If you’ve ever tried to capture splash shots of objects dropping into a glass tank of water you’ll know that catching the action at the right moment is very hit-or-miss (mostly miss in my case). While it’s something that you probably get better at with practice the hit ratio is still likely to be fairly low. This means lots of mopping up and tank-glass cleaning after shots where nothing useful was captured.
To increase my proportion of usable shots I bought a Triggertrap Arduino shield a few years back (when it was still available). In use I found it was a pain to set up, particularly getting the laser and sensor aligned properly. In fact I enjoyed using it so little it’s been sitting in a cupboard gathering dust ever since my first couple of sessions with it. I didn’t have much joy with its sound triggering capabilities either.
Fast forward to the present and I decided it’s time to set up more splash shots and do it in a more controlled fashion. There’s a small glass fish tank (formerly the residence of my younger son’s goldfish, which has gone to a better place) sitting on a shelf in the garage, and I’ve got drawers full of Arduinos and opto-isolators. And with a bit of luck a properly designed photogate camera/flash trigger can be used for more than just splash shots …
- Laser and sensor should be easy to set up and align, so they need to be packaged in a way that’s easily attachable to whatever they have to be mounted on,
- Laser module and sensor should both be remote from the control box (ie. plugged in via cables) unlike with the Triggertrap,
- There should be a user adjustable delay between beam break and camera/flash triggering,
- There should be two modes of operation:
- Camera mode: mostly for use in daylight conditions. The camera is triggered after a delay from beam break. If flash is going to be used it will be triggered by the camera.
- Flash mode: for use in dark conditions. The camera (on bulb setting) is triggered when the controller is armed, flash is triggered after delay from beam break.
Parts needed to put this together are:
- Arduino to do all the clever stuff – I used an Arduino Uno for this build,
- Rotary encoder module to set the camera/flash delay,
- Nokia 5110 LCD display to show delay value and system status,
- Push button switch (single pole) to set the mode,
- Push button switch (single pole) to arm the system,
- Arduino laser module (beam source),
- BPW34 photodiode (beam sensor),
- A few resistors: 6 x 10kΩ, 1 x 1 kΩ, 2 x 390Ω
- Two 4N35 optoisloators to trigger camera and flash,
- Battery clip to supply the power: I used a 6xAA clip, which gives 7.2V with NiMH rechargeable or 9V with regular alkaline batteries,
- All the usual finishing bits (project box, sockets, on/off switch, foam tape, etc.),
- Very small project box (x2) to hold the laser and sensor.
The Nokia 5110 LCD is a great little display if small and monochrome is what you’re looking for, but it’s a 3.3V unit so for most Arduino boards you need to add series resistors to ensure the 5V digital outputs don’t cause any damage. A 1kΩ does the job for the SCE pin, and 10kΩ for RST, D/C, DN and SCLK. The Uno has a 3.3V power output, so there’s no need for any additional regulation circuitry to power the display.
For this project I used a rotary encoder for setting the delay, as these are generally a lot quicker to operate than push buttons: just rotate the knob clockwise to increase the delay and anticlockwise to reduce it. These encoders often have a built in push switch, so I’ve used this to reset the delay to the default value.
The Mode and Arm pushbutton switches are connected between ground and the digital (or analogue) input pins: configuring the pins as INPUT_PULLUP enables the Arduino’s internal pull-up resistors, so no external pull-ups are needed.
For the output side there are two 4N35 optoisolators for the camera and flash triggers, both with inline 390Ω resistors (see PIR Trigger article for the explanation of how the resistance value is calculated if it’s of interest).
Laser and Sensor
Using a laser module like the one listed above means you only need to supply 5 volts and ground. Both of these are available on the Arduino board, so all it takes is two wires to get it connected. The module is very small and is mounted on a pcb, so is easy to box up for mounting as per the first requirement above. That’s the light transmission side sorted.
For the light sensing side it’s only slightly more complicated: the cathode of the photodiode is connected to 5V and the anode to the Arduino’s input with a 10kΩ pull-down resistor to ground. Again small and easy to box up (more about that later).
The breadboarded test version of the circuit looks something like this:
The complete circuit diagram is shown below:
The reason for this is that both of these inputs are relatively timing-critical, so if the software is polling the encoder and sensor in a loop where most of the time is being spent doing something else (like updating the LCD display, which is a relatively time consuming process) it might mistime or completely miss an important event.
The way a rotary encoder works (a description can be found here) this can cause it to decrement when it’s meant to be incrementing, which can get frustrating. With the beam break detection it’s liable to lead to inconsistency in the results, including completely missing beam breaks. Luckily the Arduino Uno has two external interrupts which is enough for this application (some other models of Arduino have more).
Arduino External Interrupts
On the Arduino Uno pins 2 and 3 can be used to generate interrupts. These can be set to trigger in four modes: on a rising edge; on a falling edge; on any change, and on a low level on the interrupt pin. In the Arduino sketch, interrupts are set up in the “setup()” routine using the following statements:
attachInterrupt(digitalPinToInterrupt(CLKpin), isr, LOW);
attachInterrupt(digitalPinToInterrupt(SPDpin), isr2, FALLING);
The first of these statements is for the rotary encoder and makes input pin 2 (CLKpin) cause an interrupt when its state is low. When the code is in the main loop, a low signal on pin 2 will trigger the interrupt and execution will jump to the interrupt service routine “isr”. It returns to the main loop when the code in “isr” has been executed.
The second statement makes input pin 3 (SPDpin) cause an interrupt on a falling edge (high to low transition), which is what happens when the beam is broken. This time interrupt service routine “isr2” is called (as you can see, I got very creative when I named the interrupt service routines!).
Interrupt Service Routines
The interrupt service routines are short, as they pass back global flags to the main loop rather than doing anything too time consuming inside the interrupt handler itself. The variables used to pass data between the interrupt service routines and the main code body have to be declared as ‘volatile’. This is so they don’t get corrupted when execution switches between the interrupt service routines and the main code body.
Apart from the introduction of interrupts the software is pretty simple: the usual setup routine and the main loop to poll the non time-critical inputs and debounce the switch inputs, process any updates passed back from the interrupt service routines, fire the camera and flash triggers and (finally) keep the display updated.
A zipped copy of my Arduino sketch can be downloaded here. This has been updated to set the display’s contrast level in the setup routine. Nokia 5110 displays vary so there’s probably no contrast value that will work for all of them. To find the best contrast value for a particular display, the Arduino sketch in this post should help.
Since this is a piece of gear that has the potential to be used outside the studio, it’s worth building it into a reasonably robust box and powering it with batteries. Mine uses a 150mm x 80mm x 60mm project box, which is big enough to fit the Uno and battery clip and gives plenty of space on the front panel for display, switches, rotary encoder and connectors.
For the laser module and photodiode connections I used phono sockets to make them different from the camera and flash trigger, which use regular 2.5mm jack sockets. Aside from what’s already on the Arduino the only additional components are the resistors and optoisolators. Most of these are mounted on a small piece of strip board that has header pins attached so that it can be plugged into the Arduino like a shield board. To save a few wires between the front panel and the circuit board, the three pull-down resistors are attached directly to the switches and sensor socket.
Regarding the front panel, as well as labeling the controls and sockets it needed something to cover up the scuffs from drilling the holes. This time rather than the laminated paper panel I made for the drop controller I decided to print out the label template on glossy self-adhesive printer paper. To protect it and avoid the labels getting smudged I stuck a piece of clear self-adhesive Fablon sheet on the top. The result doesn’t look too bad, but to be honest I don’t think I’m yet close to reaching my optimum front panel labeling solution.
The finished box looks like this:
Boxing the laser module and photodiode
The aim here was to enclose them in small, lightweight rectangular units that would be easy to clamp or otherwise attach to any support or frame round the tank, or anywhere else I want a beam trigger. My solution was a pair of 26 x 18 x 13 mm sub miniature ABS plastic boxes. These are pretty small, and the laser module needed to have the three pin header de-soldered, followed by a bit of deft Dremel work to thin it down and remove the corners so it would fit around the lid screw post mouldings in the corners of the box.
Housing the sensor was a lot easier: the photodiode just had to be mounted on a small piece of veroboard to fix it in place. The completed units meet my original spec, and don’t look too bad either – this is them beside the trigger box:
Plugging it all together and trying it out for the first time it seemed like a sensible idea to do a dry test. For this I used the classic ‘two soup can’ setup (it works with bean cans as well!). The laser and sensor were taped to the top of the cans and a ball dropped between them to break the beam. Here’s what I got in the dark using flash mode, with a 0ms delay for the first shot and 10ms for the second:
Encouraged by the results of the dry test I decided to get out the fish tank and try some splash shots.
Allowing that lots of splashes were going to be made and things could get a bit wet, I laid a towel under the fish tank before it was filled. To keep the laser and sensor in place I made up a simple frame out of offcuts of wood with clamps at each end to sit on top of the tank. Both laser and sensor were wrapped in clear food wrap to keep them dry.
A boom arm was set up as a reference for the drop point to ensure some level of consistency, as the camera framing was going to be quite tight. I used my Genesis AS250 studio heads as they have IGBT circuitry and can give a very short (about 1/8000 second) flash on low power. The flash heads were set up either side of the tank, shooting through strip boxes.
The pictures below show a wide shot of the setup, together with one of the resultant splash shot. Again this was shot in the dark in flash mode. Height from drop point to water surface was roughly 300mm and the delay set to 110ms:
Note: the above was a ‘quick and dirty’ test, without cleaning the splashes off the glass between shots, or getting the lime released from exactly the same height each time.
If you’ve gone through the process of building a unit as described above you won’t have any problem getting it all plugged together correctly. However, while it makes sense to me, the way the software works may not be entirely obvious to everybody so here are a few pointers:
- The unit always powers up in “Camera” mode. That’s the mode for light conditions, so you set the camera’s shutter speed to something that works for the ambient light and the shutter will be triggered after the set delay when the beam is broken.
- Press the “Mode” button to toggle between “Camera” and “Flash” modes. “Flash” mode should be used in dark conditions with the camera’s shutter set to “Bulb”. With this mode, as soon as the system is armed the camera’s shutter is triggered, with the flash triggered after the required delay from when the beam is broken.
- Adjust the delay using the rotary encoder. The default value is 10ms – pushing the rotary decoder’s built in switch resets the delay to the default.
- Arm the unit by pushing the “Arm” switch – display will show “Armed: 1”. Unit will trigger when the beam is broken, at which point the unit again becomes unarmed. Pressing the “Armed” button when the unit is armed will also disarm it.
That just about covers everything I think, but feel free to ask if I’ve missed anything.