Cooling fans: when NOT to cool, and why?

I'm having a little trouble understanding the "best practice" consensus for cooling fan use. The basic understand (cool the plastic once it's out of the nozzle so it reaches a solid state faster and doesn't deform) just argues to "leave the fan on all the time".

But looking at existing slicer software, there is a huge box of voodoo heuristics that can be applied: we can not cool the first layer; we can not cool at all; we can not cool "mostly" but turn the fan back on for small layers...

What's the deal here? What are the arguments, really, for leaving a working fan off? My own experiments with a new ABS spool (seems like consensus is that all this voodoo is mostly for ABS) don't really show much in the way of meaningful effect. I'm tempted just to crank the thing up to 100% all the time and leave it there.

What are the actual failures these settings are addressing?

Cooling is mostly just for PLA. PLA can create some pretty good overhangs and bridges with active cooling of the plastic as its laid down. It has little warp and so can take the cooling pretty well. ABS wants as little cooling as possible due to its high tendency to warp with temperature differences. In fact, ABS hates the cold so much that you should really print it in a heated chamber between 40-70c.

The main point of cooling is that plastic that has just left the nozzle is still very liquid and can curl upwards. The fan is designed to bring it into or just below glass point so that it cannot curl. Some PLA will always need cooling, others can live without it. sometimes making layer times longer can have the same effect to allow plastic to naturally cool.

You should never cool the first layer.... Most print surfaces will cause the plastic to peel off too soon.

Edited 1 time(s). Last edit at 03/18/2017 09:53PM by Origamib.

I thought part cooling was more for PLA than ABS, not sure its needed there.
No Cooling for first layer or 2 then ramp it up, and used as you say for overhangs to stop curling.
(oops pipped at the post)

Edited 1 time(s). Last edit at 03/18/2017 10:35AM by MechaBits.

never had an issue cooling pla for all layers.. Been printing on glass bed for years.

The reason people don't usually use a cooling fan on ABS is due to the amount it shrinks. If you cool abs then the cooler layers shrink and that can cause warping and a lot of problems with prints lifting from the bed.

Additionally, with any plastic, you want the newly extruded plastic to slightly remelt the layer below as it goes down, in order to get a strong bond between layers. If you cool any plastic too much you will end up with a very weak print which is liable to delaminate with shear force. Again, this is far more pronounced with Abs

However, cooling is very beneficial for getting sharp lines and corners on models. If you print PLA without cooling you will quite likely get curlig at the corners. This is a process that often makes itself worse,layer by layer, because a corner that has curled up might touch the nozzle on the next layer and get even hotter.

Bridging performance is much better with cooling. Even with ABS. In fact, there are lots of more complex models that are close to impossible to print in abs without support, that are comparatively easy in PLA with cooling.

I actually do use cooling on ABS prints, where necessary. The cooling I use is much more conservative than with PLA. I start at later 3 and ramp up the speed very slowly up to 40% and have the fan speed increase for short layers and bridges. It really helps for models with small islands too.

But it's a trade off because these parts are noticeably weaker than without cooling. Sometimes that's a trade off that it's worth making.

when I started printing I didn't use a cooling fan and prints were coming out very bad (looked floppy on overhangs etc....so I hooked up a fan on constant and the prints came out perfect every time after that, heated bed is off with no heat in it at all...this is for PLA and works great for me..

There is no real consensus on best practice for cooling. What works on one printer may not work on another, same goes for prints, materials, slicer settings and even the ambient temperature. Even something as simple as changing the nozzle for one ostensibly exactly the same can change bridging etc..
My method is to cool the freshly extruded plastic as quickly as possible to preserve fine detail by blowing air through an annular ring close to the nozzle. I blow no more air than is needed through the ring and supplement this on PLA with a gentle but turbulent breeze from the hotend cooling fan.
This method works for me but I don't adhere to it with any real conviction other than it seems to work better than any recommendation I have followed.

Mike

I use full fan (dual 30mm fans) for PLA and PETG after layer 3. For ABS after layer 3, I use the minimum speed that will turn the fans something like 18% all the time, unless I'm bringing or printing an overhang then it's 100%. I also use the slow down feature for layers less than 15 seconds in duration this really helps with fine deal.

I think there is often a bit of residual confusion from when hotends were J heads and did not need hotend heatsink fans for ABS, but needed them for PLA. With E3d and most other modern hotends the hot end fan is always on.

Try to find or design, a small object with some increasingly severe overhangs, a few bridges and print it with various fan settings such as none, 25%,75% and 100% (after the first few layers to ensure they adhere to the build plate). Try this for different materials and also try selecting fans just for overhangs/bridges. You'll soon see what your machine needs.

Edited 1 time(s). Last edit at 03/22/2017 12:10PM by DjDemonD.

---

Simon Khoury

Co-founder of [www.precisionpiezo.co.uk] Accurate, repeatable, versatile Z-Probes
Published:Inventions

Quote
DjDemonD
For ABS after layer 3, I use the minimum speed that will turn the fans something like 18% all the time, unless I'm bringing or printing an overhang then it's 100%.

Never really occurred to me to just use the bare minimum fan speed. I'm going to try that.

How do you find 100% on bridging though? I find that i can get parts that look as good as PLA, but which are very much lacking in intra-layer strengh.

Erm not really sure what the question is? I would never use anything except 100% fan for bridges.

Do you mean using ABS your parts look good but are weak? Try higher hotend temperature I print 1st layers at 250 then 240 after that I only lower it if I can see the filament is breaking down or smoking.

---

Simon Khoury

Co-founder of [www.precisionpiezo.co.uk] Accurate, repeatable, versatile Z-Probes
Published:Inventions

Yes, axactly that. High fan speeds make parts that look like PLA but they delaminate very easily. It's trying to find a balance between cooling for part definition and part strength.

I don't get away with 100% at all, I was wondering if your parts are a. It weak near bridges.

I guess at the end of the day there are a lot of variables in part cooling, how miuch air the fan shifts and the air pattern and proximity to the nozzle.

Well the fan is only on 100% for the first layer of a bridge, to actually achieve spanning the bridge. After that it just lays down ABS as if its a normal layer i..e minimal fan. So no parts are not weaker near bridges.

If I need a 3D printed part to be strong I use PETG, the interlayer adhesion is so good that the parts are more representative of injection moulded plastic than 3D printed. But at the expense (potentially) of increased mass, as it is a lot more dense, less good surface detail as it is far more liquid when extruded, and more stringing but thats not really a problem, hit it with a flame briefly.

I totally agree with your last point. Air volume, how diffuse or directed the airflow is, and how omnidirectional the air flow is all affect the cooling of parts. Generally I'd say you want to cool from at least two
opposite directions or material will curl towards the fan.

See the thread on pumped air cooling there's quite a bit of experience shared with different cooling techniques. Some suggestion here that more diffuse air flow seemed to work better, which is interesting and I feel counter intuitive.

Edited 1 time(s). Last edit at 03/23/2017 10:02AM by DjDemonD.

---

Simon Khoury

Co-founder of [www.precisionpiezo.co.uk] Accurate, repeatable, versatile Z-Probes
Published:Inventions

Following up with my results after DjDemonD's suggestions.

My cooling fan turns at 18% of pwm. So i'm doing that from L3 and then 100% bridging speed.

Here's a couple of photos of an ABS print including an overhang and a bridge. I'm pretty pleased with the results, it's almost as good as PLA but this is ABS and it's still really strong.

From my understanding, you can print ABS with a cooling fan, but the climate has to be controlled so that the air blown is hot enough to not warp the part but cool enough to solidify it. Basically, only in an enclosure with the ambient temp being monitored and controlled.

Cooling ABS when not enclosed would cause layer delamination and severe part warping, right?

Not exactly, the ambient air above my 330mm circular heated bed which is at 120 degrees Celsius is quite warm. You can print a moderate sized ABS part without an enclosure on a large bed. Enclosures help, and enable you to print long thin parts, large parts etc... with far less drama. But get some ABS and have a play its not just for Jedi Masters. Try printing some successively larger objects and see what happens.

---

Simon Khoury

Co-founder of [www.precisionpiezo.co.uk] Accurate, repeatable, versatile Z-Probes
Published:Inventions

Quote
rich1051414
From my understanding, you can print ABS with a cooling fan, but the climate has to be controlled so that the air blown is hot enough to not warp the part but cool enough to solidify it. Basically, only in an enclosure with the ambient temp being monitored and controlled.

Cooling ABS when not enclosed would cause layer delamination and severe part warping, right?

That's the conventional wisdom, but (certainly for me) it's just not true. The pictures i posted earlier were of a part printed with cooling on my replicator clone printer. So although not as open as a prusa i3, for example it's pretty open. The part is really strong - i've flexed it as much as i'd like to with my hands and no signs of it failing. it seems every bit as strong as non-cooled ABS, and is much stronger than this part would be in PLA.. It was printed in rigid.ink abs btw.

I get excellent layer adhesion (too good in a way, sometimes it's hard to get parts off) just printing on to clean kapton on 3mm aluminium, and i don't really get parts come unstuck unless they are large and i cool them too much.

I think the reality is that there is a balance. A little bit of cooling stops curling, and isn't so much that you get warp or really weak layer adhesion. A little bit more helps bridging no end.