This is the combined VLBI/SWARM bitcode that fixes the false coherence problem by modifying the cgain_core (changed an overflow from wrap to saturate in the negate block) and the top level block (fixed data bias by adding negate blocks at the output of the F4_3 decimation stage). Timing score from planahead was 503. The bitcode has been validated for SWARM only operations, validation for VLBI mode still pending. The parameters were
map -intstyle pa -w -pr b -t 100 -ignore_keep_hierarchy -timing -ol high -xe n -mt on system.ngd