This is an interesting take indeed. External perimeter acceleration should indeed work nicely for the smaller travel moves.

I do wonder what test results did Prusa have to indicate that a slower acceleration for the external perimeter travel was beneficial vs an approach like this which is much simpler (and consistent).

They most probably miss your eagle eyes that saw the regression on the seams on curved surfaces.

@RF47
I am eagerly waiting for you to test this too.

It seems to work well, correctly setting the acceleration for print and travel.

I am curious to see how it compares to Prusa Slicer:


I don't know if something is configured incorrectly, but it seems that short travel acceleration is not being used in prusa slicer (external perimeter acceleration is used, as in this implementation).

in this prusa comit:
prusa3d/PrusaSlicer@52f5219#diff-cb3d09d2073b6e9032c65d07aa8f10d166324a53136b2aa1288f7b76f7d94319

i found this comment:

    if (!GCodeWriter::supports_separate_travel_acceleration(config().gcode_flavor)) {
        // In case that this flavor does not support separate print and travel acceleration,
        // reset acceleration to default.
        // TODO: This doesn't seem to perform what the comment describes.
        gcode += this->m_writer.set_travel_acceleration(travel_acceleration);
    }

maybe that's my case.

When this will be merged?

What happens if the travel acceleration is less than that of the outer wall?
Furthermore, could it be useful to further reduce the displacement acceleration when movements occur inside the model and are not required to cross an external perimeter? In a test I ran in a cube, drastically reducing the travel acceleration completely suppressed the vibrations at the start of the perimeter.

What is a displacement acceleration?

What happens if the displacement acceleration is less than that of the outer wall? Furthermore, could it be useful to further reduce the displacement acceleration when movements occur inside the model and are not required to cross an external perimeter? In a test I ran in a cube, drastically reducing the displacement acceleration completely suppressed the vibrations at the start of the perimeter.

Error in translation. I corrected the previous message

What happens if the travel acceleration is less than that of the outer wall?

Highly improbable, why would you have travel acceleration lower than any print acceleration?

Furthermore, could it be useful to further reduce the displacement acceleration when movements occur inside the model and are not required to cross an external perimeter? In a test I ran in a cube, drastically reducing the travel acceleration completely suppressed the vibrations at the start of the perimeter.

Reduced acceleration before the outer wall affects the seam quality as already proven (more evident for seams on a round wall).

The whole idea is that you already have an acceleration set that will be used for the outer wall. If that is good enough for the outer wall (that might have tiny segments also) it must be just as good for the tiny travel towards the outer wall.

• If the "Wipe Before External Loop" option is enabled, we have two sections of travel near the beginning of the external loop. Both paths affected by decreased acceleration. Fine!

• In contrast, "Wall printing order" can be set to "outer/ inner" and we can have more than one object with the "Wipe Before External Loop" option enabled or disabled. This pach doing nothing in this case and question is what to do.

• The same with "Wall printing order" set to "outer/ inner", "Print Infill First" enabled and "Wipe Before External Loop" enabled or disabled.

• And the funniest thing is that without this path, the "Wipe before external loop" option will always add a short move before the external perimeter and reduce the acceleration to the external wall acceleration before this move occurs.

Technically, we already have a better setting that in any case will add a short movement in front of the outer perimeter and it will happen with the acceleration of the outer perimeter.

Could we please get this merged? I am experiencing significant quality issues in OrcaSlicer compared to PrusaSlicer on models with small features/islands

The difference is drastic. In OrcaSlicer, I am getting severe artifacts (visible and inconsistent layer lines) and blobs on the external perimeters (specifically on the antlers in the example below), whereas PrusaSlicer handles them perfectly. It looks to me that the current acceleration handling for short travels is causing these defects.

PrusaSlicer:

OrcaSlicer:

The issues on the print (sliced by Orca, doesn't happen on PrusaSlicer)

Could we please get this merged? I am experiencing significant quality issues in OrcaSlicer compared to PrusaSlicer on models with small features/islands

The difference is drastic. In OrcaSlicer, I am getting severe artifacts (visible and inconsistent layer lines) and blobs on the external perimeters (specifically on the antlers in the example below), whereas PrusaSlicer handles them perfectly. It looks to me that the current acceleration handling for short travels is causing these defects.

PrusaSlicer: OrcaSlicer:

The issues on the print (sliced by Orca, doesn't happen on PrusaSlicer)

Hi @PawiX25
Thank you for showing picture.
Is the improvement in your case due to the Travel short distance acceleration option in PS?
Did you enable that option in PS?

Yes, I have Travel short distance acceleration enabled and set to 250 mm/s² in PrusaSlicer.

To rule out any hardware issues, I went through extensive troubleshooting with Prusa Support. I replaced belts, retightened everything, and completely reassembled the pulleys. Prusa Support then sent me a G-code file sliced in PrusaSlicer to test, and it printed absolutely flawlessly.

Currently, everything I slice in PrusaSlicer prints perfectly with no defects, whereas OrcaSlicer still produces these artifacts on the exact same hardware.

I’ve attached photos below showing the issues in Orca

Is this with this PR or without?

These photos show the issues I have in Orca without this PR

Can you please share photos with the PR to show the "before/after" difference?

Sure, I am compiling the build with this PR right now. I'll print the model i was trying to print before as soon as it finishes and post the before/after photos.

Also please ensure cooling and movement parameters are identical, including slicing resolution. Just a double check as some of the above artefacts are not appearing as ringing related but have similarities to slicing tolerance and cooling artefacts I’ve encountered before :)