Hi guys,
Welcome to part 4 of this process on creating a water drop photography system.
Previous parts are here
This is the interesting one where I actually manage to get some results from my work. I'll drop one of my test images here so that it gives this post a pretty thumbnail!! It's not as detailed as the previous posts - no electronics involved in this one!!
Not bad for a first attempt with just water, food colouring, a background and a bit of photoshop work to do some colour adjustments.
So, onto the juicy details
Hardware
In total, the following cost me approximate £30 UK. That includes some bits that I already had, such as the wood (which is the carcass of an old kitchen unit) and the M5 threaded rod (but this only costs a few pounds for more than enough)
The main structure is made from industrial pipework (which is normally used for decorative purposes). The vertical structure is a 2ft bar with threaded ends and a 6 inch bar with threaded ends. These screw into "floor plates" which are screwed into the wooden base and bottle platform.
The two bars are joined in the middle with a modified connector. I tapped the connector in the middle with an M5 thread and then threaded the M5 bar into this, securing it with a long nut that I had hanging around.
On the other end of the threaded rod, I hang the solenoid which is used for controlling the water drops. This is just a small metal plate which is fastened onto the rods with a couple of threaded bolts I had around. It's pretty sturdy surprisingly. The solenoid is then cable tied onto that metal plate
I wanted to briefly also mention the syphon system. With a normal syphon (or in fact if i had just put a pipe directly from the bottle to the solenoid), the flow rate changes depending on the amount of pressure above the outlet pipe. With a change in pressure comes a change in flow rate which could potentially alter the size and predictability of the drops. In order to get around this you need to provide a constant pressure through the syphon system.
The best way to do this is a Mariotte syphon. I've just used an old squash bottle and a couple of bits of 9mm/6mm pipework and used a glue gun to secure them in the lid.
One problem I did overlook was how to get the syphon started. With the solenoid being normally closed, I had to add an option on the Arduino to open the solenoid for enough time to be able to suck or blow the liquid through the pipes. A 10 second "priming" mode was added and it's all good!!.
I thought it might be interesting to talk about the timings as well as my initial estimates on how fast water drops fall was way off.
In order to create a drop, I open the solenoid for 10 milliseconds. This still gives quite large drops but I don't think the solenoid is quick enough to close back up. There seems to be no difference between a 10ms drop and 20ms drop. I may have to look into another kind of solenoid or a peristaltic pump.
Once the drop has been dropped, with the setup in the images a flash delay of 350ms was about right to get the water drop hitting the water. 5-10ms after the water strike was about the right time to get the drop bouncing back up into the air.
Consistency is a problem. I think with the varying size of water drops due to the solenoid not being precise enough, I seem to get a variation in the region of a few ms either way. Sometimes I'll see the drop just before it hits the water, sometimes just after.
I did attempt a bit of "double drop" work, where the rebounding drop hits another one that is falling. It seems that around 15ms is a good spacing between drops to get something like this (which unfortunately is a bit out of focus). The consistency problem here showed itself - I couldn't get repeatable collisions.
The next thing on the list is to play with different liquids and see how I can improve the lighting. I also need to work out why I am having a weird issue with the camera firing randomly when I touch it - think it must be grounding out somewhere and shorting the shutter optocoupler.
Thanks for reading
Mark