it's the same thing in my bios (3.1g) :
- check if cart is present => it doesn't write an error so cart detect is right
- vdp display disable => screen is green so no problem on this one
- memoryModeReg (0x2001) = 2
- loop forever
it loops on green for ever I thing so the (0x2001) = 2 doesn't do what it should...
correct me if I'm wrong but, it should
- reset M3 on SMD port (to switch to genny mode) => pin B30
- soft reset => pin B27
so I must check if something is wrong around these 2 pins I assume
else...the EPLD is dead

especially B27 since I doesn't reset (not even in M3 mode)....but it doesn't seem to be connected, according Charles' schem...

EDIT : just got an idea...
do you think a input of the EPLD (anyone) could be wrong and so the EPLD doesn't receive input like it should ?
for example, let say, because a trace is wrong, a pin is to low (or more true, unknow status) while it should be high when the EPLD received (0x2001)=2 and so it doesn't active the good output (and so reset and M3) ?
something like (pseudo code)
if ( Adr =0x2001 AND Data = 2 AND PinX = 1) then M3 = 1