Slave Flash Controller
Better Pictures for a few dollars
I needed to take better pictures for our website. I needed better lighting, and a bunch of halogen shop lights just doesn't give the right color to things. And it's a fire hazard, not to mention an obstacle course. Here’s an inspiration- surplus flash units are real cheap from surplus electronic distributors, anywhere from $0.79 to $4.55 depending upon the source. All I needed was a trigger- a photodiode, a few discretes and an SCR. What could be simpler for a quick and useful project to post? . . . (How many times have I fallen for this?)
By detecting the camera’s master flash with a photodiode, and triggering the surplus unit with an SCR, I should be able to put together five or six slave flash units for under $20! Well, things got a bit more complicated . . .
Now for the 99% perspirations . . .
The first prototype was too complicated, so after some design reduction, I put together two slave flash units, then did the testing, documenting, and photographing.
Finally, I took a quick picture of the 'flash' as a final test and for the article. Big surprise- they flashed, but nothing showed up in the picture. They didn't work with my digital camera. My eye had seen the flash, but the camera didn't. I had thought the problem was something like the airbag that opens up after the crash. So, a lot of (and eventually useful) effort went into simplification to reduce latency. I did make the circuit faster and simpler, but it still didn’t work.
Early experiments all resulted in pictures like this (the flash tube is covered):
"When all else fails, read the directions. . ."
So I did a quick search on the internet and found this from www.srelectronics.com:
"Ordinary slave units will not work with most digital cameras because these cameras use a very rapid series of pre-flashes (we're not talking about red-eye reduction). The pre-flashes are used to set the white balance of the camera's image sensor chip - not the exposure. A typical slave unit will fire on the pre-flash it senses while the digital camera captures the image on the last flash. Thus, the extra light from the slave does not show up in the digital camera photo."
Back to the bench, and a few experiments later using a photodiode and my camera, the oscilloscope showed that, indeed, two flashes were occurring- the second one 140ms later.
A new paradigm
The next version used a CD4013B dual D-flipflop to count the first flash’s pulse, then triggered the SCR after detecting the second flash. It worked! But not on the second try; the slave flashed, but the camera didn’t see it. But then it did on the third flash! But not on the fourth, and it was fairly regular in alternating back and forth- Odd flashes showed on the camera, even ones were missed. Checking the circuit for some sort of indeterminate logic states and finding none, I figured there must be something else going on with the camera. Maybe there weren’t just two flashes, but extras that the camera used for setup.
The fourth version configured the CD4013B to provide a delay after the first flash instead of counting them. It even used most of the same parts, just configured differently. This unit finally worked correctly! The slave flash showed up consistently to the camera.
Most slave flash controller designs have some amplification for the photodiode pulse, or use a phototransistor for added sensitivity. Tests showed that the strobe’s flash is definitely bright enough to generate a rail to rail pulse directly from a filtered (black) type photodiode. Using a filtered diode also helps eliminate false triggers and the need to capacitively couple the signal as there are few sources of light other than a camera flash that could activate the diode. The high input impedance of the CMOS flip-flop allows the photodiode to be directly connected, and because of its output drive capability it can directly drive a sensitive gate SCR, so no additional drive circuitry is needed to trigger the flash.
The camera’s flash causes thephotodiode to clock both D flip-flops, latching their input data. The ‘D1’ input is tied high, so its Q1s output will be clocked high. The ‘D2’ input will still be low due to capacitor C2, so its Q2 output will stay low. It will take approximately 100ms for R8 to charge C2 to the point where ‘D2’s input goes high. During this time, additional flashes will not be able to trigger the SCR. When ‘D2’ finally does go high, the camera’s main flash occurs and sets ‘Q2’ high triggering, the SCR and firing the slave flash. At the same time R6 starts charging C1, and R5 starts charging C3. C1 will charge up to the point where it will reset U1A before C3 charges to the point where it resets U1B. This longer delay for U1B assures that U1A will be reset. (Tests with both reset lines tied together sometimes didn’t reset U1A depending upon the IC used. There are probably variations in the trip points due to process variations from chip to chip.) LED D1 is a Power-On indicator and a state indicator. If there is some sort of ‘glitch’ when the controller is turned on, D1 will be off. The usual user response is to flip the switch on and off a few times. This should reset things to their correct state.
When the unit is finally switched off, U1B’s 'S'et line is pulled high by the residual charge in C4, triggering the flash unit and discharging the energy storage capacitor. This could also be done with an extra external switch, but that’s one more part to install.
Because the circuit is so simple, it can be easily constructed on a small pad-per-hole/prototype PC board. A suggested layout, as well as the schematic, can be found on the Download page.
The controller and the flash unit use separate batteries for two reasons:
- The polarity of the high voltage drive on the flash unit is setup with a positive ground. Making the grounds common would cause all sorts of trouble. This isolation also makes it more easily adapted to other flash units.
- When the flash unit’s capacitor first starts to charge, the driver draws 1 amp from its design specified ‘AA’ battery, and almost 1.25 amps from a ‘C’ cell. By using a ‘C’ or even a ‘D’ cell, all the slave units can be powered from one source. The controller only draws a few milliamps, so it will run for quite a long while on its three ‘AAA’ cells.
A quick test picture shows how well this works. I pointed the slave flash at a white sheet hanging behind and above the subject, and partly re-directed the cameras flash with a small piece of aluminum foil taped below it- no glaring high-lights, and minimum shadows.
Other units can be used as well. (Rescue one from a disposable camera.) You just need to sort out a few details:
- Positive and negative battery connections
- Trigger contacts
- Trigger contact polarity
The trigger polarity can be simply found by partly charging the flash capacitor, the checking the contacts with a volt meter- Negative to the controllers ground, positive to the controllers SCR anode.
“Where there are no oxen, the stable is clean.
But with the ox comes much increase...” Pr 4:14
Taking Ideas into Reality
Edison was very insightful in his definition of genius. There are lots of good ideas and inspirations, but all by themselves they have little value. It takes effort and insight, tools, materials, and lots of time to bring them into solid existence. Applied Inspirations, LLC was founded to assist in bringing new ideas into physical reality by providing:
- Examples - Tutorials and projects that demonstrate ideas taken from concept to reality.
- Tools - PC boards, software, hardware and kits to explore turning new ideas into products.
- Services - For those without the skills, tools or time, we provide services to develop part of, or the whole idea.