MK4 cooling problem? bad print
I designed a small cycloidal gear. I like to print it vase mode so there are no seems.
The result is however disappointing.
I printed it in vase mode with a 1 mm perimeter thickness.
The side where the cooling van is looks as expected.
The other side however looks horrible. During print you see this side curling up.
RE: MK4 cooling problem? bad print
This is more then interesting for me, since I can see similar issues (no vase mode; from customized to standard MK4-printing profiles) when printing detailed, small, thin parts.
I was wondering why strange optics have been happening on one direction only - it was more or less, looking at the printer from the front of course, on the right side of the prints or on the upper right, difficult to say.
At least it looks similar to your effects, I have checked belt tensions, movement of X- and Y-axis, nearly everyting, but never had the idea it could be caused by cooling and the fan.
I am running MK4/MMU3 in prusa enclosure, printing with dasfilament PLA mainly.
Would also be happy for any feedback on this!
RE: MK4 cooling problem? bad print
I have had good results using this fan duct: https://www.printables.com/model/588524-prusa-mk4-improved-fan-duct-bi-directional-cooling
RE: MK4 cooling problem? bad print
I have had similar problems on gears with overhangs (not vase mode). A modified fan duct helped somewhat: https://www.printables.com/model/535705-prusa-mk4-fan-shroud-for-better-tpuall-other-filam. Changing the dynamic overhang settings in the slicer may help as well. There are some more details at https://www.stevesclocks.com/forum/general-discussion/print-quality-with-prusa-mk4.
RE: MK4 cooling problem? bad print
Thanks for the replies. I will try this.
RE: MK4 cooling problem? bad print
I know what the problems is with the MK4 cooling.
Unfortunately the 2 minute video t0o large to upload to the forum. But here is the resulting ugly print.
The left front side is ok, this is side where the cooling fan is blowing.
vase mode, 1 mm perimeter, 0.2 mm layer height
The MK4 cooling fan is blowing to one side of the model. This cools the model on one side only. Causing it to shrink one side only.
The side that is cooled shrinks more than the other side. The hot side curls up and we have a bad print.
The Prusa design choice for a one directional airflow seems to be a sub optimal choice. As I discovered later the cooling is actually causing the bad print.
So the question was will a different cooling fan duct improve the print.
I first tried the design as suggested by @walter-layher, this did indeed improve the other side of the print. But made the good side a little worse.
Then I tried another slimmer design, this one distributed the problem over the complete model. The duct on this model is a lot smaller and overall did a lot worse.
Designing a good cooling system seems to be a hard problem.
So if cooling one side is causing the model to deform, what if we disable cooling. For this model without large overhangs and bridges this should be no problem. This is exactly what I did. The result was surprising, the printing artifact was gone. The model was a bit twisted by the pulling force of the nozzle. I lowered the print speed and got a good printed model. You have to wait for the model to cooldown but that is acceptable.
I printed this model in vase mode 1 mm perimeters. The standard value is 0.45 mm this is more than 2.2 times the filament that is much more heat than the little fan can cool before the next layer is printed. So for this case the MK4 cooling is not sufficient. Let's try this, and yes this works. I tried it with the original duct design and there is almost no difference, between the large two sided duct and the original design.
@korken78
For small parts you will most likely see improvements if you just lower the print speed. You must give the filament time to cool down before the next layer is printed. For small parts and pointy details the shrinking forces are relative high. Consistent cooling for the whole model is the key here.
RE: MK4 cooling problem? bad print
@chipcode: thanks for bringing this up. I was able to identify that using Prusas standard 0.2 mm structural (=slower!) is delivering better results than 0.2 mm speed profile. Will do some more try and error.
For small parts you will most likely see improvements if you just lower the print speed. You must give the filament time to cool down before the next layer is printed. For small parts and pointy details the shrinking forces are relative high. Consistent cooling for the whole model is the key here.