The subroutine package av18pot.f returns potentials for our first
few-nucleon benchmark calculations.  The potentials can be returned in
either operator or partial-wave format.  The first benchmarks are with
only NN potentials of the v8' structure with pp Coulomb potentials
containing a form factor.  The Coulomb potential is automatically
included in the partial-wave potentials; empot must be called to get it
if you are using the operator potentials.

Neither AV8' or SSCCV8' bind 6,8He adequately for a reasonable
benchmark.  So modified versions have been made for use without a NNN
potential.  The modification consists of some additional attraction in
the central T=1, S=1 (dominantly 3P_J) channel.  The additional
attraction was picked to give reasonable separation energies for 6,8He
and 6Li.  Of course these are no longer "realistic" potentials; they
give incorrect NN phase shifts.  But, as discussed in our meeting in
January, the error is not terribly large and the general structure of
the potentials is realistic.  So these should be adequate for
benchmarks.

The source for av18pot.f explains how to use the package.  The package
contains a subroutine (consts) that returns the values of hbarc, and
nucleon masses to use.  The K.E. is computed as

   hbar^2/2 Sum_i (1/mp P_p + 1/mn P_n) Laplacian_i

where P_p and P_n are projectors onto proton or neutron states.

For AV18 mp and mn are different but using just the average value as in

   hbar^2/4 Sum_i ( 1/mu ) Laplacian_i

gives a very small error; for 8He it is ~+40 KeV.  Of course for 4He,
6Li it makes no difference.


Bob & Steve