I took a shot at this, continuing my exploration of “letting the Z resolution do all the work”. Meaning as I understand it: the X and Y resolutions may have smaller motor steps, but the detail will never surpass the .35mm of the extruder nozzle output. The Z resolution, however, will squish the filament in that dimension to achieve more resolution, theoretically down to .05mm. So, I try to orient my small prints to take advantage of the Z axis where the most detail is needed. That usually means a 45-degree rotation in one axis so the layer striations remain parallel or perpendicular to the final print orientation. Angles less than 45 degrees tend to retain some of the defects of flat printing, and angles higher than 45 degrees introduce overhang issues.
I also noticed two things about filament printers which seem universal: 1. They can’t do flat surfaces smoothly, it’s always a criss-cross basketweave texture. 2. They can’t properly finish domes, pyramids, cones, or anything with a point at the top because the print head remelts the penultimate layers into a random glob. These are items which 45-degree printing also hopes to address.
I’ve had enough success with this method to feel the theory is sound. It just seems there are other quirks of the printer that throw wrenches into the works. But a photo is worth 387 words (adjusted for inflation), so here they are:
The sword as printed, with all supports still attached. It wanted about 5 hours to print at 50 microns so I left it largely unsupervised. My lack of diligence was rewarded with some sort of flow error for the last 15% or so of the print.
View from the back, giving a better view of the support structure.
The sword removed from the supports. The detail level, where it worked properly, was what I expected, but the print itself is unusable, probably thanks to settings I need to tweak to address stringing, flow/temperature, and support overadhesion.
At this time my experiments are designed to extract maximum resolution from this printer. That means the printing is much slower than normal and clearly in need of monitoring and adjusting. This sword has a shape that would allow me to print it in two halves and glue together, eliminating the backside support damage but further doubling printer time.
Much of my printer troubles appear endemic to any filament printer. So the realistic viewpoint should be: I can be busting my butt to get these results on a 600 Euro K8400, or the same results on a 4000 Euro Ultimaker II. And when you get into those prices you can also start thinking about liquid resin printers, which will solve your resolution problems for sure but introduce new issues of their own.
I think you will arrive at the same conclusion I did: the K8400 is the best price/performance product for anyone who BOTH wants to get started 3D printing, AND is a consummate tinkerer of hobby level or better.