[3.4.0beta7] new input shaper disturb pressure advance

Argo

Hello,

I noticed that with input shaping enabled, Pressure Advance values need to be much higher and also don't produce as nice edges / corners on the print object as with the old shaper (DAA) or with the most new basic form ZVD.

For reference: PETG needed 0.08 pressure advance (with DAA). With ZVD it needs 0.12 and with el2 it needs 0.2 pressure advance.

Even with the new higher value the corners are not perfect. I printed a test cube and one corner has too much PA and the next corner not enough.

Can somebody else confirm this as I have not noticed any threads reporting the issue?!

Phaedrux

yes it has been said that pressure advance will likely need to be retuned after setting an input shaper. You may also need to adjust jerk and acceleration. Input shaping is also currently experimental, so things like this are to be expected. So thanks for reporting your findings. Could you perhaps show a photo example of your different findings and values? And some more details about your printer?

Argo

@phaedrux

Oh, sorry, I forgot.

Voron 2.4 with LGX extruder and Mosquito Hotend
Duet 3 Mini5+
Speed: 120 mm/s
Acceleration: 3.000 mm/s

And I know that it's still beta. I'm not complaining, just reporting

PLA, EI2 shaper and PA at 0.15:

Bottom right is almost ok but upper right corner overshoots. Normal PA values for PLA are 0.05 up to 0.7 for this setup.

PCR

@argo yes. PA has to be tuned again. And it depends which shaper you are using. I already opened a thread on it.

For me it's

Flow-IS-PA-Retract

PCR

@argo and yes my PA is now higher than before.

PCR

@argo https://forum.duet3d.com/topic/25831/input-shaping-and-pa Thats my thread. I done more testing too.

Argo

@pcr

Oh I did not look in General Discussion as this is a beta feature.

Were you able to produce the same results with a higher PA value?
If I increase my PA value too much with input shaping enabled some corners begin to starve.

I wonder if that's a wanted behaviour. Besides the Voron we also have a bed slinger printer with Klipper. PA almost did not change with input shaping enabled.

PCR

@argo it tuned my PA with this https://forum.duet3d.com/topic/25764/pressure-advance-tuning-with-conditional-g-code/

so first IS then this.

are you using arcwelder aswell? I had problems with PA and arcwelder before

Argo

Another discovery I made:

In order to get good sharp corners I have to bump up PA quite a bit. From 0.07 up to 0.18.

But then the infill has issues and begins to starve which is a expected behaviour from such a high PA value for PLA.

Edit:
Did another test with ZVD shaper: pressure advance is at 0.07 - 0.075 for PLA and corners look nike + infill is not starving.
I guess the shapers with longer shaper duration somehow mess with pressure advance right now?

Phaedrux

@argo said in [3.4.0beta6] new input shaper disturb pressure advance:

But then the infill has issues and begins to starve which is a expected behaviour from such a high PA value for PLA.

This is not uncommon with PA and it's recommended to compensate by increasing the infill overlap in the slicer.

Argo

@phaedrux

This would work if the effect is the same on all infill paths.
The last picture I posted has huge gap on one specific infill part. Other parts look fine though.
I don't really see a way to compensate this with the slicer without causing problems on other parts.

Argo

RRF 3.4b7:

I first calibrated my PA.

Then some real world testing with PA set to 0.064 and

To get rid of the bulging I would need to increase to very high PA values which does impact all other areas whilst printing like starving infill ends etc. (see pictures above).

Used input shaper: M593 P"zvdd" F54.32

With RRF 3.3 I had similar PA values but no overshooting corners.