Cooling with an air pump?

So I have been printing ABS for about an year and a half. Now that I think that I have more or less mastered ABS I made my first attempts at PLA. After doing some first prints, which were so-so because of insufficient cooling, I installed two radial 40mm fans blowing directly at the part. Huge improvement but still insufficient.

So I thought I could go big and use an air pump. There are huge advantages:
- no need for large fans around the hot end, just one small tube
- if a metal tube is used the air flow can be directed very precisely so there is no risk of cooling down the nozzle as well
- sufficient air flow

So I googled a bit and found that obviously I am not the only one to come up with this idea: [www.tridimake.com]

Has anyone here tried it? Could someone recommend a suitable and quiet air pump?

Or maybe a large 120-140mm fan blowing air into a hose which then brings the air to the printed part?

Edited 1 time(s). Last edit at 01/04/2017 07:51PM by icefire.

---

Self-sourced Mendelmax 2.0-based Reprap Machine -- Ramps 1.4 & Mega 2560 -- DRV8825 (Z@1A, [email protected], [email protected], E@1A) -- genuine E3D v6 direct setup -- 350W custom silicone heated bed -- ABS 1,75mm -- Marlin 1.1.0-RC7 -- Cura 15.04.6

LeadingLights has a nice looking design on Thingiverse: [www.thingiverse.com]

I've read compressor pumps from an old fridge will do fine and they are quiet.

I use aquarium membrane pumps in my DuoCube for months now. Cheap, reliable and not to noisy.
[www.youtube.com]
With this setup i can easily print at 260°C with a PEEK/PTFE hotend (Merlin).

Edit: Ups sorry. I use the compressed air for the cooling of the hotend, not the printed part. The small membrane pumps will not give enough air to do that.

Edited 1 time(s). Last edit at 01/05/2017 03:05AM by Srek.

---

[www.bonkers.de]
[merlin-hotend.de]
[www.hackerspace-ffm.de]

There is also a commercial implementation available called Berd Air.

If you connect anything other than a brushless DC fan to a fan output of your electronics, connect a flyback diode in parallel with it to protect the fan mosfet.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

I have used air pumps for part cooling for several years and think the following is an approximation of the truth (insofar as anything is in 3D printing)
Cooling the filament quickly immediately after it has been laid down improves the fine detail but also makes artifacts more prominent. In order to do this the cool air should be delivered to the print as close to the nozzle as possible. Using one or two air jets seems to leave poorly cooled areas and contriving to have an air ring close to the nozzle goes a long way although it does not eliminate the need for at least some bulk cooling of the printed parts.
An early use with two nozzles can be seen on [www.youtube.com]
The later setup which I use on all of my printers now is shown in cross section below. Note the fan on one side directed on the non-streamlined body of the hot-end jacket.


For bulk cooling I use the exhaust from the heatsink fan. If the fan is close to the outlet and partly obstructed by a non-streamlined object there will be lots of turbulence which will give more even cooling - this has been checked with a model in water with ink to track the flow.
I use an aquarium pump for the air - although I designed the Rootes blower, honesty compels me to admit that the aquarium pump is cheaper, quieter and has a better output.
The use of an air pump, air ring close to the nozzle and small fan for bulk cooling does work well although probably no better than a good standard fan setup - it may give a considerable improvement in a heated enclosure.

Mike

Quote
dc42
There is also a commercial implementation available called Berd Air.

If you connect anything other than a brushless DC fan to a fan output of your electronics, connect a flyback diode in parallel with it to protect the fan mosfet.

Thanks! Berd Air seems an interesting solution but I think that the perforated cooling tubing will get heated by the hot end and thus it will blow hot air. Not necessarily a bad thing but the goal should be to keep the air as cool as possible.


I like leadinglights' solution much more.

What about using a 140mm fan with an air duct blowing air into the tubing instead of a pump? This would make it easier compared to a mains powered aquarium pump.

By the way, what would be the best way to connect an aquarium pump? Probably by using DC-AC SSR?

---

Self-sourced Mendelmax 2.0-based Reprap Machine -- Ramps 1.4 & Mega 2560 -- DRV8825 (Z@1A, [email protected], [email protected], E@1A) -- genuine E3D v6 direct setup -- 350W custom silicone heated bed -- ABS 1,75mm -- Marlin 1.1.0-RC7 -- Cura 15.04.6

Quote
icefire

What about using a 140mm fan with an air duct blowing air into the tubing instead of a pump? This would make it easier compared to a mains powered aquarium pump.

An axial fan is not able to create the kind of pressure you need. Even a stronger radial fan won't help, it will need a pump. The fan will just sit there creating turbulences. Fans are meant to move large volumes at low pressure. Pumps are meant for low volumes at high pressure. The smaller the diameter of a pipe the higher the pressure you need to transport any usefull volume of air.

---

[www.bonkers.de]
[merlin-hotend.de]
[www.hackerspace-ffm.de]

In the process of gathering the materials to test an air pump with a tube ring. Slow going.

But it does bring me to wonder if anybody has done some real research into the cooling requirements. I have seen solutions where there is a virtual tornado going on and versions where there is no more that a trickle of air going. Both solutions were praised by the users as being perfect.
What is the reality? No idea. This is one of the reasons I wanted to test the berd air type solution.

I have seen some wonderful work on CFD to analyse the actual air flow through a printed ring. But never on the actual amount of air or amount of cooling needed.

Lykle

From past experience of thermals for small instruments the usual method involves two engineers, 5 undergraduate engineers, a supercomputer and a stack of books - and the answer will be out by somewhere between 50% and a factor of 10. I think the best way is simply to try it.

The "Berd Air" tube seems to be sensible and easier than my sheet metal solution but I do think the tube should be thicker. 5mm bore K & S tube bent with K & S tube benders is possible - with care you can get a ring about 20mm inside diameter. I would suggest drilling the holes to point down and inwards but angled so as not to point directly at the nozzle - the ideal would be cooling the just laid plastic as soon as possible after it has welded to the already laid plastic.

To avoid radiated heat preheating the air you could insulate the tube with 2mm thick "Kiln Paper" which is a sort of ceramic cardboard that insulates and survives white heat. Silicone tube of 6mm bore with a 1mm wall is somewhat anti-kink and very flexible and suitable for the supply to the tube.

The aquarium air pump that I use has an adjustable flow from 100 to 1000 litres per hour. I think it cost less than £20 in the U.K.

Mike

I tried an aquarium air pump and made a copper ring like the Berd cooler. I used the flyback diode my ramps board handled it fairly well. But the pump wasn't big enough to push much air, and did make a lot of noise.

I think I got it from banggood.


[www.banggood.com]

So bigger than this and probably noisier required.

Edited 2 time(s). Last edit at 01/05/2017 03:36PM by DjDemonD.

---

Simon Khoury

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

I am looking at this turbine blower: [www.thingiverse.com]

Seems pretty simple and easy to try.

---

Self-sourced Mendelmax 2.0-based Reprap Machine -- Ramps 1.4 & Mega 2560 -- DRV8825 (Z@1A, [email protected], [email protected], E@1A) -- genuine E3D v6 direct setup -- 350W custom silicone heated bed -- ABS 1,75mm -- Marlin 1.1.0-RC7 -- Cura 15.04.6

Turbine blowers create a lot more pressure than simple fans, but still less than pumps. It might be enough, though be prepared for some serious noise.

---

[www.bonkers.de]
[merlin-hotend.de]
[www.hackerspace-ffm.de]

One thing to note with ring-type blowers that attempt to blow air at the nozzle from all directions, by virtue of symmetry, the air velocity right at the nozzle tip must be zero. Of course, if the part itself occludes one side of the duct and the other side is unobstructed, you lose that symmetry and you do get air blowing on the critical outer edge of the part. But if you are printing something tall and pointy, you are probably better off with duct that just blows air from one side.

I think there will be a lot of smoke happening soon.
Only way to visualise what is actually happening is use smoke.
So my printer is going to vape?! Oh dear.

Would it help to use a half ring that blows air and another half ring that sucks air? Two independent fans were required, but you'd know which path the air would take.

I will follow Lykle's smoke experiment with considerable interest. I had intended to do the same but haven't even got round to making the ring yet.
With regard to o_lampe's suggestion about the half rings, I think that this would make for one side being constantly away from the incoming air. I think that with cooling, linear predictable flow is less helpful than turbulence - think of how much more effective a wet dog shaking itself is at getting everything wet than for example a hose pipe pointing in a fixed direction.

Mike

Quote
LoboCNC
One thing to note with ring-type blowers that attempt to blow air at the nozzle from all directions, by virtue of symmetry, the air velocity right at the nozzle tip must be zero.

I'm not convinced that think this matters, in fact it may be an advantage. As molten filament comes out of the nozzle, we want it to adhere to the previous layer, so we don't want it to be cooled too quickly. But when it is no longer below the nozzle, it will be in the stream of cooling air, which is what we want - especially when bridging.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

Oh thanks Mike,

Now I will have to try and generate photo's I guess.
Lucky for me, the pump is not here yet and the rest of the stuff is being shipped as well. All I have now is the metal tubing. Why lucky? Because I don't have the time at the moment to be "distracted" by research. Way too busy with the Nimble. But I will do my best to show some real results.

Lykle

Smoke is useless without lasers, if using an air pump you would need to feed it smoke, just blowing it into the path wont show up much, I tried with my latest fan, but smoke was so diffusedn't really see anything interesting, maybe from a vape because they do seem to produce a hell of a lot. But air pumps are an extra noise I wouldnt want for the sake of 2 40mm fans

CPAP machines use turbine type blowers with brushless DC motors. They can deliver a lot more air than necessary and run very quietly if you keep the speed down. CPAP blowers are usually available from American Science and Surplus for about $8-10. All you need to run them is an ESC ($10) and a servo tester ($5) to drive the ESC. See: [vimeo.com]

It's very noisy on the video but that was blowing air like a leaf blower. I use a CPAP machine when I sleep and it is almost inaudible.

Here's an example of that same blower running at much lower speed. It's blowing air into a cooler full of dry ice. A CPAP hose brings the chilled air out of the box. [vimeo.com] It's is still running faster and louder than would be needed to cool prints. More details of that setup here: [vimeo.com]

---

Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]

Quote
the_digital_dentist
CPAP machines use turbine type blowers with brushless DC motors.

Is it something like this [www.sparkfun.com] ? or much stronger?

Completely different. Squirrel cage fans can't produce nearly the combo of pressure and volume that a turbine can. I suspect that the differences are rotation speed and the design of the blades. The CPAP fans are designed for very quiet operation which is primarily a function of the aerodynamics of the blades and their enclosure (brushless DC motors are pretty quiet by themselves). CPAP blowers are used in medical devices, so the quality and the retail price are both quite high. If you look for replacement blowers for a CPAP machine you'll pay $200-300 if you can find them. CPAP machines are commonly discarded by people who can't tolerate the mask/hose/dry mouth. Sometimes you can find them cheaply at garage sales or via Craig's List. I don't think ebay will allow them to be sold because they are considered medical devices.

---

Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]

Quote
the_digital_dentist
. I don't think ebay will allow them to be sold because they are considered medical devices.

that's maybe why I could not find them. Thanks for the extra info.

Quote
Srek

Quote
icefire

What about using a 140mm fan with an air duct blowing air into the tubing instead of a pump? This would make it easier compared to a mains powered aquarium pump.

An axial fan is not able to create the kind of pressure you need. Even a stronger radial fan won't help, it will need a pump. The fan will just sit there creating turbulences. Fans are meant to move large volumes at low pressure. Pumps are meant for low volumes at high pressure. The smaller the diameter of a pipe the higher the pressure you need to transport any usefull volume of air.

Why not turning that big Fan 180° and let him suck like your vacuum cleaner?
That should be much more effectly.

---

Gruß
Siggi

Geeetech I3 Pro B 8mm Acryl - Mega 2560 - Ramps 1.4 - TMC2100 - MK42 - E3DV6-Clone 0,4 - SSR 522-1dd60-40 (A-Senco) - Marlin RC8Bugfix 26.12.2016 - Repetier Host - Windows 7 64 Bit
Vulcanus mit Linearschienen
Anycubic i3 Mega

Quote

Why not turning that big Fan 180° and let him suck like your vacuum cleaner?
That should be much more effectly.

Sadly not the way it works. Suck vs blow makes very little difference to a fan - it can't really tell the difference other than the temperature of the air passing over its bearings. High volume/low delta pressure vs low volume/high delta pressure requires significantly different design.

I have come back to looking at near field cooling and particularly the Berd ring type as this is likely to work better for cooling in a printer with a heated enclosure. As I see it, one of the purposes of cooling is to shorten the time where plastic can flow viscously as this flow can smooth out fine details. To this end, the cooling air should be delivered very close to the nozzle.

The ideal Berd ring has a number of holes but should have a bore sufficiently large to feed these without starving the last holes in the tube. The diameter of the tube should be small enough to fit in the gap between the heater body and the end of the nozzle and the diameter of the ring should be small enough to get the cooling air in close to the newly extruded plastic.

I think my method for making these tubes may be of some interest so I have put down some notes and a photo or two.

The smallest tube that I felt it worth exploring has a 3mm bore and should be adequate for feeding about 8 1mm holes. 4mm OD brass tubing with a 3mm bore made by K & S is commonly available. The same size tube is available in the U.K. from B&Q hardware shops. Picture below shows a ring with 8 1mm holes on an 18mm pitch circle diameter. Also in the picture is 4mm OD K&S tube with a length of 3mm Fry's Powerflow solder in it


Cut a 150mm length af this brass tube and anneal it with a gas torch or blowlamp. Get some 3mm diameter plumbers solder, the kind without a flux core and straighten it by stretching it. You should find that this piece of solder can be pushed cold into the tube and the tube with solder inside can now be bent to quite a small radius by hand. Cut file and tweak the end of the tube so that it forms something like the letter 6. After bending the other end of the tube so that it will meet the air supply tube the solder can be melted out making sure that it is almost all removed or at least that it doesn't block the tube. The end of the tube can now be trimmed closely to where it meets the body of tube and silver soldered. The air holes can be drilled to complete it.

I did try with 5mm OD tube but to fit my hotends it would need to be flattened to 4mm. So far my attempts at making a flattened small radius tube have not been successful, a tear can be seen in the brass where it was stretched too thin.



Mike

Here is the current state of affairs for my current project on this

Cooling with compressed air

---

[www.bonkers.de]
[merlin-hotend.de]
[www.hackerspace-ffm.de]

Clever!

Using solder as a support while bending.Way to go Mike.

I just received the pump, man it is big! Alas, no time to start using it. I want to do this right.
But since the Nimble is now selling, I am running ragged designing additional adaptors left right and center.

But man oh man, what a good idea using the solder. My tubing is 2.9 and managed to get a 90 deg angle bent without problems, but did not want to start on the loop.
And of course, I then realised I need to start with the loop to make it fit the effector I designed for it.

I put that effector up on Thingiverse, in STL and STEP format so that it is easier to modify the model if anybody wants to.

---

Lykle
________________________________________________

Co-creator of the Zesty Nimble, worlds lightest Direct Drive extruder.
[zesty.tech]

Quote
Lykle
Clever!

Using solder as a support while bending.Way to go Mike.

I just received the pump, man it is big! Alas, no time to start using it. I want to do this right.
But since the Nimble is now selling, I am running ragged designing additional adaptors left right and center.

But man oh man, what a good idea using the solder. My tubing is 2.9 and managed to get a 90 deg angle bent without problems, but did not want to start on the loop.
And of course, I then realised I need to start with the loop to make it fit the effector I designed for it.

I put that effector up on Thingiverse, in STL and STEP format so that it is easier to modify the model if anybody wants to.

The solder support worked well if it was inserted cold but the failure of the other one seems to have been caused by feeding the solder in hot and then reheating several times. The solder alloys with the brass and the result is too brittle to take much of a bend. I have managed to do this by using lead instead of solder - photo in [forums.reprap.org] I used copper tube for this but only because I had run out of brass tube.

Just an afterthought, will the gears on your Nimble take a working temperature of 80°C? I guess the body parts would have to be machined out of aluminium or PEEK or something similar but my milling machine is looking for a challenge.

Mike.