The use of the FTxx-43 should not make too much of a difference. But it does tend to fall
off in efficiency at the higher frequencies. If you stick with that, you are probably better off using
FTxx-61 material, but FTxx-10 is good overall for the full range.
Yes, the transformer windings in Farhan's schematic imply a trifilar winding, but actually
some of the transformers are different, and only 2 of the trifilar windings are used in most places.
But you got it right. The findings of lower output at higher frequencies is perfectly consistent.
I am glad you are reading of the trials of WA2EBY. Now you see what a problem it really is.
The fact that you are using mainly smd materials also contributes to the layout problem...but
you are well on the right track. Congratulations.
Now beware of ground loops and the oscillation problem. Actually, the ground loop problem is
a bigger one. The usual solution is to somehow come up with a common ground point for ALL
the signals, both AC and DC. It may be that some more shielding is needed between the PA
and the rest of the rig (a la early Yeasu rig ideas). But you will find that out.