banner perso

Some links about me. Many of my 3D designs are free. I also post on Google+ and in another blog, oz4.us

Wednesday, May 18, 2016

Cooling 3D prints with an aquarium pump

3D printing cooled with an aquarium pump!
It works well and it cannot be quieter. The pump is bulky,
but the overall is a very compact hot end.
(the above was a preliminary experiment)
Interestingly, while this trick is not new, very few people seem to use it. For years I had an air pump on a shelf for this purpose. Only recently did I use it, and I really like it better than fan on the hot end.

The good thing is that aquarium pumps do build some pressure in order to pump air down into water. No regular radial or axial fan can do this, they only move air around. Do not get me wrong: regular fans works well, but the smaller the fan the noisier and the shorter its lifespan. And they are bulky anyhow and obstruct the view.

Actually, I printed almost years without fans attached to the hot end at all. Instead, I used a large, powerful and silent 120mm PC fan on the printer chassis. It blows a lot of air on the whole part, which is often even better, but it fails to cool down tiny towers or islands when the hot end never move aside (flowing air is very lazy and will not reach tortuous parts). Also, the big fan blows air only from one direction, which shows up on opposite sides.

So I still needed air again on the head....



By the way, cooling is an absolute must for long bridges, and even though I always try to design my parts to avoid bridges, there are times where it is suboptimal without local cooling.
3D printing in air. One key point for efficient bridging
is to have fast and local cooling (by Erik Cederberg).
In real life, the part should better be printed upside down!

I had bought myself a costly Whisper AP150 for this purpose, which is a branded, solid and extremely quiet pump made for aquariums. Sadly, it only runs 110VAC and we are 220VAC here... So I hacked it with a simple power diode (i.e. it gets half waves of 220VAC)... Make sure to understand though: the signal is no more sinusoidal and it would probably break many devices. However I checked that this pump had no sensitive components (see at the end of the post). And it does get hot even on long prints. Now, the efficiency is probably halved, or worse...

My aquarium pump brings ambient air via a latex tube,
down on the freshly deposited plastic. It works well!

As you can see in the video, the significant pressure I get from these pumps allows me to use a very soft latex tube, as it will stay open up to the head: no worry that it gets pinched by its own weight when the head moves around when air gets pumped into it. Also since it is very flexible it will obviously not restrict the movement of the head (regular thick plastic aquarium tubing would be very wrong here).

How did I make the copper nozzle?

I used copper because I had no other suitable metallic tube. The best choice would be very thin stainless steel tubing, as it would not suck heat from the nearby hot end like copper. Reciprocally, copper is an easy material to shape as you can see below.



Copper tubes are easily bent (even more when heated).
Better insert something so the tube does not flatten too much,
some even fill the tube with sand before bending it.

I used this loop of thick stainless steel wire
to keep the mouth wide open

And a secondary piece. It is important so that
I can crush the middle later without closing the two gaps.

Heating it makes it softer, which makes it easy to deform at will.

Using a hammer to widen the opening.

And closing the middle section. I do not want to cool the nozzle
itself, but only to have powerful flows of fresh air around it.

This is how it looks before I remove the wire loop.

And the almost final mouth, with the two lateral openings.

Once polished.
Latex tubes are cheap, very flexible and easy to connect.
They would collapse without air pressure (e.g. regular fans).

The end result, which is screwed directly between to fins of the heat sink hot end.
Nylon washers keeps the tube airtight. Using two screws would prevent the tube from
swiveling when the bottom hits unwanted plastic blobs. Also, that end of the tube
shall not get in contact with the nozzle or it will act as a heat sink!
I also ended up filing the mouths so the powerful flows are less horizontal

An action shot! Printing CFPLA at 210°, 75mm/s, 25 mm layer height and 0.4mm nozzle.
The setup works very well, especially for trickier prints than this, with little towering islands..

Here is another setup, which shows the support John O'Shaughnessy made for his tube. It screws conveniently on the fan of his delta printer. By the way, he used an old trick: you can fill the tube with salt to avoid pinching it when you bend it. I found that heating it gives the same result. In any case, pinching the output is even a good thing to achieve a venturi/blowing effect.




By the way, what is inside an air pump? How does it work?

I opened it, you bet! And it is deceptively simple. There is only a transformer, which is opened on one side. It acts as an alternating electro-magnet which push-pulls two levers. The latter move only by a few millimeters to fill and empty rubber pouches on two valves. Since they runs at 50Hz or 60Hz according to your country it represents a lot of air -- where 60Hz will give you a slightly better throughput  ;)
Inside of an aquarium air pump (Whisper AP150).
This is only an electromagnet which relies on the alternative mains!

Also, some make their own pumps of course. Thy most often rely on regular motors, which are probably noisier and less efficient or durable than the industrial pumps, but they can be made to work very well. And, they are homemade, which does not mean it is crappy :)
Home made air pump with a regular motor, very nice howto and result.

So what is next?

A $12, dual outlet air pump,
1.5L/min (verified myself).
Since it is working better than I expected, I bought the dual outlet pump on the right (so cheap: $12 shipped on ebay). I was first disappointed by the throughput until I noticed the little magic button on the back! Switching it doubled the throughput ... and also probably doubled the noise (is still OK though). I checked it by myself and emptied a 1.5L bottle in about 50 seconds with one outlet. This is certainly not the claimed 4L/min but it is still  decent compared to my previous "real brand" one.

The pressure seems even higher, so it may be better for very focused air cooling (I am unsure is it worth though).

I also need to add either a switch to stop or start the air flow (update: check this excellent maker), or, better a PWM regulation to set the amount of air as with the default fans.

Todo: less weight, wider mouth.
Also, I wanted to give it a try: not only to cool the immediately deposited thread of plastic, but also to try and cool down the hot end itself without having to revert to water cooling. This way I would get rid of the annoyingly noisy and failure-prone but nonetheless compulsory little fan on the hot end heat sink (more on that here by the way).

Finally, my existing tube is uselessly long (hence uselessly heavy), and the mouth could be wider to cool a larger portion, to trade pressure with surface. So my next iteration will probably be based on a shorter and thinner tube, like the one on the left.





About Me

My photo

If you know me and you cannot tell exactly what my real job is, then you probably found the right Jeremie. Check zax.fr for some pointers.

I am self-employed and I help start-ups, research centers, small companies with their needs related to computers, sensors, data processing and mechatronics. If you have a project and know what "R&D" is, then you already sparked my interest ;)