[ADDED] Slicer behavior for thin walled sections

[dkightley]

Both of these deficiencies, errors, crap working, faults.....call them what you like......are features in the software that have been there for more than two years...if not longer....and many users of the software want both of the issues sorted out. I have this as my number one requirement for any software upgrade.

[S3D-Jason]

Just wanted to chime in and let everyone know what we are definitely aware of this request, and it's something our development team has already been working on for quite awhile! So we appreciate everyone's feedback and we look forward to providing more thin wall improvements in the near future!

[dkightley]

And for everyone's benefit I thought I'd share the PM I received from JasonCS....

Hi Doug,

I'm sure you know that we have quite a few people who read the forum every single day, so I saw your message this morning and I wanted to reach out. I know we see posts from you all the time, so we definitely appreciate your enthusiasm and constructive feedback on the forum! We have a lot of people here who are working long hours every single day to try to deliver many of these most requested features as soon as possible. I can assure you, we are just as anxious as you are to make these things available!

So again, we do appreciate your feedback, and I just wanted to make sure you knew we are listening and working hard to make these features available ASAP

Sad to say I remain to be convinced..... I look forward to being proven wrong.

[yonkiman]

How has this not been implemented/fixed yet? I can't think of anything simpler than slicing a wall that is the exact same width as the "Extrusion Width"...

[brian442]

I can't think of anything simpler than slicing a wall that is the exact same width as the "Extrusion Width"...

I would imagine it's actually a very difficult problem to solve. In fact, just from a quick google search there are lots of ongoing thesis projects and research papers about this exact problem. Normally, you can just trace around the outline of your part to create the perimeter, but in this case, you have to somehow determine the "centerline" of the shape, which I'm sure gets complicated very quickly. Complex shapes would have very complex centerlines, so it's not simple at all.

[mroek]

I can't think of anything simpler than slicing a wall that is the exact same width as the "Extrusion Width"...

I would imagine it's actually a very difficult problem to solve. In fact, just from a quick google search there are lots of ongoing thesis projects and research papers about this exact problem. Normally, you can just trace around the outline of your part to create the perimeter, but in this case, you have to somehow determine the "centerline" of the shape, which I'm sure gets complicated very quickly. Complex shapes would have very complex centerlines, so it's not simple at all.

For the problem of one remaining perimeter (where the wall thickness is an odd number of extrusion widths), it should be possible to just run the last extrusion with a fixed offset (of one extrusion width) to either of the last two perimeters.

And also, looking at the current (and seriously annoying!) wiggle fill that is used, which looks like short straight line segments going between the last two perimeters, the centerline that you refer to would pass through the center/mid point of each of these line segments. So if S3D can calculate these fill segments, it should be possible to also find the centerline the same way.

[brian442]

For the problem of one remaining perimeter (where the wall thickness is an odd number of extrusion widths), it should be possible to just run the last extrusion with a fixed offset

That only works if the wall is exactly 3x as wide as your extrusion width. Most real-world models are probably going to vary in size and won't be as simple as this.

wiggle fill... which looks like short straight line segments going between the last two perimeters, the centerline that you refer to would pass through the center/mid point of each of these line segments. So if S3D can calculate these fill segments, it should be possible to also find the centerline the same way.

I can easily think up a lot of situations where that won't work. For example, think of a "T" shape. The center points as you go through the junction won't be valid. So again, that only works for pretty simple shapes, and it would get really wonky for anything more complex than that.

[mroek]

For the problem of one remaining perimeter (where the wall thickness is an odd number of extrusion widths), it should be possible to just run the last extrusion with a fixed offset

That only works if the wall is exactly 3x as wide as your extrusion width. Most real-world models are probably going to vary in size and won't be as simple as this..

Yes, but then again many people (myself included) design stuff specifically for 3D-printing, and I try to use wall thicknesses as a multiple of the extrusion width whenever possible, but as mentioned, only those times where I can fit an even number of extrusion widths will give me exactly the result I want. I'd be extremely much more happy with S3D if it had coped better in this area. Dynamic extrusion width is also something that could possibly be implemented to reduce the need for the annoying wiggle fill.

wiggle fill... which looks like short straight line segments going between the last two perimeters, the centerline that you refer to would pass through the center/mid point of each of these line segments. So if S3D can calculate these fill segments, it should be possible to also find the centerline the same way.

I can easily think up a lot of situations where that won't work. For example, think of a "T" shape. The center points as you go through the junction won't be valid. So again, that only works for pretty simple shapes, and it would get really wonky for anything more complex than that.

Sure, I realize this isn't always easy, but S3D is very good in many areas, so I'm really hoping the developers will come through and give us better functionality for these situations.

[subnoize]

So, my worthless 2 pennies on this topic.

When I trick S3D to print walls the width of the extruder output the quality is surprisingly good. When I do it with other applications (read competitors) the results are less than appealing. S3D seems to have the control down to keep surface quality under control in such extreme conditions.

I understand the concepts behind the slicers and why they treat thin walls the way they do. I understand what I'm asking is difficult to do.

All that said above, the fact I'm actually printing thinner and thinner walled products with less and less infill is directly due to market pressures. People want thin and light products. Whether this is in the medical industry or the hobby industry the need for light parts with thin walls is growing.

The printers can handle it. S3D can do it when tricked. There just needs to be a means to tell S3D to ignore its normal settings about thin walls in certain areas.

I think S3D is in a great position to make this revolutionary step. I hope it happens sooner than later as my little niche of this world would certainly benefit from it.

[Lagbert]

I'd like to see at least 3 options:

Single wall extrusion biased to the outside perimeter
Single wall extrusion biased to the inside perimeter
Single wall extrusion biased to the center line

Here's a picture showing how I imagine biasing to the outside and biasing to the inside would work. The pathing would have to alternate between several different patterns to ensure structural integrity.