I’ve long admired the images of water droplets captured using very high speed flash. I suspect this owes something to the hundreds of hours spent photographing and counting water drops during my last year at university (my research topic was to describe the size distributions generated by various spray nozzles). I have taken a few shots myself in the past, using an eye dropper to make the droplets. But I wanted something more predictable and reliable.
For my birthday at the end of last year I got a kit for a trigger unit from HiViz (details here). Assembly was fun and it worked well. I could get shots like this:
or this:
What I really wanted, however, was more ambitious. I wanted to catch collisions between water drops. Here are some brilliant examples from a real master: Tobias Brauning
For those of you who are new to this, it turns out that the high speed flash part is relatively easy. Most ordinary flash units will, if they have a manual mode and their power can be wound down, produce flash durations measured in microseconds. That’s short enough to freeze almost anything. The tricky part is timing the flash to catch the right action.
When released the droplets pass through a “photogate” sensor (part of the HiViz kitset). By adjusting the delay, I could get the flash to fire when they hit the water. The results were quite repeatable.
When drops hit a liquid surface, an upwards jet is formed which, if it goes high enough, releases a drop or two at its tip. Here’s an example:
The idea is to release a second drop at just the right moment so it collides with the droplet heading upwards.
To do that I built a frame (using recycled rimu from our old villa restoration) in which I could mount a solenoid valve (from Jaycar, here). To control the valve I wandered into the wonderful, open-source world of micro-computers. I bought an Arduino (you can get one for $NZ 25-30 from here) and started to learn how to program it. I made a small control box for the valve, using the circuit described here.
I started with a water glass, then moved to a glass bowl. I really wanted something much bigger, like an old-fashioned developing tray. But I had passed those on ages ago. Eventually I find what I wanted at a local plastics fabricator: a surplus skylight for a campervan. Perfect! They also had some beautiful scraps of translucent plastic sheets of various colours.
Eventually the HiViz trigger will be superseded by the Arduino setup. But I have a few things to sort out first.
One of the problems with this exercise is that it’s quite difficult to see what’s actually going on. It all happens very quickly. I got strange results as I tried to release two droplets, about 100 millseconds apart. Then I had the idea of using the photogate to see what was happening. I went back to dropping one drop at a time, pointed the camera at the photogate sensor and set a very small delay.
Here’s the problem: There’s more than one droplet! Why?
OK, now point the camera at the solenoid valve itself. I could not use the photogate sensor for this, but the kitset came with a simple sound sensor (for bursting balloons, gunshots etc). It worked well taped to the valve, the sound of the valve opening was enough to actuate it. But what it showed was fairly ugly:
That’s meant to be a single drop, not a stream of water!
At this point I must say I got side-tracked. I’d been hooked by the soldering etc required to build the kitset. I realised that for years I had been interested in electronics (I built a kitset Heathkit FM stereo radio back in 1973, long before we had FM in New Zealand) but never really understood what I was doing. It turns out there are some tremendous resources available on the web to remedy that.
For example: MIT actually run an open (free) course on electronics. I’ve sat through the first lecture and managed, even though my maths was a bit rusty. I will see how many I complete before my mathematics runs out. (I asked my son if he would mark my homework – he has a PhD in maths. He didn’t sound very keen.)
Then there’s the practical side. Some people say you should never have to buy electronic components – almost anything can be salvaged from old gear (VCRs, CD/DVD players, etc). My wife asked me to take a load of stuff that we haven’t used in decades down to the local recycling centre. What a wonderful place! I came back with piles of good stuff (much to my wife’s distress). Then I bought a ‘box of bits’ on Trademe for $20. It turned out to be a beer carton full of more good stuff: resistors, capacitors, LEDs, ICs, some unassembled kitsets, plus quite a few items that I had no clue about.
Many days later, with everything sorted, stored and the surplus stuff back on Trademe (sold now), I can turn my attention back to the water drops. The deadline for my next magazine article is a couple of months away, so I have a few weeks free.
So now I will play with the pressure behind the solenoid valve (the ‘head’, which I can adjust very easily) and the time the valve stays open. If I cannot get good results with this valve, I may have to go to a more expensive model. Tobias Brauning uses three valves, so he can get collisions between four droplets (which may have different colours). Not sure I can go to that extreme. Walk before you run.
After listening to me describing my project, a psychotherapist friend of my wife said “you’re playing!”. I conceded that she was quite right. She approved.
I will keep you posted on progress!
