The correct pseudo-file should read:
G "Conductor A"
C "conductor A to air interface" 1.0 0.0 0.0 0.0 +
C "conductor A to Dielectric1 interface" e1 0.0 0.0 0.0
G "Conductor B"
C "conductor B to air interface" 1.0 0.0 0.0 0.0 +
C "Dielectric1 to B interface" e1 0.0 0.0 0.0 +
C "Dielectric2 to B interface" e2 0.0 0.0 0.0
G "Conductor C"
C "Dielectric2 to C interface" e2 0.0 0.0 0.0 +
C "conductor C to air interface" 1.0 0.0 0.0 0.0 +
G "Dielectric1 to air interface"
D "Dielectric1 to air interface" 1.0 e1 0.0 0.0 0.0 x1 y1 z1
G "Dielectric2 to air interface"
D "Dielectric2 to air interface" 1.0 e2 0.0 0.0 0.0 x2 y2 z2
Note that (x1,y1,z1) and (x2,y2,z2) are the reference points lying on the 'air' side of the dielectrics. Take care that this is not simply 'outside' the dielectric shell; the reference point is considered with respect to each one of the panels composing the dielectric shell, and FastCap2/Fastercap assumes that this side of that specific panel is filled up by a dielectric with permittivity <outperm> (1.0 in this case). In general, it is recommended to define a reference point on the inside of a convex shell, or to take great care on which side of each panel the reference point lies.
Said that, you can find an example of how to define conductors in contact with different materials in the FastCap sample '_1X1BUS.LST'.
Regarding last point, the 'B' element would be useful only for infinitesimally thin conductors lying on the surface between two different dielectric mediums. This is an extreme case, since in practice all conductors have finite dimensions, and this is also why in the end the 'B' element has not been implemented.
Best Regards,
Enrico
