Thx for your reply. I snapped the loose trace off. RPG on/off doesn't seem to make any difference. Also tried cleaning up with Isopropanol alcohol. All no changes to the behavior.
What did change it though were those dip-switches. I tried all combinations and those where the results. Maybe you can read something out of that. Also attached a bigger picture of the PCB and the Ducky manual entry about the dip-switches.
X= on
0= off
0000 = 1st row off
0X00 = 1st row off
00X0 = 1st row off
0XX0 = 1st row off
XXX0 = 1st row off
XXXX = 1st row off -- after USB out/in Demo Mode
00XX = 1st row off -- after USB out/in Demo Mode
0X0X = 1st row off -- after USB out/in Demo Mode
000X = 1st row off -- after USB out/in Demo Mode
0XXX = 1st row off -- after USB out/in Demo Mode
X0XX = 1st row off -- after USB out/in Demo Mode
XX0X = 1st row off -- after USB out/in Demo Mode
X000 = weird behavior (like described in previous post)
XX00 = 1st row clones 2nd row
X0X0 = 1st row clones 3rd row
X00X = 1st row clones 4th row / after USB out/in Demo Mode