Fluidic diode:trials and tribulations

Story, significance,

This story starts on wikipedia, where I discovered the Tesla valve, a one way valve for fluids with no moving parts.  I was so impressed that I designed a part in OpenSCAD and uploaded it to Thingiverse, here.  Unfortunatly, it didn’t work very well. I tried three different designs, all of which only kind of worked. So, I went back to the drawing board and kept designing.

Tesla valve

Printing my Tesla valve

But before I tell you about the designs, let me point out why this is important. Fluidics is the art of using fluids to do computation and control in the same way that electrical computers compute, but using a fluid instead. The Tesla valve is an example of what a Fluidicist, just made that up, would call a fluidic diode. You can also make the fluidic equivalent of transistor(fluidic amplifier), capacitor, resistors, inductors and more. Home 3d printers print in plastics which limits what they can do. In particular, they can’t print an electric stepper motor, one of the main components of an extrusion based 3d printer.  BUT extrusion 3d printers are awesome at printing solid parts like gears and pipes and such. With such elements, fluid stepper motors can be made along with fairly complex logic to control and drive it. In a way, fluidics could be a killer application of 3d printers because complicated ducts can be printed at no extra cost. In essence, you could print out complicated fluidic microchips at home. I should also point out that I’m not suggesting some untested crack pot wild scheme. Fluidics is a very well developed technology with a history that dates back to the 1950s and with applications that range from aerospace to biology.  For the more technical reader, I’ve collected some useful links at the bottom of the page.

Now back to the story.  So the Tesla valve wasn’t working as well as I wanted it to. I thought I’d do away with the “no mechanical parts” paradigm and include a flap that closes if air goes in one direction and opens when air goes in the other. This worked much better but still only reduced the air flow, by what felt like 30% or so, I’m guessing here. If the material was softer, I’m using PLA, I’m sure this would work much better. You can download and print it here.

One way valve

One way valve with flap:side view.

One way valve

One way valve with flap:bottom view.

One way valve

One way valve with flap:top view.

But then through my reading and on Thingiverse, I found out about the fluidic vortex diode. Sounds cool, huh?  It works the same way that draining a bath tub works. Huh? When you drain a bath tub, the water has  some angular moment and creates a whirl pool as it goes toward the drain. This whirl pool limits the speed at which water can drain. But going the opposite way, pumping from the drain, there is no whirl pool and so less resistance.   A fluidic diode works like this, only the angular momentum is maximized by pumping the fluid in perpendicular to the sink and by putting the whole contraption in a closed cavity.

I made two designs. The first one you can download here. It also only kind of worked. So I went to the drawing board and did some math.  It turns out that a spiraling fluid with no friction experiences a pseudo-potential barrier that varies inversely with the radius squared, meaning that if you double the radius of the vortex diode, it should take four times as much energy to overcome that centrifugal barrier.  So I designed a vortex diode with over twice the radius, which you can find here. It also didn’t quite work as well as expected. So now I’m calling out to the community, YOU, can you help me find a better fluidic diode?

Large fluidic vortex diode

Large fluidic vortex diode

Fluidic vortex diode

Fluidic vortex diode

Links and further reading:


1967 report by the military summarizing fluidics

Fluidic stepper motor


Lecture at the Broad Institute about microfluidics


  1. powertomato

    First of all you _can_ print a stepper-motor: http://www.youtube.com/watch?v=paT63-8DLbs 🙂

    Now to the fluidic diode part, this is just a guess, and I’m risking talking total bullshit but:

    If you apply reverse voltage with no resistor to an electrical diode, considering zero-ohm conductors, the dominant resistance would indeed be the reverse resistivity of the diode (which should ideally be infinite but this is not the case, even with electrical diodes). In your case this resistance is too low so the dominant resistance is the resistivity of material (in fluidics that would be the size of the holes). If you make them bigger you may reach a point where the resistance caused by your diode becomes dominant, but still will be too low: i.e you will notice asymmetric current limiting when blowing. Beside of lowering the material resistivity you should also try to increase the backward resistance, which you’ve tried by doubling the radius, but it still doesn’t seem to be enough. My idea here would be: combine the two: create a spiral-shaped tesla valve and place it around the opening of the vertex valve. But I think without moving parts it will never or (only with much work) reach your expectations.

    The moving part doesn’t have to be anything complex. You could just recreate a balloon valve. Those things basically just use a small circular plastic pad. Something like this: http://imageshack.us/f/849/foow.png/
    Green is the position when the flow is in forward direction, red the position in backward direction. I think that would be a much better approach as it would work more “digital” whereas a tesla or vortex valve will always have a little current in reverse direction. Unless you’re interested in this effects for current or pressure regulation (analogous to electrical circuits which use zener-diodes)

    • independentindustries

      You make a really good point that the resistance of the tube has to be small compared to the resistance of the vortex diode. I’ll have to think about how to do that.

      The stepper motor also looks pretty awesome but the nice thing about fluidics is that you could print the driver and logic out of plastic.True though, it’s not that hard to find magnets and copper wire.

      As for the valve with the flap, I’ve already tried it here:http://www.thingiverse.com/thing:96479. Now, you’re probably suggesting that I use a different material for the flap. If I were making just one diode,that would work great but I’m interested in potentially printing out thousands of these at a time, an automated process, so I’d like to avoid any manual labor.

      I like your idea about combining the ideas of using a vortex diode with the Tesla valve. I’ll have to think about this a little more. I think the answer will be subtle, depending on a good knowledge of fluid dynamics. In the end I need a device for which as a I change a parameter, like the radius, the resistance in the backwards direction grows more quickly than the resistance in the forward direction. These are non-linear elements so I don’t think resistances add in series. I’ll probably blog on this soon. The math speaks for itself.

  2. Nate Burley

    I think PT is right, the issue with most of the valves with no moving parts, at least the ones I’ve seen, is that they’re not 100% effective in the reverse direction. If you want to obstruct pulse flow say in a dosing pump or something similar they’ll be great but in terms of a back stop in positive displacement environment, the ones I’ve seen don’t really cut it. Three options then…

    1) Try enhancing the Tesla valve by wrapping it in 3 dimaensions. This was never possible before 3D printing, might have to get it done in SLS or SLA though. There’s one on Thingiverse here: http://www.thingiverse.com/thing:15511 very interested in how well this would work!

    2) The success of your flapped valve might not be down to the design as much as it is down to the material. Shapeways have now brought out a flexible Elastomeric material which I’m going to start looking at using for similar devices. Try this perhaps: http://www.shapeways.com/materials/elasto-plastic

    3) Perhaps the idea of using 2 different materials in the model wouldn’t be such a bad idea IF you didn’t have to do any manual assembly. Like, say, IF the 3D printer were capable of printing in a solid plastic AND an elastomeric material as well? http://www.youtube.com/watch?v=pbjcfplk8Ig

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s