Do you have a smooth plate to test with?
The divots in a textured plate are nearly the height of the layer height on default 0.20 nozzle’s layer height of 0.10. I print 0.2 on smooth plates or something like a frostbite satin plate.
Got a chance to try a different approach this weekend. Turns out. If I bypass using the AMS and route filament directly into one of the spots in the 4-way splitter, choosing external in print settings, there is absolutely NO Issue. So not an issue with extruder pursae, just odd that the AMS causes an issue with .2 nozzles? I verified there was plenty of slack when using AMS and two different positions and it caused an issue. If I bypass the AMS entirely, using the same formerly failing filament, no issues on multiple test runs.
change the g-code:
;===== for Textured PEI Plate , lower the nozzle as the nozzle was touching topmost of the texture when homing ==
;curr_bed_type={curr_bed_type}
{if curr_bed_type==“Textured PEI Plate”}
G29.1 Z{0.03} ; for Textured PEI Plate
{endif}
;===== for Supertack Plate, adjust nozzle offset ==
{if curr_bed_type==“Supertack Plate”}
G29.1 Z{0.03} ; for Supertack Plate
{endif}
Change the offset from -0.02 to, for example, 0.03
Been fighting the same battle. I have literally rebuilt everything, then along came a hero
This video fixed my problem, it’s not the 3 screws on the hot end, it’s the 3 on the back of it. It the sensor that reads the plate, which is why it will extrude freely until it hits the plate.
youtu.be/NheLZYVRdPA?is=XECYJnFfR5-0K5UH
I guess copy and paste 
it won’t let me post a link.
I was encountering the same issue, so I tried a few things listed here and it worked perfectly. First, I watched the YouTube link and tightened all the screws, but it didn’t work on the first try. Then, I changed the G-code as mentioned in the thread, and I think that fixed it!