Perimeter bonding problem on concave perimeters

Hi, I'm printing ABS on a Mendel90 at temp. 244, small or visible perimeters at 30 mm/s. After playing around with various numbers I'm mostly satisfied with my print quality, except concave perimeters get no good bonding to the rest of the object, especially small holes for screws or similar look messy. Any more experienced 3D printer expert got an idea? Would be really appreciated, as I'm very new to this kind of stuff.

Which slicer are you using, and are you setting the extrusion width manually or not? The small holes are fairly typical, I normally live with it and clean them out with a drill, but the larger U-shape does seem a bit off.

I've been using Slic3r 1.29 for now, and after some more reading wikis etc. I read that Slic3r has (had?) a problem if the extrusion width is set to "auto", so I changed it to be 0.4mm - my nozzle width. It didn't really change a lot. After some more reading I changed the layer heigth from 0.3mm to 0.2mm. Things improved, but of course printing time shoots up through the roof. It got somewhat better, but not really good. It may have something to do with filament quality. After maybe 10 or 12 hours of printing and testing I noticed the small holes getting better, without changing a parameter really. Right now I switched Slic3r back to 0.3mm klayer heigth, I want to see if the small holes get worser again.
And yes, I read that in certain conditions Cura has less problems with small holes, so I'll try Cura later this weekend.
Thank you for your suggestions!

Addition: My nozzle looked like it was shoving melted plastic when moving in +y direction, but after some more investigation it looks like there is a leakage at the nozzle, about 2-3 mm distance from the nozzle tip. After disassembling & cleaning I notice 2 partially circles at the nozzle tip - looks like it's cracked. Anyone recommends a good hotend? ;-)

Edited 2 time(s). Last edit at 01/15/2017 05:11AM by Inhumierer.

OK, so the 'sweet spot' (and people will probably argue with me here) for ease of printing is roughly layer height of half the nozzle diameter and extrusion width of between 1.2 and 1.5 times the nozzle diameter (the 1.5 is on the high side, but a lot depends on nozzle geometry).

If you go squarer than that, things tend to get difficult because the flow of plastic is becoming under-constrained and so more variable, and the ratio of contact area between layers to volume of molten plastic is getting smaller. Which is another way of saying that 0.3 height by 0.4 width on a 0.4 nozzle is starting to get into the wild zone. Yes, print times are long with thin layers, but the answer to that is bigger nozzles.

Interestingly, I think cura defaults the extrusion width to the nozzle diameter, and Slic3r does the same for external perimeters which seems to contradict my advice, but I think they are trading off print detail vs ease of printing. I'd recommend trying thin wide extrusions to see if it helps with print integrity, and then tune for appearance from there. (Or not, if you're after strong functional prints and aren't so worried about what they look like.)

Good advice, thanks again. Unfortunately my next experiments will be delayed somewhat. After cleaning & reassembly I manually moved the nozzle short way above an unheated bed and pushed some filament into the heated extruder, carefully by hand. ABS flakes aroused at the side of the nozzle. Did a video with my endoscopic cam, but the quality is very very poor. Afterwards did some photos - not the best quality, but anyway I'd like to share.

What an interesting looking hotend you have! You're using a collet vise to hold the throat? Looks like a very neat idea Without a low conductance heatbreak and active cooling, you may find you have rather a long transition zone in your hotend which might give you a problem with ooze control, but let's not leap ahead.

Your ABS looks very strange indeed. I'm not at all sure what happened there, but I'd try pushing a few 100mm of filament through and see if it becomes more normal - perhaps there are some contaminants from the cleaning that need to be purged. Extruding into free air can come out looking very strange if you go either too slow (the plastic starts to decompose in the hot-end) or too fast (I suspect turbulent flow instead of the desired laminar flow). It's always nice if the filament extrudes straight down, but I find it rare with anything other than the least viscous of plastics. It doesn't seem to make a lot of difference to actual print performance.

Not really a collet vise. There's a ring at the top of the hotend, 3 screws with washers press it against the holding. Hotend should be Wade compatible, it's the one that came with my Mendel90 kit: building instructions.
The thermal barrier is made of some very hard duroplaste or ceramic, stays pretty cool on top even without a fan. And yes, you're right, it's oozing.
It's not cleary visible at the photos, but the plastic flake came out at the flat side of the nozzle, about 2 or 3 mm away from the tip, left side at the 2nd & 3rd picture. It builded up some kind of circle and touched the other flat side then. At the tip almost no ABS came out.
So I'm pretty sure I'll need another nozzle, and I want to test some other extruders/hotends, so I already ordered some parts. And maybe I'll get a new nozzle from the shop, as I suppose the plastic coming out at the side of the nozzle is probably a manufacturing problem.

Oh, the nozzle sprang a leak from the side? That's unusual, must have been drilled too close to the wall I guess. What is the large knurled section above the hotend, below the heatbreak, for?

It's a brass union nut - hope this is the right translation
It fixes the nozzle and is screwed from below against the heatbreak. You can see it in the building instructions linked above.
The shop will replace the nozzle, and I'm waiting for my ebay orders to arrive.