Internal Channels are undersized too?

[Quark]

I'm trying to understand a problem I am having. I designed a razor holder to hang off a soap dish I have. All the dimensions came out pretty accurate, except for the channel that goes onto my soap disk. I know that holes comes out undersized, but do U channels as well? I'm designing in Fusion 360, converting to STL using Autodesk Print Studio, then into Slic3r. That back side channel area supposed to be 7mm wide, but its printing close to 6.5mm. Big difference, everything else is spot on though. I've attached a model render for reference. I can attach the step or stl files too. Any tips?

[insta]

(500 microns isn't a huge discrepancy in the XY plane, FYI)

Your software stack is fine, I'd look at slicing settings and perhaps print speed. Force your extrusion width down to 0.35mm for a test print and see how it behaves.

[jsc]

You know you can right click on bodies in Fusion 360 and Save as STL?

Have you run an extrusion calibration? And where are you measuring the channel, perhaps it is a thin lip on the first layer that is too close, is it 6.5 all the way up?

[Quark]

You know you can right click on bodies in Fusion 360 and Save as STL?

Ugh...didn't know that. That just made things a little easier.


Yes, I have done the calibration tests and those are good. I setup the PETG to under extrude a bit, calibration widths set to 0.4, actual is at 0.395.
I'm measuring about 3/4 ways up. I couldn't get an accurate measurement at the bottom opening as I think i stretched that open while doing a test fit. The entire thing is tapered up now.
Now that I've slept on it, I'm wondering if the thin wall is getting pulled together during the bridge operation on that first bridge layer (I print PETG with fan on). I've measured the imprint on my glass surface and it is very close to 7mm, definitely not 6.5mm. I'm going to try and model in a thin center brace in that channel to give lateral support for that thin wall, then cut it out. Thought about running without fan, but I've done that, and it's easier to just run the fan, than having to think if my models have overhangs or bridges.

500 micron may not be much of a discrepancy, but when all other dimensions are within 100 micron, and one is 500 micron, that raises an eyebrow for me.
What does lowering the extrusion width from 0.4 to 0.35 do in terms of dimension accuracy?

[ednisley]

What does lowering the extrusion width from 0.4 to 0.35 do in terms of dimension accuracy?

That tells the slicer to put the threads on 0.35 mm centers, so you get marginally better XY resolution for thin peninsulas and towers. Other than that, thinner threads have no effect, because the slicer insets perimeter threads so their outer surface matches the model's surface.

Now, should you misadjust the Extrusion Multiplier to produce 0.35 mm threads when the slicer thinks they're 0.40 mm, adjacent threads will have 0.05 mm gaps between them: 0.35 mm (actual size) threads on 0.40 mm centers (intended size). Whether that produces an actual air gap depends on other factors, but flat horizontal surfaces will look at least slightly underfilled.

Because the slicer insets vertical surfaces based on the intended thread width, external vertical surfaces will be inset 0.025 mm too far: solid columns will be 0.05 mm too thin, and openings will be 0.05 mm too large.

Remember that you're squeezing molten goo onto a flying platform, so tolerances within ±0.20 mm are about as good as it gets on a long-term basis. Designing your objects with that "manufacturing tolerance" in mind is part of the design process. Requiring a part with an exact fit is possible, but it always drives up the production cost and increases the scrap rate...

[Quark]

So I redesigned the part with a lateral support for the channel, didn't make a difference, still smaller. What is puzzling to me is that only that channel spacing is undersized, but every other external dimensions are well within .1mm. I haven't revisited this again, and just put the parts to use. It isn't perfect like how I would like it to be, but petg is bit flexible, so it just flexes, its functional, and that's what I cared about. Time to move on to other parts.

[Bratag]

So I redesigned the part with a lateral support for the channel, didn't make a difference, still smaller. What is puzzling to me is that only that channel spacing is undersized, but every other external dimensions are well within .1mm. I haven't revisited this again, and just put the parts to use. It isn't perfect like how I would like it to be, but petg is bit flexible, so it just flexes, its functional, and that's what I cared about. Time to move on to other parts.

You might want to play with horizontal compensation in the S3D settings. You may be able to tweak for the tolerance you want.

[Quark]

You might want to play with horizontal compensation in the S3D settings. You may be able to tweak for the tolerance you want.

I'm not using S3d for this, but wouldn't that setting just give me thinner walls altogether? My walls are good at 3mm +- 0.1mm.

[Bratag]

You might want to play with horizontal compensation in the S3D settings. You may be able to tweak for the tolerance you want.

I'm not using S3d for this, but wouldn't that setting just give me thinner walls altogether? My walls are good at 3mm +- 0.1mm.

Not exactly - this is the thread I always go to when I am trying to remember how to use it

https://forum.simplify3d.com/viewtopic.php?f=8&t=2965

[Jules]

Not exactly - this is the thread I always go to when I am trying to remember how to use it

https://forum.simplify3d.com/viewtopic.php?f=8&t=2965

Very good! i hadn't seen that, and i like the sponge analogy!