Time estimate variation from 20m to 1h34m for top layer 0 to 1 config change on small part on Stable release. (Dev release is around 40-45min)

[carlmorgan]

Version

Stable 2.3.56 and Dev 2.3.57.1_win64_211008

Operating system type + version

Windows 10

Behavior

Wildly different time to complete between 20m and 1h32m for only a change in Top Layer from 0 to 1 on a small print.

Observation that stable release:
Estimates 20m for with TopLayer=0
Estimates & Take ~1.5hr for with TopLayer=1

Observation that 2.3.57.1_win64_211008 release:
Estimates 41m for with TopLayer=0
Estimates 44m for with TopLayer=1

Seems inconsistent / unexpected variation - normally I trust the estimates pretty well.

Project File (.3MF) where problem occur

V0.1_Door_Handle_1toplayer.3mf.zip

Thanks, Carl

[neophyl]

Seems to be related to ironing speeds. if you turn ironing off then both are roughly the same but with ironing on then one is way different. That would also explain why the top layer 0 or 1 changes it as if zero then you cant iron either.

[carlmorgan]

OK - thanks for the details. For my education
a) is there any way of showing the 'limiting' function at any point in the print, ie is it flow, speed, temperature / cooling (if not might be some kind of feature request).
b) why does ironing try so hard all the way up the virtical sides of the print? I thought it was for 'top surfaces' only.
c) the difference between the stable 0 top (20min) and beta/dev of 0 top (41min) - is that just better estimation or something doing 'worse'.... (I'm kinda tempted to through filament at the two configs and time them....

[carlmorgan]

Follow up to c)
Dev did take 43min (without ironing), Stable (without ironing) 24min

So time estimate is accurate, Dev is taking twice Stable.

V0.1_Door_Handle_3toplayer_noiron_comparision_GCode.zip

[supermerill]

There is some things going on here.
On both:

• don't let gapfill & thin walls to 0 speed, unless you can have a really big speed, as these can have very thin extrusions

Why dev behave differently than stable:

• in dev, the ironing from the old "pattern" use now the same type as the new "post-process" taken from prusa, and it's now not taken into account for auto-speed compute as these almost don't extrude anything. They are using the ironing speed (and i just see that i forgot to activate it if ironing pattern is selected, have to fix that).

oh, and you may want to use thin_perimeter_all, at least on the top layer.

[carlmorgan]

Thanks for getting back to me - for clarity there WAS a 20min difference between stable and prod for the actual printing. I'm still learning the whole 3DP ropes. Related questions (or how do I help myself to get to the same info)

a) Are there any tools to visualise / show what the calculated speeds / accelerations are viewing a peice of GCode (assuming there are no machine limits)
b) Concept wise, if I set layer hieght/nozzle, filiament & bed temp, vol flow, max speed and max accell - with SS can EVERYTHING else be defaulted to '0' and it will go as fast as the 'slowest thing'? My little 120mm^2 printer seems to be capable of 'ludicrous mode' for accel and speed is limited by flow primarilly - hence question a) to understand what is the 'bottleneck'.

PS: Really enjoying the SS - took a bit to get my head around the new tech in general, but the access and development is great and calibration elements are really helpful.
PPS: Did you want "They are using the ironing speed (and i just see that i forgot to activate it if ironing pattern is selected, have to fix that" raised separately or will that be rolled in with other generic work?

[supermerill]

a) Are there any tools to visualise / show what the calculated speeds / accelerations are viewing a peice of GCode (assuming there are no machine limits)

select "speed" on the bottom left box, in the gcode preview
note that the time for each feature is available on the top left thing

b) Concept wise, if I set layer hieght/nozzle, filiament & bed temp, vol flow, max speed and max accell - with SS can EVERYTHING else be defaulted to '0' and it will go as fast as the 'slowest thing'? My little 120mm^2 printer seems to be capable of 'ludicrous mode' for accel and speed is limited by flow primarilly - hence question a) to understand what is the 'bottleneck'.

autospeed try to get the same volumetric speed on the whole object.
Max speed can be the limiting factor, or volumetric speed. If a feature us thin, its max speed is reached before the rest and will limit the speed of everything else.
That's why you have a 20min diff between beta & stable, because the 'ironing' very thin extrusions were taken from the 'top speed' to the 'ironing speed' and the ironing speed isn't controlled by the autospeed thigny.

pps: will be fixed now