Actual Infill Density vs Slicer Percentage

Discussions about ideaMaker and other printing software.
User avatar
Julia Truchsess
Posts: 208
Joined: Mon Mar 21, 2016 1:30 pm
Location: Connecticut
Contact:

Actual Infill Density vs Slicer Percentage

Postby Julia Truchsess » Sat May 07, 2016 9:36 pm

I've noticed that sometimes a small change in the percentage fill value used can make a seemingly huge difference in the subjective fill density. I haven't done the actual math to know whether the slicer is operating correctly and this is a non-intuitive subjective thing, or whether it's a slicer bug, or even an intentional "feature". Anyway, in the two attached screen shots we see 20% fill and 25% fill. I would really not expect to see such a huge jump in apparent density between these two values, especially as going from 15% to 20% makes very little difference. It seems there are discontinuities in the mapping (or it's highly nonlinear), and this makes it difficult to estimate the percentage setting one wants, necessitating a lot of trial and error: slice, preview, cancel, edit profile, slice, preview, cancel, edit profile, again and again.

20 pct fill.JPG


25 pct fill.JPG

User avatar
dhylands
Posts: 119
Joined: Mon Mar 21, 2016 9:11 pm

Re: Actual Infill Density vs Slicer Percentage

Postby dhylands » Sat May 07, 2016 9:40 pm

I've observed similar in Cura.

If you look closely (best is using a gcode preview), infill which is 24% or less has the infill running in both directions on all layers.
Infill which is > 24% has the infill running in a single direction on each layer, which is why the holes get a whole lot closer together.

User avatar
Julia Truchsess
Posts: 208
Joined: Mon Mar 21, 2016 1:30 pm
Location: Connecticut
Contact:

Re: Actual Infill Density vs Slicer Percentage

Postby Julia Truchsess » Sat May 07, 2016 9:48 pm

Thanks, Dave - maybe there's a coding-related reason for it, but it seems like a bad thing from a UI perspective. Honestly, sometimes I wish I could just specify infill (and support) spacing in absolute terms, like "4mm" instead of percentage. Infill and support spacing needs in many cases are dictated by bridging capability, and this is more of a fixed distance than a percentage - you can't bridge twice the normal distance just because your print doubled in size. In terms of load carrying capacity, maybe percentage makes some sense, since it's a weight-per-area thing. It'd be nice to be able to go either way.

EL Cuajinais
Posts: 177
Joined: Wed Apr 06, 2016 8:48 pm

Re: Actual Infill Density vs Slicer Percentage

Postby EL Cuajinais » Sun May 08, 2016 3:22 am

I've also observed this in S3D but never mentioned it. It was very apparent between two different printers as well. One was reprap, the other was a delta.

User avatar
walshlg
Posts: 1339
Joined: Wed Mar 23, 2016 12:01 pm

Re: Actual Infill Density vs Slicer Percentage

Postby walshlg » Sun May 08, 2016 6:23 pm

Julia Truchsess wrote:Thanks, Dave - maybe there's a coding-related reason for it, but it seems like a bad thing from a UI perspective. Honestly, sometimes I wish I could just specify infill (and support) spacing in absolute terms, like "4mm" instead of percentage. Infill and support spacing needs in many cases are dictated by bridging capability, and this is more of a fixed distance than a percentage - you can't bridge twice the normal distance just because your print doubled in size. In terms of load carrying capacity, maybe percentage makes some sense, since it's a weight-per-area thing. It'd be nice to be able to go either way.


Great idea, please suggest this on suggestions page

User avatar
Vicky@Raise3D
Posts: 2425
Joined: Fri Mar 25, 2016 3:54 am

Re: Actual Infill Density vs Slicer Percentage

Postby Vicky@Raise3D » Mon May 09, 2016 10:10 am

Thanks for you suggestion. We will take it into consider.

User avatar
Vicky@Raise3D
Posts: 2425
Joined: Fri Mar 25, 2016 3:54 am

Re: Actual Infill Density vs Slicer Percentage

Postby Vicky@Raise3D » Mon May 09, 2016 10:43 am

Julia Truchsess wrote:I've noticed that sometimes a small change in the percentage fill value used can make a seemingly huge difference in the subjective fill density. I haven't done the actual math to know whether the slicer is operating correctly and this is a non-intuitive subjective thing, or whether it's a slicer bug, or even an intentional "feature". Anyway, in the two attached screen shots we see 20% fill and 25% fill. I would really not expect to see such a huge jump in apparent density between these two values, especially as going from 15% to 20% makes very little difference. It seems there are discontinuities in the mapping (or it's highly nonlinear), and this makes it difficult to estimate the percentage setting one wants, necessitating a lot of trial and error: slice, preview, cancel, edit profile, slice, preview, cancel, edit profile, again and again.

20160509183014 (2).png


20160509183014 (2).png


Hi Julia,

Please check the attached pictures. Just enable the "Only Current Layer" you will find difference
20160509183014 (1).png

20160509183014 (2).png

The infill is 25%.

We change the infill direction from two directions crossing in one layer to print all the lines in the same direction in one layer from 25%. Because the crossing point will be bulge when two lines cross in one layer. And the higher infill density, the more bulge points. We set 25% as a watershed.

20160509184107.png

The infill is 20%.

Zettlinger
Posts: 412
Joined: Mon Mar 21, 2016 11:48 am
Location: Roosendaal, Netherlands

Re: Actual Infill Density vs Slicer Percentage

Postby Zettlinger » Mon May 09, 2016 11:43 am

Thank you for the explanation Vicky :)

User avatar
Julia Truchsess
Posts: 208
Joined: Mon Mar 21, 2016 1:30 pm
Location: Connecticut
Contact:

Re: Actual Infill Density vs Slicer Percentage

Postby Julia Truchsess » Wed May 11, 2016 4:04 pm

I can see that the infill algorithm changes from crosshatch-on-each-layer to spaced-lines-rotated-90º-each-layer, and I understand why this is done, but I still don't understand the rationale for the sudden change in line spacing. Ignoring the layered nature of the structure and regarding the infill as solid vertical walls, the X-Y spacing between the walls, i.e. the infill density, i.e. the parameter that's being controlled, makes a big discontinuous jump at the point that the algorithm changes. The algorithm's workings are its own business, and rightly made invisible to the user, but the target parameter, infill density/spacing, should change smoothly as the control value is changed by the user.

20 pct lines.JPG


25 pct lines.JPG

User avatar
dhylands
Posts: 119
Joined: Mon Mar 21, 2016 9:11 pm

Re: Actual Infill Density vs Slicer Percentage

Postby dhylands » Wed May 11, 2016 11:16 pm

Let's pick some arbitrary numbers to make the math easy. Take a 10 x 10 grid (say a piece of graph paper) and assume that the filament is one square wide.

If we do a cross hatch, then there will be a 9x9 area with no filament and a 1x10 and a 1x9 area filled with filament. So 19 squares out of 100 or 19% infill.

Now if you run the infill in only a single direction, then only 10 squares are filled in, which is 10% infill. If you put a second line down the middle (so parallel to the first), then you'd have 20 squares filled in (or 20% infill) but the distance from line to line would be 5 squares instead of 10 squares.

User avatar
Julia Truchsess
Posts: 208
Joined: Mon Mar 21, 2016 1:30 pm
Location: Connecticut
Contact:

Re: Actual Infill Density vs Slicer Percentage

Postby Julia Truchsess » Wed May 11, 2016 11:49 pm

I see, you're looking at it in 3D and assuming that there's air between successive layers of "alternating" single-direction infill, and that this gap nullifies halving the X-Y spacing. I was seeing it in 2D only and assuming that the layers would bond into vertical "walls" even though there's a "skipped" layer between them. I'm gonna cut open some parts and have a look :)

User avatar
Vicky@Raise3D
Posts: 2425
Joined: Fri Mar 25, 2016 3:54 am

Re: Actual Infill Density vs Slicer Percentage

Postby Vicky@Raise3D » Thu May 12, 2016 3:35 am

As Dave explained, we didn't increase the fill density. Just change the shape. For example, before there are two lines crossing making a 1mm square. After there are still two lines in the same direction dividing the 1mm to two 0.5mm space. Still same quantity of lines in one layer. But looks more concentrated.

User avatar
Julia Truchsess
Posts: 208
Joined: Mon Mar 21, 2016 1:30 pm
Location: Connecticut
Contact:

Re: Actual Infill Density vs Slicer Percentage

Postby Julia Truchsess » Sat May 14, 2016 9:28 pm

Well I'll be darned, there are indeed air spaces in there. I had a very wrong mental picture of the infill structure - I thought the every-other-layer infill would somehow droop or be intentionally overextruded to form a solid vertical wall rather than a 3-D lattice. Thanks Dave and Vicky, for enlightening me :)

I still feel that the UI for this could be improved, as in many cases I'm more interested in the X-Y spacing for supporting horizontal shell surfaces than I am in the 3-dimensional density of the fill.

DSC_1159.JPG

Hescockrl
Posts: 26
Joined: Thu Mar 31, 2016 4:43 am

Re: Actual Infill Density vs Slicer Percentage

Postby Hescockrl » Sun May 15, 2016 1:17 am

I think you are all going a bit far with this. Think of it this way, 25% is 160% of 15%, so the difference between the two is quite large. Going from say 50% to 75% will be less noticible to the eye because 50% already looks darn near filled in. I'd chalk it up to a visual oddity kind of like when you try to line up two similar curved pieces of toy train track one above the other and the bottom one looks longer.

User avatar
Julia Truchsess
Posts: 208
Joined: Mon Mar 21, 2016 1:30 pm
Location: Connecticut
Contact:

Re: Actual Infill Density vs Slicer Percentage

Postby Julia Truchsess » Sun May 15, 2016 1:40 am

The discussion was about the difference between 20% and 25%, not 15 and 25%. Visually/subjectively from a "top-down" view it's a huge discontinuity, but now I understand that it's not discontinuous in terms of volumetric fill.

Hescockrl
Posts: 26
Joined: Thu Mar 31, 2016 4:43 am

Re: Actual Infill Density vs Slicer Percentage

Postby Hescockrl » Sun May 15, 2016 1:50 am

It's the same principle, the smaller the infill % the bigger visual difference you will see with small changes. Larger infill % will produce smaller visual differences with the same amount of change. Kind of like nobody notices 500 million dollars added to the debt now that its in the trillions, but 30 years ago 500 million would have been a big deal.

User avatar
Julia Truchsess
Posts: 208
Joined: Mon Mar 21, 2016 1:30 pm
Location: Connecticut
Contact:

Re: Actual Infill Density vs Slicer Percentage

Postby Julia Truchsess » Sun May 15, 2016 2:10 am

I understand the concept of relative change you're describing, however, that's not what's happening here. If you read the thread you'll understand that there's a fixed threshold in the slicer logic set at 25%, where the infill algorithm changes radically, and the spacing of the lines in X-Y jumps to 50% of what it just was to compensate for the fact that the spacing in Z doubles at that threshold. It's not an optical illusion. If the X-Y spacing was 6mm at 24.999999999% infill, it would jump to 3mm at 25%. Now that I understand that the Z spacing has doubled, the rationale for the discontinuity in X-Y spacing is clear to me.

User avatar
dhylands
Posts: 119
Joined: Mon Mar 21, 2016 9:11 pm

Re: Actual Infill Density vs Slicer Percentage

Postby dhylands » Sun May 15, 2016 2:31 am

Actually just look at the difference between 24% and 25%

Hescockrl
Posts: 26
Joined: Thu Mar 31, 2016 4:43 am

Re: Actual Infill Density vs Slicer Percentage

Postby Hescockrl » Sun May 15, 2016 4:06 pm

Well then, I definitely see what you are saying there now.
2016-05-15.png
2016-05-15 (1).png

ChrisDagher
Posts: 1
Joined: Wed Sep 13, 2017 6:05 am

Re: Actual Infill Density vs Slicer Percentage

Postby ChrisDagher » Wed Sep 13, 2017 6:17 am

I recently put together an equation to determine the line spacing for 90-degree alternating lines.
Infill To Line Spacing Analysis.png

I basically printed cubes varying from 0 to 100% infill in increments of 10% and measured the spacing of the lines center to center, parallel to the extrusions. For smaller spacing, I used magnification to help. This seems pretty accurate when I model it in my CAD program. Hope this helps in some way. Pretty happy with the results, considering I did this late at night in my cramped dorm room.
Attachments
20170913_001358.jpg


Return to “Software”

Who is online

Users browsing this forum: No registered users and 4 guests