Top Layer Problems

Ok, trying another one now with some different settings.  I've never actually done anything at 100% before.  Will make a solid one also and report back with results - maybe tonight, maybe tomorrow morning.

Let's say this was a larger/thicker part where weight and material usage would make 100% infill unfavorable - what suggestions then?

I'll post some pictures in a moment when my phone recharges (good camera is at work).  My next-to-top layer is nearly perfect as you'll see in a picture soon but what is so frustrating is that the top layer is absolute junk.  That layer (next-to-top) is at default and gcode shows it at 0.42 (same as single wall tuned). Currently I think that the extrusion width is too far from tuned values. The pic in the post above is at 105% width for top infill and the coming pic (pretty bad but enlightening) is at 80% (whice gcode indicated was 0.24mm!!).  I have just enough time to try the tuned value explicitly at 0.42mm (nope - going with Ward's suggestion) and will report on that when it is done.  Really, my gs4 was at 2% when I was last begging it to take a picture and it refused - back soon.

Thanks for the insight Ward. I'm going to hold off for just a moment to try out this line of reasoning and then I'll start seeing what 100% looks like and show any progress.

Food for thought:
I pretty much always use 3 perimeters.
I'll try the 0.36 and 0.48 on this next run.
I'm down to 40 mm/s - do you really suggest slower?

Here's the pic of nearly perfect next-to-top and crap top layer. I stopped the print shortly into the top layer.

EEPROM is calibrated at roughly 208 steps/mm  for 1.81mm (1.75mm) Octave Orange ABS with digital calipers over 10 cm.  Single wall is calibrated to 0.42mm. I'm surprised to hear 105ish for E.  I would believe mechanical except that everything except for the top layer is manageable.

*the current print at 0.36-0.48 isn't doing so well

I'll double check E steps and let you know

On a side note - when I was printing at 80% top layer width it seemed as though (and really before then also) that the z-height for the top layer thickness was too high such that there was no adhesion between layers.

Also - regarding E-steps - I'm at 1/32 on the RUMBA so 208ish should be fairly close to your 105ish at 1/16.

EDIT - in the pic above, I stopped the print shortly into the top layer

What is your temperature of the hotend? Had a problem myself with the top infill like your pictures. Had to set the temperature to 230 degree's. Started to go away at 223 degree's but the best was 230.

I think I have this as solved as it can be currently.  Also, the 0.36-0.48 print turned out better than it looked at first but showed a wider range of minor defects in other layers and still had the surface problem (albeit greatly improved) - I'm thinking that these are just too far from my tuned width of 0.42mm.

Slic3r seems to have issues once I start using widths more than +-10% from my tuned single wall width - so... I have now set ALL of my widths to that (0.42mm in my case) and will slowly try changes.  The defaults and text-over suggestions were just not working out.

I should note that I still have the original extruder and 0.35mm nozzle.

Here are some pics:

First, my Anti-Backlash Z-Spanner - no indication of binding at all.  I'm thinking about updating the design soon now that I have some good CAD/CAM to work with. The current design had to be filed down on the linear shaft facing surfaces at first because the squished first layers messed up my close dimensions. Since it was then (and now) it is slightly undersized it actually gives a small clicking sound when reversing Z direction as the spanner rotates a bajillionth of a degree and switches which side of the shafts it is contacting depending on the direction of rotation.

Second - the 0.36-0.48mm width print (0.36mm for top and bottom solids):

My final acceptable print at 0.42mm width for all layers:

I may add my current settings to this thread later tonight for posterity's sake.

0.35mm nozzle

Z-lift is enabled at 0.2963mm - the same as the corrected "0.3"mm layer height. I'm definitely not taking the spanner off - at least not until I make the next one. Between the spanner and the corrected layer heights the slight z-banding I originally had (when I first got the printer) has essentially vanished - and this is with the stock SD3 5/16" Z-screw.  I have several boxes worth of upgrades sitting around (including a PEI plate, linear bearings and shafts, a Bulldog XL, and an E3D V6, and other stuff...) but as soon as they arrive, I find a good configuration with the stock setup and don't want to change anything.

In the case that I want to publish a Z-spanner on the forum - do the SD2, 3, and 4 all have the same Z arrangement? As in: shaft diameter (3/8"), distance between shafts (71.75-72mm), screw size (5/16")? I can make versions for M3 and M5 as well.  This would be at least a couple of weeks away since I have to travel for work over the next two weeks.

Solution below.

Um... I do consider it fixed. Thanks for the help - I definitely appreciate it.  Not trying to beat a dead horse but the spanner actually greatly improves Z repeatability and isn't binding at all.

Check out the photo of my last print in post 15 and let me know if you see anything that looks like it still needs attention. The last two pics are from different prints and the bottom one is pretty good if you ask me.  More pics - first is all of the prints (leftmost is first test print, rightmost is final good print after the bad ones), second is close up of good print.

[SOLUTION]
Slic3r > Advanced > Top Solid Infill was trying to use unacceptable widths.

TL/DR: Default extrusion width was bad and using the % modifier in slic3r instead of telling it real mm value hid the problem until I looked at the gcode to see that widths were very different from values acceptable for my nozzle size.

Long Story: My Slic3r was just outputting some unexpected widths and setting all widths explicitly back to my tuned width solved the wierd behavior on the top layer. I'm not sure why it was making weird widths except that I had managed to change settings enough over time that it wasn't obvious to me that it was sometimes going down to widths of 0.24mm!! No wonder it wasn't sticking to the previous layer!  History of the issue - I had changed back to all default "0" values when doing maintenence last week and instead of using explicit mm widths, I had started using % widths.  The 0.24mm was from a Slic3r>Advanced> "top solid infill" setting of 80% which means that my 100% was apparently 0.3mm width and my original (when the problem first appeared) of 105% was only 0.32mm.  This also explains why the other infill layers (which stayed at 0.42 for some reason) still looked good.

Using Z-lift just became part of my preferred settings over time to minimize knocking around any slight warped area but I'll play with it today on the next part of this project enclosure.  The next part will be the bottom which will hold the Arduino whereas the top is what you've been seeing