Presuppositions As noted elsewhere, this solution has been generated under two assumptions:

1. the "upgrade" under consideration is to be implemented BEFORE machine studies characterizing CSR and space charge effects are completed, and
2. the upgrade path is to entail minimum reinstallation of machine components

Design requirements are similar to those in the 42 MeV driver:

• deliver the specified phase space to the wiggler (now at 75 MeV)
• energy recover power from "spent" beam
• address fundamental physics issues - CSR, BBU, space charge, beam loss, RF stability,... all of which are more or less the same as at 42 MeV (though space charge maybe reduced due to higher energy)

Design Variances An increase in the final energy to 75 MeV through use of a second cryomodule drives three significant deviations in design features from those present at 42 MeV. These are as follows:

1. RF Focusing
   • The use of RF focusing for beam confinement is integral to the single module 42 MeV design
   • In the 2 module 75 MeV design, RF focusing is essentially absent in the second module during acceleration (from 42.5 MeV to 75 MeV) and the first module during energy recovery (from 75 MeV to 42.5 MeV).
2. Injection/Extraction Chicane Geometry
   • the difference in injected/final and final/extracted momentum ratios (75/10 vs. 42/10) leads to geometry differences in the injection/extraction chicane layouts
   • bend radii/angles, path lengths, dipole focusing effects and momentum compactions change
3. Backleg Transport
   • the longer linac means that the backleg is longer (by ~10 m, the pitch of the transfer line feeds)
   • we try to use the same backleg design concept (periodic FODO transport with M=-I), as it helps suppress aberrations
   • we can retain the 42 MeV design of 6 90 degree FODO cells, lengthening them (thereby increasing betatron mismatch) or put in more FODO cells with lower phase advance to keep the overall transfer matrix M=-I, with less mismatch

We have a reasonably detailed upgrade design solution addressing all three of these differences. A conceptual layout is shown above (with one for the 42 MeV system, for purposes of comparison). In the solution, we split the linac and stretch the backleg to make space for the second module and the additional focusing required to offset the loss of RF focusing at higher energy. For the acceleration cycle, a triplet is placed between the two cryomodules. As noted in the description of the 42 MeV lattice, reinjection matching has been included in the 42 MeV design in support of this upgrade. This match is not necessary for the 42 MeV transport; it was included to reduce the reinstallation required during this upgrade to 75 MeV.

This upgrade solution addresses the issue of mismatch in the backleg by increasing the periodicity of the transport. Design studies for 42 MeV demonstrate that a backleg with M=-I is desirable so as to suppress aberrations. This is accomplished in the 42 MeV system by using 6 90^o FODO cells, which, at the cell length for the 42 MeV backleg leads to peak beam envelopes of ~10 m. Were this solution simply stretched to fill the longer backleg, peak beam envelopes of ~20 m would arise due to aggravated betatron mismatch in the longer cells. To avoid the mismatch, we increase the transport periodicity and commensurately shorten the FODO cell length, generating an M=-I module using 9 60^o FODO cells. The cell length is in this case nearly the same as in the 42 MeV system, and the peak beam envelopes stay at the original value of ~10 m. Moreover, the lower phase advance at higher energy leads to quad pole tip field strengths that are roughly the same as the higher phase advance solution at lower energy. The required quadrupole dynamic range is thereby reduced.

The use of such a solution requires complete reinstallation of the backleg. An alternative solution is therefore under investigation. In this approach, the backleg is split at the center, and the upstream half reinstalled (together with the first arc, FEL, and new module), with the nominal 42 MeV cell design pitch that is also retained in the (now unmoved) second half. The 3 cells in each half are now run at a reduced phase advance of 60^o/cell. The gap between the two halves is to be filled with 3 more FODO cells (now of somewhat different, most likely shorter, length) that will also be run at a tune of 60^o/cell. Because each "flavor" of cell is folded into an M=-I module, the overall transport matrix remains M=-I; because all cells are of the nominal 42 MeV length or shorter, the mismatch will remain small. The quadrupole dynamic range may be somewhat increased (the short cells requiring higher strength), but a third of the backleg no longer needs reinstallation.

The final design variation - the reinjection/extraction chicane geometry - requires a change in the element layout to accommodate the change in beam momentum ratios. The overall slot length and end magnets are unchanged; the center dipoles, dipole locations, and path length change. Focusing differences are compensated in the upstream match into the wiggler and the reinjection match. Between the 42 MeV and 75 MeV layouts, these changes lead to a path length difference of ~6 mm (modulo the RF wavelength) and an M₅₆ difference of ~6 mm. These are readily compensated by the available correction systems.

The end loop design and the "first light transport" are the same as in the 42 MeV system. No new elements are required. The total element difference is as follows:

• 9 more quads
• ~8 more correctors
• ~7 more BPMs
• 2 different injection/extraction chicane dipoles

The magnets are the same as for 42 MeV. The machine dynamic range therefore allows ramped operation to lower energy, even down to 42 MeV or below. This may require the injector energy track the machine energy (a practice that could be space charge limited) or may require minor reinstallations of injection/extraction chicanes to accommodate changes in injected/extracted/full energy ratios.

Last modified: 10 March 1997
http://www.jlab.org/~douglas/ is maintained by: douglas@jlab.org