help analyze this print

This print is a box with 2mm sides. The bottom is open except for a 3mm overhang.
I stopped it at layer 14. The 3mm overhang is on the inside,
At this point the height is 2.90mm +- 0.06mm - 13 @ 0.2 + 1 @ 0.35 = 2.95mm
The 2mm wall is about 1mm high on the inside.

My questions:
1. why am I getting those separations? - 2 on the 3mm overhang and 1 on the 2mm wall

2. why is the separation on the top wall wider? All the walls have a different amount of separation. This corner has the most variation. The top wall is about 2.50 wide and the side wall is about 2.10mm - should be 2.00 - there are 10 traces there.

3. The blobs here at the seam are not very large, but is there anything I can do to reduce them further?

The inside dimension is 51.56mm, outside is 56.04mm == design is 52.00mm in, 56.00mm out. So it seems that separation is affecting the inside dimension.

CoreXY machine, PLA. no heat, lots of fan, Marlin 1.0.2 RC3 firmware w/Arduino/Ramps
Repetier Host 1.6.1
Slic3r 1.29 settings:

layer height = 0.2mm
first layer = 0.35mm
perimeters = 5 for vertical, top & bottom
seam = aligned
infill = 100%
speed = 60
advanced settings = all 0

temp = 190 (first layer = 195)
retraction = 4mm
lift = 0.2
speed = 80
extra length = 0

accel = 6000 - except for 1st layer = 4000

I'm not sure what I am looking at with that photo - could you maybe post a photo of the whole print so I can work out the perspective?

Dave

that pic is showing the upper rt corner of this one.

Makes a lot more sense with that pic!

Maybe look at your G-code file and locate the lines that print the outer perimeters on each side of the inner rectangle and check whether they are about 52.5mm apart (it should be exactly 52mm + extrusion width). If the G-code file places them in the correct position, then your printer is not positioning the head correctly and I suspect the problem is slack belts on your X & Y axis. That would cause positioning errors, with the error changing depending on the direction the head approached the start of the line, which could explain why there are consistent gaps if Slicer uses different directions at those points.

If the G-code file is wrong, then the issue is either in your STL file or a bug in Slicer. I assume from the photos that you have set a concentric fill pattern for the solid layers - maybe try changing that to rectilinear and see what result that gives you. I have seen gaps in the concentric fill on a number of occasions, and I suspect that to avoid them the dimensions to be filled must be an exact multiple of the extrusion width. You could therefore also try changing the extrusion width (usually set to automatic but you can enter your own value). The extrusion width chosen should be equal or greater than your nozzle diameter.

Dave