Crossflow Cooling for UM2

[avogra 68]

Just thought that the discussion which came up there might be better continued here in the forum.

@00D00B, got your pictures ;)Same picture with light coming from different directions.

Not too impressive yet. I think, the fanshroud currently blocks a lot of air from the blowers. Hopefully removing it will help.

@gareee, I also see the problem, that when the buildplate is lowered, the air won't flow straight under the nozzle but instead sloped down towards the edge of the buildplate. Also when the buildplate is very far down, the blowers won't suck the air from under the buildplate anymore. I thought about some airguides, that force the air to flow at the height of the nozzle and back under the buildplate. Simple sketch:

No idea if this works out and if my understanding of flowing air is right at all :PAt least the current setup doesn't create a laminar airflow yet. From what I understand, that is the main point in foehnsturm's setup. Maybe I can get that too with a better nozzle geometry. If not, those air guides might still give a decent cooling. I will try on and keep you updated!

Edited April 5, 2016 by Guest

[avogra 68]

Oh, forgot to comment the pictures. These are 45° overhangs printed at 0,1mm layer height with 40mm/s and 205°C.

[Labern 775]

You can also rotate one fan on the fan shroud so it sucks instead of blows, so the air will blow down on one side and suck up the other.

[avogra 68]

good thought! i will try it.

[00D00B 58]

Thanks for pics ! Not too impressive that's true but I'll continue to follow your topic

[tommyph1208 55]

Pics from the side would be better, no? To show the quality of the overhang?

These top down ones look pretty much the same to me...

You talk about removing the fanshroud from the printhead which I think is a good idea...

This is what I did on my UMO, and to me that is half of the benefit from the crossflow approach; A less bulky and lighter hotend carriage with (at least) the same cooling capabilities and print quality as the carriage mounted fans and shrouds...

Edited April 6, 2016 by Guest

[00D00B 58]

Not sure we can say crossflow would totaly replace the fanshroud. Don't forget it blows only from one side (from now at least) and the filament needs to be cooled down fast and precisely.

In fact, this system just cools down the print area, which is intresting for the first layers and printing hard parts but fanshroud seems to be important. I mean a labern version, not the old UM2 one.

[tommyph1208 55]

Not sure we can say crossflow would totaly replace the fanshroud. Don't forget it blows only from one side (from now at least) and the filament needs to be cooled down fast and precisely.

In fact, this system just cools down the print area, which is intresting for the first layers and printing hard parts but fanshroud seems to be important. I mean a labern version, not the old UM2 one.

 

Yes I have heard this argument about the "good" and "poorly" cooled side of prints quite frequently...

I have myself removed the carriage fan and fan shroud entirely, and gotten used to orienting my parts so the "hardest to print" (most overhangs, details etc.) side is turned towards the cross-flow fan, also often printing in that side of the print bed if the model doesn't take up the entire build space.

For me I would say that this approach has been quite sufficient in almost all cases.

Edited April 6, 2016 by Guest

[avogra 68]

@00D00B, if you have a look into the original thread by foehnsturm, they achieve super good results with only the crossflow fan. I think foehnsturm also added a small wedge to the printhead to deflect some air to backside.

@tommyph1208, I did some pictures from the side, but I think the topdown versions show more detail. Have you looked at them in full resolution?

[tommyph1208 55]

@tommyph1208, I did some pictures from the side, but I think the topdown versions show more detail. Have you looked at them in full resolution?

 

I have yes, but it is hard (impossible) to see the quality of the outer walls (which I am most interested in in relation to overhangs... )

from your pictures I can see that you have fine top layer quality in both cases...

[avogra 68]

oh sorry, my fault. what you see is the bottom side, the top surface of the prints is laying on the paper.

[tommyph1208 55]

oh sorry, my fault. what you see is the bottom side, the top surface of the prints is laying on the paper.

 

Aha.... The plot thickens

I though that what I now understand are overhangs were actually a brim

I would still love to see a side picture though, I think those show the "sagging" errors of overhangs well...

If you want to share the models I will be happy to give them a whirl with my crossflow setup and post some results here...

Edited April 6, 2016 by Guest

[avogra 68]

I can share it tonight. Or you create it yourself It's just a cylinder with diameter 20mm, height 4mm and then an outward cone on top with 45° and 6mm height. the square one is 20x20mm with 4mm height and then a capped pyramid with also 45° and 6mm height. Can you also print with 0,1mm layers, 40mm/s, 20% infill and 205°C? Then the main difference should be cooling.

[Labern 775]

I think you are better off with an overhang of 60 deg +

I can get really good results at 45 and you want a bad - better result no?

[avogra 68]

@Labern, you have quite a point. With my current setup even 45° on those smaller parts become mediocre, but of course that is a weak benchmark. Up to now I didn't bother to print a better fanshroud like yours, as I didn't have high Tg filament at home and wanted to try the crossflow approach anyway.

Anyway I don't think it makes too much sense to look into those prints too close now, my implementation is still far from satisfying, don't need the prints to tell I would rather improve it and then come back with hopefully some better results

[tommyph1208 55]

I can share it tonight. Or you create it yourself It's just a cylinder with diameter 20mm, height 4mm and then an outward cone on top with 45° and 6mm height. the square one is 20x20mm with 4mm height and then a capped pyramid with also 45° and 6mm height. Can you also print with 0,1mm layers, 40mm/s, 20% infill and 205°C? Then the main difference should be cooling.

 

Sounds like a plan... However immediately after posting, I remembered that I currently have a 0,6mm. nozzle in my machine... That will have to be changed too as wider nozzles also greatly improve prints in the overhang department

Edited April 7, 2016 by Guest

[avogra 68]

 

@tommyph1208, do you still want to try those prints? I think it's not too useful. My current design has already shown to be only that useful, no need to prove it I would rather bring the topic up again when It shows a real advantage over stock cooling or other fan shrouds.

[tommyph1208 55]

@tommyph1208, do you still want to try those prints? I think it's not too useful. My current design has already shown to be only that useful, no need to prove it :PI would rather bring the topic up again when It shows a real advantage over stock cooling or other fan shrouds.

 

Alright cool I won't bother you guys in here with it ... I might try it still at some point just out of my own interest...

But please keep us posted on your progress, its always cool to hear about new mods and development, and your solution seems, if nothing else, magnitudes cheaper than mine.