[Cura 4.6.1] Calculation of Body height with Use Adaptive Layers wrong?

[DivingDuck 105]

Hi,

Somehow it looks like the high of a model is wrong calculated in Cura 4.6x. Can someone please take a look on this too? Maybe I oversee something. I attached a simple project file with a  body that should have a hight about 25 mm. Actually I see a hight of 24.75 mm in the G-code file when I select [Use Adaptive Layers]. Deselecting [Use Adaptive Layers] I see the exact hight of 25 mm (=125 layers).

Please find attached the project- and gcode file.

 

 

 

BugelKunstoff_v461.zip

[DivingDuck 105]

11 day's old and not even a comment. By the way, there is also a problem with retract and G92.

A frustrated Cura user.

[Guest maht]

Sorry this must have slipped through the net. Thanks for the project file. How are you measuring the dimensions with gcode?

[DivingDuck 105]

Hi @maht,

Thanks for your response.

I "measure" this with the z-values in the gcode (regarding hight). I can give you a link for the CAD file, if needed. I had some serious quality complains with parts from the last two versions. I was out of ideas and finally start to analyze the gcode of some projects with technical parts.

 

Where to find it in code: when you look from starting line 90380 (last but one layer):

 

G1 F1350 X103.051 Y63.825 E1319.56458 <--line 90380
G0 F7200 X103.617 Y63.825
G1 F1350 X104.228 Y63.214 E1319.60661
G1 X104.298 Y63.144
G1 F2400 E1317.40661      <--line 90384 retract
G1 F300 Z25.45     <-- line 90385 move head 1mm up
;MESH:NONMESH
G0 F7200 X66.412 Y73.713 Z25.45
;TIME_ELAPSED:10663.609631
;LAYER:138
;MESH:B??gel Kunstoff.stl
G0 X66.412 Y73.713 Z25.75
;TYPE:WALL-OUTER
G1 F300 Z24.75   <-- line 90393 start last layer, but last layer should be 25.00 mm
G1 F2400 E1319.60661
G1 F1350 X66.399 Y74.474 E1319.64363
G1 X66.399 Y88.495 E1320.32565
G1 X53.799 Y88.495 E1320.93855

 

This happen in all models I had test after I recognize this and I was wondering why parts starts not fitting correct. You can verify the hight with deselecting [Use Adaptive Layers] in the project file and generating the new gcode file. In that version the body hight is correct. I came to this because if have some complex parts where multiple level highs have to match the accuracy of 0.1 mm. My prints usually fulfill this accuracy in hight, but as I recognize now, not for gcode generated from at least with the last two versions (beta and 4.61) of Cura.

 

The other point is right in the beginning with retract:

 

M82 ;absolute extrusion mode <--line 10
M140 S81.0   ;Start heating bed
M104 S235 T0 ;Start heating extruder
M190 S81.0   ;Wait for bed to reach temp before proceeding
M109 S235 T0 ;Wait for extruder to reach temp before proceeding
G28 ;Home
G1 Z15.0 F6000 ;Move the platform down 15mm
G92 E0
G1 F200 E3
G92 E0
G92 E0  <--line 20, not part of any start or end gcode
G92 E0 <--line 21, not part of any start or end gcode
G1 F2400 E-2.2  <--line 22 retract, not part of any start or end gcode
;LAYER_COUNT:140
;LAYER:0
M107
G1 F300 Z1.2
G0 F3600 X107.139 Y45.552 Z1.2
;TYPE:SKIRT
G1 F300 Z0.2
G1 F2400 E0 <--line 30 where is the compensation of retract from line 22 ??
G1 F1350 X107.629 Y45.678 E0.01641

 

Now imagine what happen if your retract is far longer than 2.2 mm and you are using no skirt or the part is small and so on. This will result in a damaged first layer (missing length of filament, no steady flow in the beginning of a first layer)

 

In addition an little irritation starting line 20:

There are too many "G92 E0". The extruder have no start and end gcode, but the engine put in these codes. In the end it do not matter but there shouldn't be any additional code if there is no reason for it.

 

 

 

[DivingDuck 105]

Ah, I just forget a last thing:

Is it maybe possible to change special characters (in this case german umlauts) in gcode descriptions in the same way as for filenames (or in its asci equivalent like ü to ue)? It shouldn't be a complex problem to do this automatically in a future version 🙂

 

 

[DivingDuck 105]

@maht, any new cognitions on this?

Edited May 19, 2020 by DivingDuck

[Barry37 0]

On 5/19/2020 at 8:24 AM, DivingDuck said:

@maht, any new cognitions on this?

I've noticed this, too (Lulzbot Taz 4).

Printing thin strips 110 x 3mm with cylindrical "bolt heads" which should be 0.8mm high for the strip, and 0.7 for the "bolt heads". There are no sloping verticals. With 0.25mm 1st layer, and 0.12mm other layers, slicing with Adaptive Layers ON, only has 6 layers - 3 + 3. and is only 1.2mm high (the strip itself is only 0.54mm thick. The layer height doesn't seem to have been varied, there are just less of them

Without Adaptive Layers, there are 10 layers, 6 + 4, and the print is just over 1.4mm high.

I've just been back to Cura 4.0 and replicated the print settings - it does exactly the same.

My understanding was that when Adaptive Layers is used, there will be thicker/thinner layers, but only if there are sloping or rounded vertical surfaces to improve their appearance, but overall dimensions  (well, height) would stay the same.

Doesn't seem to be the case.

[DivingDuck 105]

A little bit late, but thanks for your feedback.

Your observation is the same I and other users of my printers have seen. And it is wrong with 4.6.2 too as well as the retract issue I reported on may 13 here.

Anyway, I had a lot of hard discussions regarding Cura last and this and even more this year (I'm supporting some small companies here in my local area). We took a decision and changed the slicer since beginning May.