Corners a bit rounded

I’m working on a large model that will need to be joined in a few places as too big for print bed (plus I like to keep individual prints to less than 8 hrs).

The joins are showing up a little bit due to the exposed corners being slightly rounded rather than be clean right angles.

This is on Generic PLA, 0.2mm standard.

Have tried project a small square out of the corners a little bit (0.4mm) - then sliced/filed off the excess to make the corners more square. This works quite well - just wondered if anyone had any better ideas for really clean joining of multiple parts.

Example with (left) and without (right) the extra projections attached.


PS/ I guess now I think of it - It I projected the corner extensions at an angle to the joint it might work better - because then I could slice off the excess when the two parts are joined together - rather than before putting them together.

Update - tried that and no better - but quite happy that the little projections allow the corners to be completely square when cut and filed off - just need to make some sort of jig to allow for consistent straight square cutting.

I have not been able to use my bambu lab printer yet but from my past 3D printers you will need to change the jerk settings or acceleration for outer walls this will help with round corners looks like the soft fever fork will let you adjust the jerk setting hope this helps you out with your project.

1 Like

Thanks will give soft fever a try at some point. I guess though that due to the fact that the nozzles are round there is always going to be at least a 0.2mm radius on each corner when using a 0.4mm nozzle.

Think I have settled on the approach of extruding out about 0.5mm at right angles to the cut then across and back. Then shaving the little tab off to give a precise right angle. Examples in the attached photo.

Right hand side is parts as they print with the little extra tabs on the corners.

Front left is parts joined after shaving off the little tabs to get right angled corner (using jig behind).

Left Middle is the original test parts with the rounded corners.

Also decide to bolt through as well as dovetail in order to press the joints as close together as possible.

With a bit of sanding I think I will be able to get the joints almost invisible.

1 Like

Fairly happy with squared corners in real model - but probably need to create a better jig to get really clean cuts.





Left half of last picture is an example of old rounded corners. Reprinting this part to get better corners - plus to improve colour matching - not sure if it was due to a different filament used, or just a smaller layer height.

1 Like

Have been using my overlapping joint technique on my cruise ships - as had trouble getting the overlapped parts consistently cut cleanly.

Finally resolved this problem with the aid of a Dremel type device with a 3mm router bit and a small 3d printed fence part.

Joints not completely invisible - but fairly close.
Left hand side is printed with 0.6mm nozzle - showing the slightly rounded corners.
Right hand side is printed with a little 0.6mm square protruding and overlapping the joint at both sides, which is then shaved off to make a completely square corner.

2nd picture is of Dremel with 3d printed jig - then its just a matter of sliding the parts along and letting the router bit slice off the excess plastic.


1 Like

A few more pictures - will probably publish the Dremel add-on on Printables once I have tried it out on a real print. First photo is updated parts with tabs, compared to a standard part with rounded corners - due to round nozzle.




1 Like

FYI, I believe I hit a slightly different version of the same issue: Dealing with Z-layer failures - #17 by 3dsurfr
Have you tried putting a modifier cube around your corners to slow the speed down to something like 10mm/s?

Thanks - i has assumed the corners were inevitable due to the round nozzle - but i will try some single and multi colour testing with normal and slowed down corners to see if i can improve things without having to resort to post processing.

1 Like

Slowing down to 10mms didn’t make much difference unfortunately - and also I wonder how you stop the modifier cubed sections from having walls all the way around them.

So for critical parts I think I might stick with my print and little bit of extra and shave it off solution.

Printer nice and quiet though when it is going that slow.

10mms print

Right hand side 60, left had 120

Joint between 2 at 10mm/s

1 Like

I had that problem too. Add one cube and got walls around it. Added a second and that one didn’t. Moved the first cube around and the walls went away. (Edit: I fooled myself, can’t eliminate the walls around the modifier) Seems like a slicer bug…

I did some tests on my own with a dovetail. I’m using Bambu PLA basic, 0.6 nozzle, “0.3mm standard” settings, standard Bambu filament profile. The result:

Rounded corners… so you’re not alone…
After some knife action one can see the 3 walls and how they make a round corner:

I then printed a second sample with the following changes:

  • no top layer so one can see the walls clearly
  • set pressure advance (orca slicer) to 0.001 (wasn’t sure whether 0 would be interpreted as zero or “auto”)
  • added a modifier to slow printing of inner&outer walls to 5mm/s (it looks like a separate object attached to the unmodified half, that’s a slicer bug/feature)

The result shows that PA=0.001 results in bulges at the corners (one purpose of PA is to avoid that), see red arrows. Also, there are no bulges where the speed is lowered (blue arrows), which is expected (I read that PA really only matters >20mm/s).

So I don’t know whether this helps anything except confirm the situation. It seems that you may be able to dial in the PA to improve the corners by reducing it from the standard value (dunno what that is). Unfortunately it doesn’t look like you can change PA with a modifier, so you’ll have bulging corners elsewhere in your model.

Update:

I did one more try, which is to add a little point similar to what you do:

I projected out by 0.6mm (0.3mm line width) in both X&Y. The results:

Still some roundness, but I believe this is quite a bit better than initially.

Note that you may want to dial in some elephant foot compensation on the quality tab.

Update 2:

On a hunch I reprinted the previous test with the following changes:

  • outer wall speed 10mm/s
  • pressure advance 0.01
  • elephant foot compensation 0.2mm

I had wanted to try the outer/inner/infill order but forgot, so it’s the default inner/outer/infill…
I still got a bit of elephant foot and did a quick (insufficient) clean-up.

That looks pretty close to me. It’s not a square corner but I think if you do any sanding of the outside wall you end up removing the small bulges and it’ll come together pretty well. I’m sure it can be tuned a bit more. I can push the 3mf & step files to printables if you care.

Thanks for the suggestions, you are right about the elephants foot - got a bit lazy last night as it was getting late - think I will try a few more tests with some of your settings.

Also I think that sanding rather than shaving is worth a try - will see how a small projection in line with the joint rather than towards it works - then try sanding the surface.

Or maybe one side projecting out in some sort of circular path that precisely hugs the curve on the other side - then just sand off or iron out the excess.

Also watched an interesting new video yesterday from slack3d about joining parts - his examples all seemed to have rounded corners too - but there were some interesting alternatives to dovetails - such as a sprung T joint that could hopefully pull the parts more tightly together without having to resort to bolts.

2 Likes

Tests from today

Where
1 - 0.6mm rounded corners on both size - due to nozzle diameter

Rest are a. - As printed, b. - After post processing

  1. Project 0.3 out for 1.2mm towards joint - then shave or sand off excess - on face not facing outside

  2. Project 0.3 out for 1.2mm at right angles from the joint - into the visible part - then sand/route off

  3. One side leave 0.6mm rounded corner, on other side project out to hug the other sides rounded corner.

Then shave off the projected out bit to give a flat surface.

I think I will probably stick with option 2 - due to lack of sanding required on exposed surface

However I think 4a actually looks quite nice without the sanding - so might consider that as a ‘no work required’ option.

Also quite lack the Slack3d suggested springy T joints - might switch to those instead of Dovetails for future projects.

PS/ Final picture from top to bottom, 4b, 3b, 2b, 1, 4a.

I think for existing prints I might go for 4b - as the additional part is only required on one side I would only need to reprint half of the parts. Also the sanding might possibly look better on white than green.
NB./ On white print shown I know the windows don’t quite line up - it is just some old test prints.




3 Likes