Single Wall Hex. Cells - How Small Can I Go?

[DeftCAD]

See image below, and yes dimensions are mm!

This is looking from top down. I've only drawn up a few for testing, but what I actually want to do is print a pattern of these 'cells' to form a honeycomb block of about 50mm². Z height overall would ideally be 200mm. I would like the walls to be as thin as possible and ideally the circle in the centre to be about 0.75mm, but I can live with more. I've modelled what you see in the image, but Ideamaker says it's too small: it will process it, but in the preview there are elements missing, as below:

So I suppose my question is, how far can I push this? What is possible and how can I set it up?

Oh, and once I can do it in PLA, I'd like to make one in Tungsten filled material...

[Jetguy]

I'm just going to be dead honest here with you, it's about using the wrong tool for the job.
The workflow for most 3D printing is, you have a CAD shape, you export an STL, that STL is then interpreted by the slicing software and the slicing software creates the toolpath gcode.

When you get into something this precise, you are working right on the edge of the very rules and limits of the tool chain of software.
In other words, think of the STL as a color between the lines outline drawing. Then you are hoping that the slicing engine can produce the path you desire within these right on the limits color between the lines. Also keep in mind in an STL how a wall with thickness is represented. It's actually 2 logical walls, one faces inside, one faces outside, the middle space is infill but remember the rule that makes it infill, the area is manifold and water tight. The last rule is that the extrusion width has to fit down this infill wall. That's why you are having issues, variations on these 3 problems. Example is, the middle is missing from your gcode preview. If I was to guess why, it's one of 2 rules being broken. You have overlapping geometry in the cad design and when you created the STL, you violate the manifold and watertight rule so the slicer cannot determine what is infill area in this overlap (flipped walls, hidden overlapping geometry, and others) or the curves and such break the width rule.
If you had posted or linked the STL we could check it for rule violations https://www.shapeways.com/forum/t/overl ... try.29105/

When you get to this kind of engineering, you may need to go more direct to machine path creation and not be limited by the rules.
Again, with such a simple vertical shape, this is just going to be a repeating tool path on a layer and then Z height changes, rinse, repeat.
So for ultra precision, sometimes hand writing the toolpath gcode, or a CAD program with CAM function is more direct than the STL export, import the STL into a slicer, and then create the toolpath.

[Jetguy]

Here is another guide on STL and "the rules" http://www.meccanismocomplesso.org/en/c ... -printing/

[Jetguy]

Here is a discussion on STL model repair functionality in Ideamaker, but in this case with these tolerances, I don't know if auto repair is going to fix this STL and give you a good toolpath (gcode preview).
viewtopic.php?t=3290

[Jetguy]

In fact, this post right here is the one to see about overlapping STL geometry viewtopic.php?t=3290#p18605

[Jetguy]

Let me just recap this because it's a good discussion to have with the group at large.
In general, 3D printing and the slicing software was an attempt to make it easy for the average person.
STL file format does have some limitations in how it represents the original CAD object. It is meant to be a small file size and contains basically the bare minimum geometry to describe the shape. As we get to more complex and precise shape that begin to push the very limits of extrusion width, feature size, and single wall, the STL file format is VERY limiting and unforgiving of mistakes and errors. Couple that with the fact that 3D printing slicing is meant to be fast and easy and is not really honestly designed down to the same precise toolpath for single wall objects that push the very limits of minimum feature size and extrusion width.

When you get to a certain point, you are literally just wanting to control the very path of the nozzle. At that point, STL, interpretation, and slicing software are likely to expose these limitations inherent in the entire STL to gcode function.

[Jetguy]

Again, your CAD drawing showing clear potential for overlapping geometry if not properly Boolean joined and internal hidden geometry removed before exporting the STL.

Again, you did not post the STL but a good guess is this is what is going on when looked at from a side view.

[Squenz]

while i don't concur with Jetguy's postings, i'd try it this way....

You need turn down shells to 1, no infill, disable "check thin wall" and enable "fill Gaps in shells". This leads to a 0.4mm thick single wall.

Then you set Extrusion width to 0.5mm and the Extrusion Multiplier in the Filament settings to somewhere around 10 to 15 percent more than usual value, test print and play around with this value until you get exactly 0.5mm shell width. 200 mm hight .... if you print a 50 sq.mm part in 200 mm height, it seems to be feasible in my opinion.

Of course, the outer edges of the shells won't be that sharp, but it think you could live with that.

[DeftCAD]

In terms of 'wrong tool' I understand what you're saying. I know I'm pushing it, but I regularly work in this kind zone; with folks that are pushing the envelope of what can be done and as such I get some very odd but interesting engineering problems to solve. If I can make this work I will have some very happy customers, a few respectable sales, and a successful product. It's all good news.

Thank you for your feedback. I thought the overlapping areas would be ignored when the stl was created... Every day is a school day! I've remodelled:

... and in preview:

So, that's one step closer. Will it print? Oh YEAH! Here it is resting on the end of my mechanical pencil!

I printed it in green PET-G just because that's what I had in the machine. Looks pretty good. I can see clear through two of the tubes easily, the third would probably clear out if I pushed a piece of wire through it. This is really promising. I'm highly impressed that the printer was able to do this and it was even stuck quite firmly to the bed!

[Jetguy]

Again, by "wrong tool", I meant that in some cases, a free slicer may not auto generate the optimal gcode path you desire for a given model. The printer will attempt to do whatever you feed it for a print file and is quite a capable tool, but is near the last end of the total tool chain to go from CAD to actual object.
More importantly, what I was not saying or implying and could have been misunderstood, that the Rais3D printer might be the wrong tool or incapable. We might be pushing limits here in this example, but it's still inside the envelope of possible.

To the original problem here, it was overlapping geometry in the STL.

Absolutely, number of shells and other slicing parameters are required to produce proper automatic gcode path creation in complex and tolerance pushing thin wall objects. But even then a non-conforming STL file with overlapping geometry is highly likely to cause errors in slicing.

[Squenz]

while i don't concur with Jetguy's postings, i'd try it this way....

You need turn down shells to 1, no infill, disable "check thin wall" and enable "fill Gaps in shells". This leads to a 0.4mm thick single wall.

Then you set Extrusion width to 0.5mm xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx, test print and play around with this value until you get exactly 0.5mm shell width. 200 mm hight .... if you print a 50 sq.mm part in 200 mm height, it seems to be feasible in my opinion.

Of course, the outer edges of the shells won't be that sharp, but it think you could live with that.

I have to correct my self here. If extrusion width is changed, Ideamaker corrects automatically the extruded amount of filament as well as the tool path of the nozzle.

Which means, if the printer is properly dialed-in(E-Steps) and the Flowrate is also accuratly tuned, changes in extrusion width will be adapted by Ideamaker without further need of user interference.