Tuning overhangs

I’m trying to understand how to tune overhangs for my Polymaker ASA filament. I had dialed the fan speed way down to promote layer adhesion, except for overhangs. But overhangs came out shitty:

I had:

  • 0% fan @ 35s
  • 20% fan @ 5s
  • 100% fan for overhangs

I then adopted the bambu ASA fan settings (leaving everything else the same):

  • 25% @ 35s
  • 90% @ 3s
  • 80% for overhangs
    The result is much better:

Notice that the bad model has 100% fan for overhangs and the good one only 80%. So is less fan better?

Well, looking at fan speed, here’s the bad and the good versions:
(green is ~100%, red-brown is ~20%)
(green is ~80%)

Nozzle speed is identical between the two versions:

Sooo, my conclusion is that the overhang speed is not really effective, unless the model has quite long overhangs. Possible explanations:

  • the fan can’t change the airflow as quickly as would be needed here (dunno whether the head briefly stops for the airflow to change or whether it just keeps going)
  • the second wall loop, which is printed first messes up the outer wall loop, e.g. by still being too hot if the fan is essentially off

here is my settings not sure how good will be , and can you share the model i can print with my settings and see what speeds are you using for ASA?
Noticed that i reduced from 100% to 70% the overhangs and increased from 20% to 40% the max fan the last print , and reduced the 6s to 3s as i change sometimes based on the model not sure which model triggered that change

EDIT Assumption : I actually too much fan on the overhangs causes the ASA easy to warp(shrinkage) and depending on the percentage of overlap will detach and goes up sometimes down , then the top layer will sag will be no good . Good example is 55d and 75 d section of the failed one
I think i had it to 60% at some point , originally started with 100% and was bad
Also what i found on big models i print really hot110C BP/260C head , started building too much heat inside the object and some bridges and overhangs started failing so increased a bit the max fan speed to 40% but for shorter 3s.

1 Like

The model is the OverhangTest from CNC Kitchen found in Dropbox - FilamentTestModels - Simplify your life

Polylite ASA:

The actual nozzle speeds:

Your fan speeds are a bit higher than mine. I had them way down. I have some Z-layer strength hooks printed and will test them tomorrow to see whether I can find a difference vs. the “no fan” setting… Z-layer adhesion was why I dialed the fan speed down originally.

Printing …
These are the revised one , i hade the same problem with layer adhesion and pushed it down to 0-20 , and 100% for overhangs ,but then some issues on big objects , and overhangs.
Also did not solve completely the layer adhesion strength , improved it a lot

for the layer adhesion i found out that the best out come is when i reduce the speed bellow 100mm/S some times down to 70mm/S.
Now I use 140 walls/solid infill/top surface, 155mm/S sparse infill when want a quick print and no need for good layer adhesion. And 100 for final print or 70mm/S or when really needed. Also noticed that bumping the nozzle temp to 270C and 70mm/S makes it perfect for ABS+ , have not tried it for ASA yet.
And i just change the max volumetric to reduce the speed

My current settings are no good with this speeds especially above 65degrees,
WiIl play a bit and let you know how it goes , it was better at some point :frowning:

My print was perfect up to 65degrees , with reducing the speed to 100mm/S for after that is still an issue: Any ideas see bellow

Ok small update , i found what is the actual problem it was not only the speeds and temperature and fans. It was the slicing it self, Hm no wonder some overhangs were a problem from time to time

Over hang line printed as arcs needs tiny support under(preferable 25% or more but even 10% works of line width) here we have 2 rings with the second completely in the air and first also have part in the air. Reducing the speeds almost solves the issue , but …not a full proof solution , probably with a bit extra play with the fans will improve it but not solve it . Actually the line is supported only by side line which usually does not work well, As a result tries to do a ARC bridge in the air with potentially tiny support from the side but as the no pressure under , can’t make the line width any way

in Orca Slicer there is an option which solves it , but not sure for the BS as i have not seen it there
But it is not suppose to create an Arc 100% in the air
I will update with the best settings to print it - reducing the speed helps a lot, decreasing the layer height solves it, Line width 0.4 probably will improve , but have not tested it

I also reduced the Line height to 0.08 , which produces a better output and end result, probably increasing/decreasing the line width 0.4 for 0.4mm nozzle will help a bit as well

here is example from the prusa slicer and the problem from Orca/Bambu
Prusa Slicer : The correct one

Orca/BS slicer with 0.2 layer height really bad and :

This is Orca with 0.08 layer height looks correct:

This is Orca with 0.2 and make overhang printable option looks ok and printed no problem

On the good side if we can print a circles or arcs in midair would be great printer :slight_smile:
We need to find why the hell the slicing put almost 2 print lines in the air
I printed it perfectly with my settings and just reduced speed to 100mm/S to 65degress then a lot of issues

I have noticed the fans ramp up a little late.

1 Like

I’m not sure I want to get into g-code hacking, but for testing the hypothesis one could insert a short 500ms or so pause after the fan speed change. But this should really be taken care of by the firmware. (Better yet would be for the firmware to look ahead in the g-code and ramp up/down the fan ahead of time.)


Got around to testing a couple of Z-layer test samples (CNC kitchen ones). Conclusion: choose good Z-adhesion or good overhangs but not both!

Specifically, testing Polylite ASA using 20% fan speed for the CNC kitchen test samples results in ~28kg of breaking strength while using 60% fan speed results in ~18kg of breaking strength. XY samples sustain ~37kg breaking strength. A significant difference.

Breaking the XY sample by bending in my hand also causes the 60% fan samples to cleanly break at a layer line while the 20% fan samples bend and stretch more and eventually show an irregular fracture.

So there’s a clear trade-off here as long as the “overhang fan speed” setting is ineffective.


Is this tests with 140mm/S default settings or reduced to bellow 70mm/S i guess in the cross section would be percent of that ? And did you try 40% fan
I was able to print perfect with the 40% fan, and changing the slicing model, as per previous comment
After work will print one more and share the resuts

I did not try 40% fan, I only printed the tensile strength test pieces with the same profile I used that gave good overhang results. The nozzle speed is very slow due to the short layer time:

1 Like

Thanks , think i may have figured it out how to improve it but updated to the latest firmware and having heaps of issues with the models. i printed one with this firmware and saw improvement but as everything is … not sure yet . Will try to see if i can make the new firmware working but if not will just revert back to the last one

the idea is to use 30-50% fan speed and really to slow down the overhang probably around 40% fan
and have a balanced layer adhesion/overhangs .

Also want to see if there is easy way to start the fan a bit earlier as per @StreetSports comment but could not find a way with out editing the g-code


Not sure if its helpful, but Im printing the 1st test print right now and it got me wondering if infill pattern could also help. Large portions of the print dont have anything holding the tip of the overhang from the inside. If that makes sense. Probably a dumb idea and probably only works sometimes, but im wondering if different infill would help. Im suprised the software doesnt always assign an infill line to the sharpest points and then vary the rest of the jnfill around that point. I just woke up so im sorry if this makes no sense.

1 Like

Best of 12 tries. Tried alk kinds of tricks too. I did forget arachne walls on this last one, so im doinv 1 more lol
Arachne made it slightly worse.
With new firmware.


Pretty cool! Medium fan, lower temp, slowing down. Overall definitely an overhang vs z-adhesion tradeoff…

1 Like

Yeah. If I needed this part to be structural, without supports, I would raise the heat back up, run .08 layer height all the way and run another wall.

I am running test 13 now , and the best so far is very similar to yours, that is with 0.02 and with out the special Orca setting make overhangs printable
the best two are :

  1. overhang speeds did not matter so much as far as it is bellow 30mm for higher angles
    40mm/s, 25mm/s, 25mm/s,25mm/s,25mm/S
    fan 10% to 40% and and overhang fan 80% with 0% eg all outer walls included
  2. 0-30% fan and overhang 70% with 0% , with external and internal walls speed 40mm/s
    overhang speeds : 40,20,20,20,20
    Note that most outer overhang line is considered as a bridge so external bridge speed was used

Difficult to find the balance between really hot and overhangs above 65deg , but this test is really pushing it

Tried to lower overhangs speeds bellow for the same fan but got worse results
Tried fan only for overhangs above 25%, worse the fan starts really late

I did some more strength tests with my settings but slower fan speeds. Basically I have:

  • 260C, 20% fan, 0.3 layer @ 0.6 nozzle → ~28kg Z-force
  • 260C, 60% fan, 0.3 layer @ 0.6 nozzle → ~18kg Z-force
  • 265C, 40% fan, 0.3 layer @ 0.6 nozzle → ~24kg Z-force
  • 265C, 40% fan, 0.2 layer @ 0.6 nozzle → ~21kg Z-force

So Z-layer adhesion is basically inversely proportional to fan speed and overhang quality is basically proportional to fan speed :cry:

Something that would be interesting is being able to set different fan speeds for exterior wall vs. interior wall (& infill). Specifically, high fan speed for exterior wall to get good overhangs and slow speed for interior to get good adhesion. With 3+ wall loops this could be pretty close to best strength and close to best looks…

I think you are right, but there is may be one more option , i am experimenting with line width for increased layer adhesion noticed that bambu are using 50% fan and increased Print temp , which i did for ABS+ to solve adhesion problems

Some one 'borrowed my scales and cant do a stress test
can you please try just for caparison with you digits the following:
260C, 60% fan , 0,3 @0.6 and setup the line width for sparse infill to 110%or 120% and solid infill to 120% or 130% and may be the inner wall to 110% or any of your tests only change the Line width to see if there is difference . But do not bump the inner walls past 120% and outer wall should stay around the 100 or keep it what you have it at the moment

I just started my 16th test with 270,50% ,0.02, @0.4mm and changed the line width , a bit stollen from Prusa mk4, but also got a bit reading and for the outer wall kinda got to conclusion should be around 100% , but for the inners increased over 110% should give better layer adhesion, Also combined with my ABS+ print settings and Bambu ASA profile which are printing at 270C with 50% fan and 80% overhang
Not sending a lot of photos as nothing really good the best the same as [StreetSports], unless i really bump up the fans

1 Like

No matter what i do can t make it past 65 degrees including started to add AUX fan 30% and fan up to 100% different line widths, checked the filament calibration 0.02mm height , just to see,
did any one succeeded with some kind of settings to achieve 75 no matter the layer adhesion at this stage?
Some examples from the 22 prints and counting , some 65 is not ideal some is better some is almost perfect , but 75 never good, and always the warp on top

All the above is experimented with high temps (260/270 of the nozzle to keep the layer adhesion , but now will try with lowering it first will reprint the temp tower ,

Interesting observation
For test i did print the same test on the Prusa, with PLA with Orca Slicer and With Prusa slicer
With Prusa slicer perfect result , with Orca slicer much smaller but the same problem on the 75 degrees , and tiny warp on top with Orca sclicer, no warp with Prusa slicer . Same filament , same settings , same machine same environment

I wish i had a chamber for Prusa to try the ASA there , but i may be able to make Prusa Slicer work on the Bambu
The one on the right is Prusa slicer ,
The Left is Orca slicer not very clear from the photo if visible but the same problem very close to the best ASA result

1 Like

curious results on the prusa… is the wall order the same?