Underextruding Manually

[Neuffexx]

Im trying to print a grid with a .1mm gap in holes, now theoretically according to the slicer with my settings it should be possible to do this:

However the actual result doesnt line up with the slice, because the gap dissapears when printing, I assume it pushes out too much filament causing it to spread over time and close the gaps of the rows.
I figured the easiest way to handle this would be to manually underextrude the material so as to prevent that.

But I dont really know how to do that, never messed with settings like these before, would appreciate it if someone could help me out here and tell me how to set this up, or know of a better or easier way to handle this case.

Tight Net.3mf (44.1 KB)

[blchandl2]

Ypu could set top and bottom shell layers to 0 and the infill to a high percentage (I.e. 85%) to get something like that. It would be trial and error to achieve 0.1mm holes.

[Neuffexx]

I did not consider this option, but I made a parametric tool to create any type of mesh/net to use as dust covers, repair my fly nets, or whatever other need I have for it.
So I want to print at specific sizes and make print profiles for certain ‘size ranges’ rather than a one size fits all print profile (since its clearly not going to work to for something like this case).

[blchandl2]

Achieving 0.1mm features is asking a lot from FDM printing. Have you tried reducing line width settings and maybe reduce the temperature? Using a 0.2mm nozzle may help.

[Neuffexx]

Ive been experimenting the last two days with how to achieve this.
Doing some simple numbers tweaking controlling the flow rate.

Once I reached a certain threshold of underextruding, it seemed feasible to get a grid of that size, since the lines no longer bleed together after printing them.
However, it was not enough, despite the thinner lines it proved difficult to get them to stick (I was using the Textured PEI plate that came default with my P1S). So I started increasing the bed temperature, to try and improve the adhesion.
Going up from the default 55°c that was default for the Bambu PLA Basic profile, to 70 seemed to stick well enough to minimize the loss of inital layer lines.

The biggest problem I then noticed is that certain line’s sides were sticking closer together (bleeding into each other) than other, so I looked at where the toolhead started printing the lines and where it stopped, noticing the pattern that wherever the seams were placed the most bleed happens.
Setting this to nearest improved the results and made it more balanced, but still shows clusters of them bleeding together.

I therefore attempted to do this manually in an alternating patern, however for some reason it just refuses to adopt the drawn seams? Maybe I dont really understand how the tool works, but I assumed that the areas marked would be seam placements…. here two different test cases wher ethe slicer just refuses to alternate between the sides:

• Grouped:

• Individual:

Is there any way to get it to actually follow these paterns for seams?

[Neuffexx]

Also for some reason unknown to me these lines are completely unable to stick to the Super Tack Cool Plate… I tried different heatbed temps just in case pushing it up to 70 as well. Made sure to clean it and let it dry before making another attempt, but maybe there is just simply too little material for that plate to ‘grip’ because of the underextrusion?

Maybe someone has some experience with this they can share.

[Neuffexx]

I also tried setting the whole grid on a raft layer, since that might make it easier getting consistent lines on the bottom side, however, the raft layers do not at all stick to the build plate, but the model’s lines do for some reason.

They get printed at the same speed as well.

Is there a way to make the raft layer print at a different flow rate from the rest of the models layers?

[Bullocks]

You’re really doing this the hard way!

If you want to try to fool the slicer to make your top and bottoms line up like this, you need to understand whats happening with line width and flow:

By increasing line width, you technically increase the distance between extrusions, but the solver ALSO increases the flow to match. So if you want to just simply make it anemic - increase line width AND reduce flow dramatically.
This quick and dirty example is a cube, 0 wall, 0.8 width top and bottom, 50% filament flow, and monotonic line (to make it look nicer but is not important)

If you do some geometry volumetric math you can probably predict the numbers to use to achieve exactly 0.1mm, but this can be tuned extremely quickly with trial and error

[Neuffexx]

I dont fully understand this to be honest.

By increasing line width, you technically increase the distance between extrusions, but the solver ALSO increases the flow to match. So if you want to just simply make it anemic - increase line width AND reduce flow dramatically.

I reduced the line width initally to exactly .4 (default .42) because thats the nozzle size so it wouldnt push out more material (higher flow) to compensate for the extra width.
And later on did the tests in which I showed the images, where I reduce the flow rate down to ~.5, I include the file this time as well with which I did the testing.
Maybe you can demonstrate what you mean with the model included?

Tight Net.3mf (360.8 KB)

[Bullocks]

Sure, here you go

Note the filament profile in this 3mf - it’s flow is set to 0.500 - if you modify this project/change filament to what you use, be sure to retain this setting - it must be low.

Sparse_Net_From_Cube.3mf (28.5 KB)

Stole this infographic from prusa’s forum but the concept is same - if you declare your line width to be really wide, the slicer is still going to want to fill it all with plastic - thats what it does! And it calculates that by the VOLUME of the space.

But… we dont want that filled up. We want gaps on the sides. So if you make that multiplier low, it’ll still position the lines far away, but not extrude enough plastic to fill the area.

If you do that twice, rotating by 90degrees each time [slicer does this by default], you make a net.

This method works reliably because the nozzle is still pushing enough plastic to “fill” the nozzle aperture just like a normal print. It just doesnt know its walls arent connecting =p

[Neuffexx]

Interesting, I think I understand why this works, your basically only pushing out half the required material to make the full line width happen by making the the flow rate .5, thereby making a .4mm thick line instead of .8mm thick which you set for initial / top layer width.
That unfortunately creates .4mm gaps/holes though, but theoretically that can be adjusted to make .1mm sized gaps if you tweak the flow.
I made a test print using those settings (changing the line orientation though to make un-interupted lines.

.8mm Lines | .5 Flow == .4mm grid | .4mm gap

.5mm Lines | .75 Flow (conservative) == .4mm grid | .1mm gap

(Not 100% consistent, might have to get even more conservative with the flow value)

And this does work for the example I gave, but unfortunately, I am unable to make the same thing for my test model. The reason for example why I want to use that model instead of a simple cube, and maybe you know a way to do this via settings too, for example, have the chance to make the grid stronger / without compromising grid density by creating multiple layers per line.
Because I want the grid to be adjustable for whatever needs I might have for it.

As well as wanting to generate grids that are of certain patterns instead of a standard 90° overlap, like triangle, circle, honeycomb, or custom SVG patterns for example.
(Potentially having a base / top layer of one pattern, then an additional alternative pattern sandwiched between them)
Im using this simply now to try and get such small tight-packed lines to print consistently.

The problem im facing is, if I have the line width defined by model, I cant set initial / top surface layers to be a larger line width than the model lines are wide, because the slizer will just leave them empty.
Of course I can compensate this by just removing the gap to make a solid cube and then simply underextrude, effectively doing what you were but with extra steps.

But this falls appart the moment other paterns would be introduced.
Maybe extra planning in the pattern creation can make up for that, would just make it a bit more tedious.