The physics of cooling?

[madoverlord 0]

TIL, excellent bit of info. I'll have to add this to my evil bag of tricks.

 

[madoverlord 0]

OK, I cooked my silicone shield for 30 hours in the rice cooker and it came out fine. https://www.dropbox.com/sh/ou1z52iwtq4hbtl/XhhmSoJJEN.

Interestingly, I got a bit of delamination on the mold, but not on the z-axis layers. Flakes of the surfaces in direct contact with the silicone detached (but were easily removed). One thing I did do is dip the mold in water after lubing it up with vaseline, in the hopes this would speed the cure.

I cut a rectangle out of the top of the shield so that the top edges overlapped the top of the hot-end, then punched holes for the heater and thermocouple wires, and added a slit up to the top so that the shield could be wrapped around the hot-end. I also had to chop off the bottom of the nozzle part to get adequate clearance.

After putting the shield in place I added a tie-wrap for extra security.

Temperature measurements (sorry for the mixed units) using the stock fan and duct at speed 255 with Z=0, using a crappy meat thermometer.

Hot End Off: 78F

Hot End On, No Shield: 92F

Hot End On @ 210C, Shield: 80F (!)

I'm currently doing an overhang print test to see if the shield actually makes a significant difference.

 

[madoverlord 0]

I have updated the gallery in my previous post with some pictures of my v2 shield, which more completely covers the hot end and the wiring. I need to make some updates to the design (for example, the tunnel for the wires is too small), but it appears to be an improvement.

Things I've learned:

 

• You need to completely clean the inside of the shield and remove all traces of the vaseline, otherwise it will burn and cake on your hot end.
  
• The cornstarch trick for curing works. Go easy on it, it works really well. I'd suggest maybe 10-20% cornstarch mixed in.
  
• Tiewraps become brittle when kept warm.
  
• The shield made only a minor improvement in the quality of supported overhangs using the stock fan.
  
• Those big triangular paperclip things make excellent clamps.
  

 

[madoverlord 0]

Did another pass at the mold, it's getting close to publishable. This version (https://www.dropbox.com/sh/ou1z52iwtq4hbtl/XhhmSoJJEN) is a boot that slides around the hot end and wires.

Noticed that the nozzle of the tube containing the silicone appears to be threaded 7/16NF20; if so I should be able to make an injection mold version, screw in the tube, and fill the mold that way. Might improve the surface quality a bit, and eliminate mold lines and flash.

Comments / suggestions?

 

[owen 19]

Looks much better MadOverlord. Do you have vents for the rubber to exit from your mould? It's a bit hard to tell. Thanks for you efforts.

 

[madoverlord 0]

The entire top (it's made out of two-subparts) of the current mold is open, I was just spreading silicone into the bottom part, adding the insert and screwing it down, adding more silicone to the edges and the top parts and clamping them in place, and then just scraping off the excess.

I am currently printing a completely enclosed injection mold. When finished, it will have an injection port that is threaded 7/16NF20 to fit the tube of silicone, and several small 1mm exit vent holes on the diagonally opposite faces. Hopefully this will be enough to avoid trapped air bubbles.

I also printed a toothpaste key to use as an injection device.

I'll do the injection test tomorrow morning so it will probably be Wednesday before I can report results.

 

[foehnsturm 969]

Tried some air ducts now ...

However, with the racing vase 90-100 mmm/s and 1 sec layer time seems to be the maximum for my setup.

Immediate cooling of the extruded filament works very well, which allows to print bridges of arbitrary length without any sagging (each bridge consists of 2 layers with 0.12 mm height and 4 lines).

On the other hand, heat accumulation caused by the print head returning before sufficient heat dissipates, is difficult to tackle. Some times, small objects or tiny details stay within the cooling air flow long enough, sometimes not. It just depends on the object geometry and the print head path.

 

[illuminarti 18]

Wonderful results on the bridging. Did you come to a conclusion as to whether any particular nozzle design was better or worse than any other?

 

[madoverlord 0]

My injection mold had a few problems. A couple of the bolts interfered with the insertion of the silicone nozzle (which did work) which meant there were some leaks, and I need some more vent holes (and I'm going to thread them this time so I can use small bolts to seal them).

On the bright side, the silicone was definitely flowing through the mold.

Anyone got any tips before I print up another mold?

WRT cooling fan design, it occurs to me that assuming you have good general area cooling and have minimized the heat transfer from the hot end -- and foehnsturm seems to be on the right track in this regard -- the real gains are going to be in making the slicer more aware of the physics. In the meantime, there may be a bit of a win by using more of the air for general area cooling.

 

[madoverlord 0]

PS: foehnsturm, have you posted your design files for your fan setup anywhere? I would certainly like to take a look and experiment with them.

 

[foehnsturm 969]

Thanks Simon!

Well, I'm using a radial fan, so static pressure is an issue. Some designs failed because the air duct added to much resistance to the overall system.

The two-folded ducts worked almost as well as the four-folded. But as you already mentioned, they are more sensitive to part geometry and orientation.

The best design has medium-sized orifices. Larger openings seem to reduce flow velocity to much.

I tried to increase the distance to the nozzle in order to cover a larger area with cold air - no success, maybe due to the reduced flow velocity at the nozzle as well.

I would like to aim directly at the tip of the nozzle, but this is always a compromise. You want some clearance to the print bed as well as having a chance to see what's happening when printing the first layers.

I have like 3 designs quite similar to the one which is mounted, aiming a little higher and a little lower. But this one performs slightly better. Honestly, I don't know why.

 

[foehnsturm 969]

@MadOverlord

I will post the designs. But stupidly, the file for the best working duct has been overwritten. So I will have to "reverse engineer" it.

You would also need the print head, fan mount, aluminum parts, see http://umforum.ultimaker.com/index.php?/topic/2435-new-fancooling-design/?p=18362

The center block of the print head will be redesigned as well as I switched from FreeCAD to Moi3D (Moment of Inspiration) thereafter, which I'm absolutely in love with now.

 

[foehnsturm 969]

I was wondering how bridging works when closing large open areas:

Box: 40 x 80 mm wide, 4 infill layers

KISSlicer, 0.12 mm layer height, infill / solid infill speed 50 mm/s

Ultimaker: xy-jerk set to 4

Jerk has considerable influence, with the default setting some 5% infill lines didn't bond to the perimeter.

... no lines lost.

 

[drayson 75]

Whow - great work!

I like this "tentacle" design - I also thought of something like this but with 4 small 25mm fans in it. I have only a rough concept sketch up to now.

Do you think that might worth trying? How much air flow was necessary to get your bridging result?

 

[foehnsturm 969]

I doubt if 25mm fans give enough air flow.

I have one here, the blades are ridiculous small and it produces just a very, very gentle breeze. 4 x 25mm will be inferior to 1 x 45mm.

The fan I use is rated with 33m³/h (19,42cfm). I assume that flow resistance of the flex hose and the print head kills about 30 to 50% of that.

 

[gr5 2,168]

The fan I use is rated with 33m³/h (19,42cfm). I assume that flow resistance of the flex hose and the print head kills about 30 to 50% of that.

 

More like 90%. I'm always amazed how an open tube can create such back pressure. You need a special kind of fan (more like an impeller) to fight back pressure. Like those blowers they use to inflate a bouncy castle - that kind of fan wouldn't be affected much by a hose. But the fans used for cooling electronics can't handle hoses or restrictions well.

But 10% is better than 0%. Much better. And is probably enough in many situations.

 

[madoverlord 0]

One rough way to guesstimate it is to feel the relative pressure coming out of the end of the hose vs. what's backing out of the fan itself.

An Adda AD2012LB-K70 20mm fan generates 0.6CFM unconstricted. and a AD0312HB-D50 30MM x 15MM (pretty chunky) fan gives 4.2, and an AD0212LX-K50-LF 25MM gives about 1.8.

Given the constraints on the stock hot end, you could only mount 3 fans on it, of up to 40mm size. Your big issue is going to be mass. The bigger fans are very heavy. But maybe 3 of the 30mm fans blowing in a triangle?

 

[foehnsturm 969]

That's why I use a radial fan vs. axial. Radial fans can build up more static pressure. But given the original 12 V fan supply it is a compromise of course.

But you are right with the tubes. I already designed 2 high-volume downdraft hoods for my wok cooking stations and I know, flow resistance of (low diameter) flex hoses can be a nightmare.

I tested the setup with fan, adaptors, various tubes (30, 40, 50 mm and without tube at all for comparison) and print head before mounting. That's how I came to that guesstimate. If it had been inferior to my normal setup with 2 axial fans and ducts, I wouldn't have mounted it.

 

[drayson 75]

Thanks foehnsturm for the explaination.

My intention was to have less flow but from 4 sides - but following your infos now I'm also think thats not worth trying...

My current fan shroud produces a lot of back flow so I'd like to change it to a system cooling more than only one side... Any suggestions what's worth trying?

[msurunner 2]

I have been thinking that one of those camping mattress inflatable pumps would be a decent means of moving a lot of air quickly and it's already designed to work with some sort of tubing setup. I'm assuming that's similar to what you have running on your's foehnsturm? BTW, I'm getting our budget numbers shortly and then I think I can order up a set of those pulleys if you're doing another order...

 

[foehnsturm 969]

@drayson

I used 2 40 mm fans quite similar to the UM2. Worked fine, but (PLA) ducts desintegrated over time and I wanted to get rid of the bulky parts moving with the print head.

@MSURunner

Those pumps would be fine but they draw considerable current (additional power supply). As far as I know, most compresor-like systems can only be switched on and off. Unfortunately, there are only a few 12 V, low amp, speed-controllable fans /pumps, which provide sufficient air flow and static pressure.

There will be a possibility to order pulleys in the future. I just will have to switch to a kind of webshop in order to reduce the efforts for messaging and waiting for answers.

 

[madoverlord 0]

I used 2 40 mm fans quite similar to the UM2. Worked fine, but (PLA) ducts desintegrated over time and I wanted to get rid of the bulky parts moving with the print head.

I wonder if the silicone hot-end booty that I've been working on (lastest version, injection-molded, is in the dishwasher right now) would let you get closer to the hot end with a PLA duct.

Or perhaps a couple of these with printed mounts that put their outlets close to the hot end: http://www.amazon.com/Brushless-Blower-40x40x20mm-Sleeve-bearing-Skywalking/dp/B00BKDV7JO

 

[madoverlord 0]

Here are https://www.dropbox.com/sh/gwvrk66fgntoe9e/VQS-8oIHaB#/ of the injection-mold variant of the booty. I used a higher-temp (750F max) silicone, and got a fair amount of delamination of the mold surfaces. This may be because this type of silicone releases acetic acid during the cure, or because I cooked the mold in a mason-jar inside a rice cooker. I'll do an experiment to see if I can improve on that.

On the plus side, I got a complete casting that literally fits like a glove.

Side note: you can make plastic parts close to the hot end more heat-resistant if you stick on some aluminum foil tape. It reflects some of the heat and rapidly conducts some of it away from the heat focus.

 

[owen 19]

Looks great MAdOverload, thanks for the updates. Looking forward to seeing the files.

 

[madoverlord 0]

I'm doing one more cast to refine my instructions (hoping I can reduce the delamination issue), and absent any disasters the files will go up on thingiverse middle of the week.