Generating microsurfaces  

RE: Generating microsurfaces

You are running up against the resolution of the printer and possibly the practical limitations of the technology.

Currently theSL1 is advertised as having a 'pixel' limit of 0.047mm and a minimum layer height of 0.025mm - so anything finer than that is, for the moment,  impossible.  Pyramids, or any other shape, have to be calculated in terms of stacks of these pixel blocks so the nearest, smallest, approximation of a pyramid (one block placed on nine) would give a pyramid of 0.141x0.141x0.072mm with a 0.045mm seperation.

Part/feature repetition is much easier programmatically, such as this OpenSCAD script:

// many pyramids

xnumber=10;
ynumber=10;
side=5;
spacing=10;

for( x=[0:xnumber-1]){
  for( y=[0:ynumber-1]){
    translate([x*spacing,y*spacing,0]){
      pyramid();
    }
  }
}

module pyramid(){
  cylinder(r1=side/2,r2=0,h=side,$fn=4);
}

Which gives you the attached image ...but if you're planning millions of features you'll neet plenty of memory and expect huge object files.

RE: Generating microsurfaces

I figured something like:

color("green")
    linear_extrude(height = 1, scale = 1.0)
            square([20, 20], center = true);

color("red")

for (i = [0 : 1 : 199])
    for (j = [0 : 1 : 199])
        translate([-20/2+0.05 + i*0.10, -20/2+0.05+ j*0.10, 1])
        sphere(r = 0.05);

Thanks @diem for code for generating pyramids. With spheres it was easier for me 😉

But even with array 200x200 objects it take some time to generate object...

I know that I'm pushing the printer limits - I'm trying to achieve something like: https://www.sciencedirect.com/science/article/pii/S0264127519302771

RE: Generating microsurfaces

@m-grzegorczyk

You'll always have problems generating a structure with so many faces - the file will be huge.  In this sort case it would probably be quicker to generate the image files for the LCD directly, this is really a set of patterns creating a texture so creating the geometry is an extra, computationally intensive, step.

The LCD in the SL1 doesn't have the ulta-fine resolution of the example in the paper you linked - it's an order of magnitude finer than you would expect in any hobby printer.

It may be possible to modify your SL1 to accept a replacement LCD or to insert a lens but that's a whole new project.

If you want to experiment with hydrophobic surfaces it might be easier to phot-etch a sheet of metal and use that, heated and then cooled in contact, to emboss a thermoplastic surface but there is a chance that a hydrophobic surface structure might also be molten-plastic-phobic too which would interfere with the process.