I changed the Custom '139' and still have the same issue.
After tracing all the tracks from the unlabeled pins on the 139 I've narrowed it down to a couple of suspects.
Either the Custom '148' at 4F or the PAL at 6J.
From Namco C148 Driver:
Namco C148 - CPU Bus Manager
Starblade relies on this for showing large polygons
Does some Memory Decode, Interrupt Handling, 3 bit PIO port, Bus Controller
Is there anyone who has a Starblade PCB that has the ability to dump the PAL at 6J on the CPU board for me?
I'd like to rule it out as a possible suspect, output pins 18 & 19 (which are connected to pins 54 & 57 on the Custom '139' respectively) are stuck high and I have a feeling they should have data coming out of them.
On Starblade this is a PAL20L8ACNS, image below is from a System 22 schematic but but both Systems are very similar.
Can anyone guess from experience whether they feel these outputs should have data?