Trying to achieve thinner layer heights but still maintain usable support...

54 views
Skip to first unread message

Chris Keller

unread,
Jun 9, 2014, 5:54:31 AM6/9/14
to kisslicer-r...@googlegroups.com

Hey folks, 

A few questions


In the past I've had great success using a layer height of 0.175mm, and extrusion width of 0.49 mm and infill ext widths of 0.49 mm.
I am now working on getting a layer height of 0.08 mm to work, with stacked sparse infill and stacked support layers set at 2
- Problem is, I'm getting gaps in infill and support (incomplete extrusion paths - iow not enough material flowing out the nozzle to string it from a to b)
Also this occurs at any speed.  it's just too little material coming out to bond to the layer 2 layers below (actually isn't it 4 layers below? because the last layer it's only crossing, and each layer is 2 layers thick)


INCOMPLETE INFILL PATHS


I've now changed the infill extrusion width to 0.8mm and the infill now looks great.  
But I'm still having the same problem as before with the sparse support.

Is there any way to make the support follow the infill width? is supposed to already?  As I said, I've increased the width of the infill and now it looks ok.  But it definitely leaves me wondering if the calcs for the width/height are correct.

I'm trying to dial in a fine-resolution set of settings.  the 0.175 layer height with 0.49mm ext width works GREAT (with the exception of same material support which isn't great, but it's still 100x more outstanding than Slic3r's poor excuse for same material support!)

Lonesock, would it be a huge pain to make the support thickness a separate parameter? - I HATE to ask because I know most of these things, it makes things worse to open it up to the end user.  But I can't find a way to go below 0.15mm layer heights with support and have the support be as good as it was in 0.175mm layer heights.

OR, and I don't know how difficult this would be to check, but perhaps the extrusion width math on stacked layers isn't what it's supposed to be?  <<--- I'll go look at the gcode, but in the meantime, in case someone can help me out... I'll post this without that investigation (will add it later)

Thanks
Reply all
Reply to author
Forward
0 new messages