Homebrew Waterfall Display HOWTO

All - I wrote the following up for the Electro Tech Online forums, for a thread discussing such a system; I thought maybe some of you might want to see it. I’m hoping that it may help lead someone to build an inexpensive homebrew version:

Homebrew Waterfall Display HOWTO @ phoenixgarage.org

The thread on the Electro Tech Online forums is here:

Electro Tech Online: Waterfall Printer



I’d start by building around the TPIC6B595 chip.

I’d start by building around the TPIC6B595 chip.

While that would be a good chip to control the system with, I think ultimately one needs to start by first verifying if my valve idea would even work!


The controlling code and hardware can come later; what should be done is to build a simplified singular form of the valve. The idea is to have a valve that operates so that singular drops can be formed, with no pressure (other than gravity) behind the water droplets. Plus, the valve needs to operate at high speeds/cycles.

Such valves are possible to purchase, and I am sure that the videos of these displays use such valves; but they are anything but cheap. So my goal was to come up with a solution that could be implemented cheaply (though not easily).

If a singular valve worked (or if the idea could be used to make a cheap valve that worked), with perhaps it being controlled by an Arduino via a simple transistor switch - it would prove the concept, and that perfected valve could then be used to build the larger device.

Which leads me to wonder if a singular form of the valve wouldn’t perhaps be useful in some manner to those who do water-drop photography…hmm.


how about something using a similar principle to bubble jet printing technology?

how about something using a similar principle to bubble jet printing technology?

Based on what I read here:

I don’t know whether such technology could be adapted to a continuous circulation fountain (which is what one of these displays ultimately are), or whether such technology could be fabricated at a larger scale by a hobbyist, or whether such technology would even work at larger scales.

Furthermore, since inkjet technologies rely on propelling droplets, whereas in the videos of the waterfall displays, you can see only gravity is being used (plus there’s the fact that if the droplet nozzles were pressurized, the change in pressure from opening one or more would have a cascade effect on the amount of pressure behind others falling - which may or may not be difficult to compensate for in software).

But the idea is interesting - right now, no matter what, the goal to achieve this at a “homebrew” level is getting the cost of the valves to a low price per valve; such a system would need at minimum 5 valves to make characters (5x7), but more valves would be wanted to make the kind of effects shown in the videos (probably 64 at least on the low end, 100+ on the high end). If a valve cost $5.00 to make, such a fountain/display would get expensive quickly (though still much cheaper than the shown fountains, most likely).

I think my idea, if it worked, would bring the cost down to less than $5.00 each, possible much less (maybe even $1.00 each). But all options and ideas are worth trying and investigating, because they might lead to cheaper options.

One possibility might be to instead replace the ball bearings I noted with perhaps short lengths of a nail of similar diameter, with the tip of the nail ground or filed to a rounded point; that would entail more work, but might be cheaper than the ball bearings (though the steel bearings from smallparts.com are pretty cheap already).


My initial thought about the actuator is that it still sounds kind of leaky.

My initial thought about the actuator is that it still sounds kind of leaky.

Yeah, it really does, doesn’t it? I tried to find rubber coated steel bearings/balls but had no luck, thinking such a thing could be used to seal it better. Perhaps if the idea of a cutoff nail were used instead of a ball bearing, you could dip the nail into some silicone glue?

Also - I’m not sure if minor occasional drips, at least for a homebrew system would matter in the end. As long as it wasn’t a steady drip, nor a stream, it probably wouldn’t overall matter (think of it as a bit of ‘static’ in the display).

But I am not trying to defend my design; it was the best I could come up with in a very short amount of time. Maybe a two-way magnetic driver, with a second coil around the bottom of the tip, to draw the ball downward to seal the tip? It would be current intensive, but possible…?


As it’s purely gravity how about using an electrostatic force to retain the drops until needed?

  • this being based solely on experiments with a tap (faucet) and a balloon :-[ :-[

I’ve never seen the advert before. Very clever indeed. I cant add anything useful to this but this may be of some note to your/anyone’s efforts.

Using such small bore valves may not work as well as you expect. I’m working on the theory that the viscosity of water remains the same. Scaling down the apparatus will present a higher resistance to the flow of water through the smaller bore of the valves.

So there may be a limit to the minimum bore needed for the water to flow fast enough to give good delineation of patterns. This may be overcome by having a higher volume of water in the ‘trough’ so that the water pressure, caused by the mass, over comes the smaller bore.

I’m not trying to rain on anyone’s parade…you see what I did there ;D I just wonder if it may be a problem. One area that makes me wonder is model boating. One book I read pointed out that a model boat, even if scaled to the nth degree, will still act differently in the water, to it’s full size counterpart, because the waters density remains unchanged it is denser in relation to the model.

It would be a great feat to see though and the solenoid system sounds feasible if the time/distance and water flow can be sorted.



That’s a very good point, Fenrisulfr - something worthwhile to keep in mind…

Too leaky with inkjent…


Hydraulics does not always scale down well. the Reynolds number changes with pipe or tube decreasing diameter and the flow profile can change from laminar flow to turbulent flow, which changes other properties.