Puzzling "extra" features appearing after slicing

[bob cousins]

Hi all,

I noticed some odd extra bits appearing after slicing, a small extra sliver of print. I noticed it because of the support generated.

I had this on previous release version, after updating to latest beta I thought it had gone away, but it comes and goes depending on the stl. Naturally I assumed it was faulty STL generated by OpenScad, but looking at the wireframe and checking with NetFabb, the stl looks ok to me.

Here are some screenies showing the problem

Advice on STL debugging or gathering diagnostic info welcome....

Cheers

[frozen...@yahoo.com]

That is an .stl error. Netfab is not particularly great at mesh repair, and has always made more errors than it fixes with any file I've run through it.

Does KS flag/highlight errors in that area of the model before you slice?

[bob cousins]

Not in that area, there are 3 shown in another part.

Is there a better tool for mesh analysis/repair?

[frozen...@yahoo.com]

There seems to be some mesh errors that KS doesn't detect/flag, possibly due to error scale. There must be intersecting or flipped faces in that area, can you post the .stl?

I have had some success with Meshlab, but I mostly manually repair mesh errors in 3DSMax..

[Ante Vukorepa]

On Saturday, May 31, 2014 11:51:19 AM UTC+2, bob cousins wrote:

Not in that area, there are 3 shown in another part.

Is there a better tool for mesh analysis/repair?

Yes, but it's extremely pricey.

Materialise Magics seems to do a pretty nice job of cleaning up meshes Netfabb doesn't.

[PenskeGuy]

On Saturday, May 31, 2014 2:51:19 AM UTC-7, bob cousins wrote:

Is there a better tool for mesh analysis/repair?

Yes. You. I never resort to automated repair attempts. Waste of time and you get more errors, if the repair application actually found any in the first place. Let KS show you where the errors are, go in and fix the problems, regenerate the STL. Repeat until they are all gone.

That said, I have had some models that have these floating bits outside the perimeter, most of which generate support beneath them, and have done the above; scouring the model for anything that might be confusing KS. Happens sometimes in spectacular fashion; perhaps 25% of the model area is showing in these erroneous paths outside the part. There's something in the code that just isn't able to handle some parts, no matter how simple.

[Franz Scherz]

Oh, one example from 2day:

My only problem is that 99% of the models which do not work with Kiss work with Cura, sli3r or CraftWare (other slicers I use if Kiss does not work). It's faster to switch slicer instead editing model (especially if you were not the creator of the STL).

On Monday, July 14, 2014 1:24:34 AM UTC+2, PenskeGuy wrote:

On Saturday, May 31, 2014 2:51:19 AM UTC-7, bob cousins wrote:

Is there a better tool for mesh analysis/repair?

[PenskeGuy]

Might be "faster" but we'll just have to agree to disagree on how best to go about getting good slices that are reliable. I prefer not to drop to the "fast" approach, instead learning why it is failing and prevent it in the future. But, that's me.

[funBart]

I agree with PenskeGuy, I like it that Kisslicer isn't 'interpreting' my model and is pretty absolute with slicing. That way you won't get unexpected changes / artifacts in your slices. I admit it takes some time to learn how you have to output your models from the modeler, but that pays off because you -yourself- have control over the complete process and not some fuzzy algorithm in the slicer.

But Kislicer is developing as well: now it's able to recognize different meshes within the model, and slices them correctly: even when meshes are sharing surfaces (triangles) it's now capable to make distinction between them. As I output architectural models (were for example, walls and floors are sharing the same surface) that's a huge improvement for me.

But that's only when printing own models: for that occasionally Thingiverse print, I have encountered many times Kisslicer is trowing in that many errors, that it's not printable. A couple of that Thingiverse models are that bad, so that even Magics (=$2000 repair program) couldn't repair it.

I simply use another slicer for that: Slic3r has a fairly good stl repair option and does a great job for interpreting single wall's in a model for example (but is slow). Cura has it's own repair options (on Gcode level) and is fast. The new CraftWare has the best (auto)repair with the (bad) stl's I have tested, and is ultra fast (yes faster than Cura and Kisslice), but (although I was very enthusiastic at first) the rigid support structure isn't doing the job well. (it takes a lot of time to manual adjust them). Simplify3D has it's own repair module, but -as far I can tell- it's on the same level as the free version of Netfabb. 

I can imagine that, for users who are mainly interested in printing downloaded models, Kisslicer isn't the right choice. For a more general use Kisslicer could benefit by an option to be more forgiving about bad stl's, as long it stays an option we can turn on or off. 

Or we have to accept that there is no slicer that 'does it all' and we choose the slicer that fit's the job the best for a specific model. On the downside: you have to spent more time to tweak other slicers.

Bart

[PenskeGuy]

On Monday, July 21, 2014 5:06:48 AM UTC-7, funBart wrote:

I can imagine that, for users who are mainly interested in printing downloaded models, Kisslicer isn't the right choice.

THIS is the most cogent, on-target response I've read all day! Can't please all of the people all of the time. Trying to do so is a fool's errand and a bottomless pit that'll suck the life out of you. For me, anything worth doing is worth doing well. If you don't want to be bothered, choose a dumber slicer.

This isn't to be elitist. There are different applications that are for different levels of competency, are for different purposes, have different capabilities, use different approaches and fill different segments of the market. The ones that excel are the ones that remain focused and true to themselves; without getting distracted. If that focus is to be "good enough", that's fine. Nothing wrong with that at all. If that focus is to be precision, it should be allowed to be that and serve the people who appreciate it. Trying to do both is a trap best avoided, as the former will eventually drag down the latter. Let each have its space. We'll all be better for it.

[frozen...@yahoo.com]

Ditto PenskeGuy!

Personally avoiding getting into any of these mesh error issues, again and again........

Wish the '3d printing for everyone' hype would just die off.

Personally, I would NEVER of built a printer if I couldn't model for my own needs.

[PenskeGuy]

On Wednesday, July 23, 2014 5:50:46 AM UTC-7, frozen...@yahoo.com wrote:

Ditto PenskeGuy!
Personally avoiding getting into any of these mesh error issues, again and again........
Wish the '3d printing for everyone' hype would just die off.

Just read a post where MakerBot has "partnered" with Home Depot. MakerBot = Stratasys = 3d Printing Borg with lawyers... or is it 3D Systems? I get so confused with all of the buy-outs.

With our system, which admittedly is a step above the desktop variety, my SO remarks that having 3D printers in libraries, etc., sounds great on the surface but with the amount of fine adjustments that need to be made between material changes (even within a type), cross section, volume & overall intent of the part, your average person off the street wouldn't be able to get an acceptable print the majority of the time; but they make it sound falling off a log easy. And we're only taking about the hardware, here.

[frozen...@yahoo.com]

I saw that a few days ago. Not going to go well I am sure!

Have you noticed that there is almost no mention anywhere that you better not let your home insurance provider know you have a rather unstable fire hazard in your home?