viernes, 7 de enero de 2011

High speed flash (part 4)

For this picture I needed an extra feature on my circuit. I needed to be able to set up a delay.
Even though I'm getting a really fast response from my circuit, it's not absolutely simultaneous. Besides that, I've discovered that pictures involving splashes can change dramatically in less than 1 millisecond, so I wanted to take control of that.
I configured the PIC so I was able to display the delay on an LCD screen, and increase it or decrease it with two buttons. This way I could also place the laser above the impact area, then add around 10 milliseconds to wait for the marble to touch the water.

I learnt that flashes control their output by the length of the burst, not the power (brightness) of the light. This means flashes have (ideally) constant output. A 1/2 burst is half the length of a 1/1. This is not exactly true, since flashes (like any other electric device) have start up and cool down delays.
In the end, the effect is the same: longer equals brighter. The speed of the flash burst is usually irrelevant, but when doing high speed photography, long flash bursts mean blurry pictures. Logically, more powerful flashes will allow you to get the same amount of light in a shorter amount of time.

I found some really detailed information about the Canon Speedlite 580EX flash durations. It's worth reading.

According to this page, the duration of the 1/1 flash burst of a 580EX is 1/833 seconds. Back to the example I used before about waving your hand at 1 meter per second (which is not fast at all), you would be moving your hand 0,12 cm in that period of time, which will probably result in an blurry picture.
At power 1/128, the duration of a burst is about 1/35.000 seconds. In that period of time, your hand would move 0,029 mm, which is hardly noticeable.

The picture above was taken with a 480EX II at its lowest setting (1/64).

Below is the code for the PIC with the LCD display. I'm sorry I can't post the code for the LCD, since it's copyrighted:

jueves, 6 de enero de 2011

High speed flash (part 3)

Besides capturing the moment of impact, as in my previous picture, I thought that I could also shoot exactly when the subject is in focus, which is specially difficult when the depth of field is small and the subject is moving.

In this picture I wanted to use a macro lens at the highest aperture in order to leave my hand (and maybe the second dice) out of focus.
I set up the laser and the sensor circuit using a mirror, mostly because I don't have much space on my desk:

I painted the laser beam on the picture. By the way, the laser is actually an infrared thermometer, though I'm only using it as a laser pointer. I'm using one clamp to hold the laser button down so it's on all the time, and another to hold it in place. The circuit is out of the picture, on the left side, behind the camera.

So I put the dice blocking the laser and prefocus the camera on them. Obviously the camera has to be in MF mode. Manual mode, 100 ISO, f/2.8 and 0.5 seconds exposure so I have time to throw the dice. 2 sec. delay mode helped since I was holding the flash with one hand and throwing the dice with the other. The long exposure is the reason why there is a red light in the picture. I'm not sure if it's light hitting the dice before the flash or a lens flare coming from a reflection.


Continue reading in part 3.

High speed flash (part 2)

Continued from Part I.

I thought that with light I could anticipate the moment of impact and take the picture in time. I went to the Don Electrón store again, where I had bought the relay and the microphone, and I explained to Juan (the shopkeeper) what I wanted to do. "I need it to be fast" -I said. He's a really nice guy, always giving advice and great explanations to his customers (almost any college student from my university that has ever had to build an electronic circuit knows him).
He just said "you need an LDR". After showing me how an LDR works (he took out a meter and a battery), he drew a little circuit for me and told me that it was probably the fastest circuit I could build.
I first tried building it with only one resistor, because I thought I didn't need the second one, but that didn't work, so I built the circuit the way the guy told me. I played with the resistors to make the LDR sensitive enough to see my laser, but not be bothered by ambient light.
It sure was the fastest circuit I could build, but it was too fast for the relay! Holding my hand to block the laser would trigger the relay (and the flash), but waving fast it didn't do anything. I needed 5V delivered to the relay for at least 0.5 milliseconds, and stay there enough time so that the flash reacts. So anything that interrupted the laser under 1 or 2 ms didn't do anything.

I had in mind putting a PIC processor in the circuit later, to insert programmable delays, but I thought I could also use it to detect the LDR output. I set my PIC to detect really small voltage changes in analog mode (again, digital didn't do anything, since I wasn't generating a real complete pulse).

Click on the above diagram to see it bigger. I found the hotshoe diagram here. Below is a simple code for the PIC to get this thing to work:


By the way, I found how to display code in this blog here.

Continue reading in part 3.

sábado, 1 de enero de 2011

High speed flash (part 1)

I always wanted to make a device that would allow me to take a picture at an exact moment, but I wasn't sure how to do it.

A couple weeks ago I decided I would give it a try. First I bought a little electret microphone kit for 11€ (the one on the left). A bit expensive, but probably cheaper than buying the components separately. I also bought a 5V relay to trigger the camera. I built a simple circuit that would take the amplified microphone output and feed it to a transistor that would trigger a relay.
I'm sure it's not a good design but it worked:


It was cool to see it working, but it wasn't fast enough! First of all, I'm taking the picture exactly after getting a sound, so I'm always getting a picture after the impact. I had everything on manual, so the camera would shoot as soon as possible. This is a picture of my hand knocking on the table:

I had enough time to get my hand out of the picture! I did a little research, and discovered that even if I'm doing everything on manual, the camera always has a long delay.
For the Canon 500D, there's a delay of about 110 milliseconds, which adds to the delay of the circuit itself. Removing the relay didn't do much, since it only takes about half a millisecond to change its state.

110 milliseconds is a very long time at high speeds. Though you can probably do much faster, let's say you can wave your hand at 1 meter per second. In 100 milliseconds you're moving 10cm!

I decided to trigger the flash instead of the camera. Still, I was getting some delay. Maybe around 20 milliseconds, but still too much:

I realised I needed to make a sensor that predicted the moment of impact. Sound wouldn't work.
So I started thinking about light.

Continue reading in part 2.