Proton Driver Meeting

17 - May - 2001
Fermilab

Presentations:-


Weiren Chou: Announcements

Cost Estimates for the Reduced Proton Driver Upgrade

Recent work on "Appendix A" - the 8 GeV version

The group has investigated many different lattices, and found two acceptable ones. At present these two look equally good:

  1. "triangular" 474m, two doublets, 1.5 Tesla fields
  2. "new-2H" same as before, but chop off about 1/3 of the lattice

The magnet inventories for these two lattices were given to obtain new cost estimates. People generally did their work very quickly. Only two areas need new estimates, but they will not impact the costs estimates much. Reminder: the original cost estimate was $242M

  1. "triangular" ("doublet") $163M
  2. "new-2H" ("racetrack") $157M
These costs could come down a few million once the updates estimates for the RF are available.

The new lattices and cost estimates will form the main content of the new "Appendix A". A comparison and discussion of the pros and cons of the two designs also will be provided. Once the writeup is completed, then this task is finished.

Note: There is no major LINAC upgrade included in this proposal. Including one would increase the cost substantially.

Note: To completely rebuild the RF from scratch would increase the cost estimate by only $4M. Is it surprising that the magnets turn out to be quite cheap: roughly $50k/each. The power supplies are quite expensive.

Superconducting LINAC Upgrade

Steve Holmes asked for an investigation of the benefits of installing superconducting RF cavities in the LINAC. He was hoping to boost the energy of the LINAC by 200 MeV (from 400 to 600 MeV). The idea is to use cavities from SNS, rather than design something from scratch.

There was a meeting on May 9 to discuss the technical aspects of this idea. After much discussion, the best option seems to be to take out the existing cavities in stations 6&7, which provides a linear section 83m long for installing SC cavities. Over this length the beam gradient is 285 MeV. The current cavities provide an acceleration of 4.5 MeV/m. One would hope to double this to 9 MeV/m. The SNS cavities, however, give a gradient of only 5.2 MeV/m. The study group devised two techniques for boosting this gradient:

  1. increase the peak surface field (to 35 MV/m)
  2. increase the fill factor (which is the ratio of the accelerating length to the physical length), by adding two more cryostats.
The overall gain factor would be 1.27 x 1.36 = 1.72 .

The increase in the energy of the LINAC would amount to 90 MeV -- far short of the 200 MeV Holmes was aiming for.

The study group sent this result to Holmes. He and Marriner will think about this.

There is no firm cost estimate yet. Weiren pointed out that the AGS is planning a linear collider of 1.2 GeV energy using the same SNS cavities, and they estimate a cost of $90M. The SNS is spending $350M.

Discussion:
Given the lack of a LINAC upgrade, and the pessimistic projection for the SC Linac, there seems to be no potential gain for Run II. The director has explicitly cut out Run II from consideration of a proton driver upgrade. Given the modest gain and high cost of the SC Linac, it seems unlikely that it would be funded and executed in time for Run II. It is more interesting as a development parallel to the booster upgrade.

Given the space constraints, installing SC RF cavities to achieve the 90 MeV gain would require nontrivial R&D, which could not be concluded in time for Run II. So the SC option makes more sense in a post-Collider era, in which case one should consider a 1 GeV LINAC.

Why not pursue the `curved LINAC' idea? There are concerns about the very small aperture. Opinions differ about how much of a problem this is. It is not obvious feasible, so work is required before making the case for such an upgrade. To be fair, there are problem of similar magnitude in the SC proposal, at the junction between warm and cold sections, for example.

The cost is a major problem, also, as far as Run II is concerned. The relevant scale for more upgrades is $20M, and $90M is far above that.


Peter Kasper: Proton Economics With and Without the Proton Driver Upgrade

Transparencies

Peter has taken a long term view of what the Lab might be doing. He notes that an important problem is radiation control, which may prevent running at 15 Hz.

Peter built a concise table of machine characteristics comparing several scenarios: the present, the present with 15 Hz instead of 7.5 Hz, the original proton driver upgrade ($242M), alternate design 1 (with an upgraded LINAC), and alternate design 2 ($160M, no LINAC upgrade). There are then three `proton spigots': pbar production (BTeV), fast extraction (NuMI), and slow extraction (miniBoone, KAMI). See his transparencies for details. We try to summarize the main points here.

First, assume no proton driver upgrade. What would the situation be?

case 1: pbar production (collider, BTeV)
dedicated 1.17 x 10**16 p/hr (2.08)
+ fast extraction 0.96 (1.64)
+ slow extraction 0.63 (1.13)
One sees the clear 20% cost for each of NUMI and miniBoone. The numbers in parenthesis are for slip stacking. The collider would have trouble reaching 15f-1 if there were less than 10**16 p/hr.

case 2: fast extraction (NuMI)
dedicated 3.2 x 10**20 p/yr (4.1)
+ p-bar 2.7 (3.2)
+ slow extraction <1.7 <(2.2)
NuMI requests 3.6x10**20, which would require slip stacking.

case 3: slow extraction (miniBoone)
dedicated 2.1 x 10**20 p/yr (2.8)
+ pbar (BTeV) 1.7 (2.2)
KAMI requests 2x10**20. Clearly there is a problem in trying to run BTeV, NuMI, and KAMI together. At best one could devise a complicated scheme which would run alternatively fast extraction, slow extraction and pbar production more or less concurrently.

case 4: low-energy source (miniBoone)
+ pbar
5.2 x 10**20 p/yr
+ MI fast extraction
3.2
+ MI slow extraction
4.5
miniBoone requires 5.0x10**20 p/yr, so running MI extraction would clearly cause big problems.

Suppose now that miniBoone confirms the LSND result. This would be a major discovery, and of course the Lab would like to pursue it vigorously. With the proton economics outlined above, there would be a major conflict between physics goals. Does on postpone the NuMI and KAMI programs? What about BTeV? Does one wait several years before following up major new physics discoveries? Running at 15 Hz would help, but it seems very unlikely that the radiation problem has a viable solution.

In short, Fermilab is in a position to run only one high-intensity physics program at a time. Consequently, unless one pursues the proton driver upgrade, the Lab cannot react effectively to discoveries.

Now assume the proton driver is upgraded.

In this case one could comfortably run BTeV, NuMI, and KAMI at the same time and also follow up any discoveries at miniBoone. In terms of proton economics, Fermilab would be rich, and could seriously consider building a neutron source as discussed in a previous meeting. Also, the Lab could add an antiproton production facility on the side with relatively little trouble.

These arguments are based on physics programs and experiments already under discussion. With the proton driver upgrade there would be room for new ideas and expansion into new areas over the next decade.


Send comments about these pages to Peter Kasper