTEX proposals for instruments at Los Alamos, and two conceptual designs for the SNS. The effort now is to propose A high Resolution Chopper Spectrometer, ARCS, for the SNS. We began with a workshop at Caltech on Monday, April 30. This meeting was intense, productive, and ultimately convergent.

The technical challenge for the ARCS instrument is not so much energy resolution, which is below 2% at the elastic line for all instrument concepts. The greater challenge seems to be getting good Q resolution at low Q, while still having a broad range of Q. Given the geometrical constraints of the SNS target building and some financial constraints, we decided that most science can be accommodated with a horizontal instrument having two secondary flight paths. The detector bank at low angles may have a 5.5 m flight path, and the high angle bank will be at 3.0 m from the sample. The detector pixel density at the lowest angles may also be increased by using some narrower tubes below 5-10 degrees. (This instrument is essentially model "D" of Doug's "Possibilities..." document distributed before the meeting.)

The presentations on software development from the high energy physics community revealed a sociology unknown to me. It is tricky to get software engineers and physicists to cooperate. Physicists are perceived as hacks who write unstructured code that cannot be integrated. Software engineers are perceived as producing stultifying rules for software design. The only way this seems to work is if physicists write code, but pass it on to a software engineer who doesn't mind working with others' code and other people. It seems necessary for higher-level software development to be associated at least in part with individual instruments, although many modules should be portable.

We have put together two ARCS mailing lists. This message was sent to everyone on the long list. The short list will contain working documents and discussions on a daily basis, often with large image files. (If you don't get two messages today, you are probably not on the short list.) Let me know if you want to be deleted (or added) to either list.