IN SEARCH OF: Meaningful Linear Advance Calibration  

Re: IN SEARCH OF: Meaningful Linear Advance Calibration

I may have figure out part of what's going on. Hollow prints always print external perimeters at slow speed, regardless of speed or linear advance settings:

One primary reason hollow prints print slow is due to this setting:

Re: IN SEARCH OF: Meaningful Linear Advance Calibration

Ah, good point! I am well aware of that setting, just didn't consider it. Checking this out, it's a bit interesting. I set min print times to 500:

But print speeds remained low. I'd expected the min print speed to only kick in if times were exceeded:

Upped the min print speed:

And now it prints fast (still respecting max speeds and max volumetric speeds):

I'm learning more about SLic3rPE at the moment than LA.

Re-generating my test tower, but have to deal with the dentist for a bit...

Re: IN SEARCH OF: Meaningful Linear Advance Calibration

Active English: SLOW DOWN IF TIME IS BELOW [xxxx]

I set mine to [ 1] and it worked well and kept printing at my higher speeds.

Re: IN SEARCH OF: Meaningful Linear Advance Calibration

Active English: SLOW DOWN IF TIME IS BELOW [xxxx]

I set mine to [ 1] and it worked well and kept printing at my higher speeds.

I knew that. Doh. OK I blame the dentist!

I am getting some observable results now. Will update once my brain isn't numb.

Edit: Success! I'm still tweaking sizing and wall thickness, trying to keep the print below 30 minutes, but this is encouraging:

You can see the bulging corners gradually decreasing from left to right as LA is applied in increments of 5.

I'm going to try to expand this idea to jerk & acceleration tuning. I want to try it for speed (M220) and extrusion rate (M221) tuning as well. I want to find optimal values for a variety of materials and nozzle sizes.

Thanks for the feedback guys! It's been very helpful. @tim.m30, I'm going to stick with novocaine poisoning as my excuse for the brain cramps.

Re: IN SEARCH OF: Meaningful Linear Advance Calibration

Perfect! You cracked that problem.

Re: IN SEARCH OF: Meaningful Linear Advance Calibration

Just a final update. I've found a good combination of test piece parameters to show the impact of Linear Advance quite clearly.

You can see the "dogbone" effect in corners with LA set to 0 at left, transitioning to rounding off as LA moves higher to the right:

The impact on the z-seam is also quite apparent, moving from puckered with LA set to 0 at left, transitioning to split as LA moves higher to the right:

I've attached configs and samples for a smaller PLA test tower for LA values 15 to 45 if anybody's interested.

Re: IN SEARCH OF: Meaningful Linear Advance Calibration

I basically never print PLA. Here is how I attempted to change it for PETG.

Opene your amf file
Importing the ini file
Change the filament to PETG

G-Code looks right. Is infill supposed to be 15% gyroid?

Re: IN SEARCH OF: Meaningful Linear Advance Calibration

[...] Change the filament to PETGYes and be sure to set that cooling minimum layer time to "1" or some suitably low value. I now make a point of checking speed in preview before saving the gcode out. Wasted a lot of time using lower speeds. I find that external perimeter speeds below 60mm/s don't really show much effect.

G-Code looks right. Is infill supposed to be 15% gyroid?
Yes, but there really shouldn't be any infill. I've created a smaller PETG tower, but no test prints yet. See attached.

I've generated towers testing the impact of acceleration, jerk & speed on otherwise-identical prints. Nothing earth-shattering, but it is interesting see each parameter's impact on print quality in isolation.

Re: IN SEARCH OF: Meaningful Linear Advance Calibration

I may have figure out part of what's going on. Hollow prints always print external perimeters at slow speed, regardless of speed or linear advance settings:

One primary reason hollow prints print slow is due to this setting:

Thanks for the hint!

I requested a feature some months ago, which would help to resolve these situations: Show limiting factor for printing speed
https://github.com/prusa3d/Slic3r/issues/1490

Unfortunately, this request didn't get much attention.

Re: IN SEARCH OF: Meaningful Linear Advance Calibration

I requested a feature some months ago, which would help to resolve these situations: Show limiting factor for printing speed
https://github.com/prusa3d/Slic3r/issues/1490
Unfortunately, this request didn't get much attention.

A bit off topic, but a drill down to find the limits would be a nice secondary feature; but what I want is something that explains - in plain English - a simple gcode command like G1 E0.80000 F2100.00000 : G1 F982 : G1 X121.959 Y112.292 F10800.000 : G1 F8640 ...

I want a command to set acceleration over a distance, not some random percentage of some other trait like base speed or max speed minus base speed after applying some other slow down multiplied by a tic-tok timer of third relevance in the command queue.

venting complete

Re: IN SEARCH OF: Meaningful Linear Advance Calibration

Why not request it on github and start a discussion about it in a separate topic here in the forum?

RE: IN SEARCH OF: Meaningful Linear Advance Calibration

Looks like the project file upload was lost. I've attached it to this post.

RE: IN SEARCH OF: Meaningful Linear Advance Calibration

In case you're having trouble finding the LA Test Tower, I'm attaching it again here.

RE: IN SEARCH OF: Meaningful Linear Advance Calibration

Nice work. Will this be updated when 1.5 is officially supported? Or do you already have an LA1.5 version?

RE: IN SEARCH OF: Meaningful Linear Advance Calibration

Posted by: @vintagepc

Nice work. Will this be updated when 1.5 is officially supported? Or do you already have an LA1.5 version?

I'm on stock firmware so whenever that gets updated, I'll try to revisit everything. The gcode is fairly straightforward to edit if someone wants to give it a spin with 1.5 in the meantime. Let me know what values you want to test and I can easily generate a new test model with the proper labels. Happy to share the OpenSCAD source as well.

RE: IN SEARCH OF: Meaningful Linear Advance Calibration

Now that 3.9.0 with LA 1.5 is out, have you reworked your test project at all?

RE: IN SEARCH OF: Meaningful Linear Advance Calibration

Posted by: @david-w33

Now that 3.9.0 with LA 1.5 is out, have you reworked your test project at all?

I'm doing updates on my notes pages but haven't re-printed the tower with new values yet. It's in my pipeline.

I have found that the automatic translation between LA1.0 and LA1.5 isn't ideal, so I'm going to do my own calibration for each filament over again.

RE: IN SEARCH OF: Meaningful Linear Advance Calibration

You may want to hold off a bit. 3.9.1 will change the values again now that they've fixed some bugs in the initial implementation. 

It should also fix the conversion to better values.

RE: IN SEARCH OF: Meaningful Linear Advance Calibration

Posted by: @vintagepc

You may want to hold off a bit. 3.9.1 will change the values again now that they've fixed some bugs in the initial implementation. 

It should also fix the conversion to better values.

Good to know. I've got 3.9.1Rc1 running now, but I'll shuffle other updates ahead of the update. Thanks.