The light source
One thing that could seem easy to find is a light source. But the thing is not so trivial.
I started using a simple fluorescent low power bulb from Ikea. Cheap, it generated a nice and wide beam of intense light.
But after the first tests I found out that from the Photoelement there was going out a square signal of around 50Hz,
instead of the flat line I was expecting when the shutter was closed.
It was not difficult to correlate these 50Hz with the power line. What I was watching on the oscilloscope was the power line
trough the fluorescent lamp, flipping up and down 50 times each second.
So I decided to use a DC powered light source, a hi-bright LED.
But when I was trying to build the thing, I found accidentally in the net a clever article on using a laser pointer to do the same job.
So I went to the chinese all-in-one shop in the village and bought a couple of laser pointers at 2,75€ each.
They are absolutely under minimum, but come with a set of different collimators, so I could choose one with a tiny beam that
is able to bounce on the window of the building in front of my flat.
The photoelement
I call "photoelement" to any electronic device that turns itself conducting when receives a light beam over.
Photoelements can be LDRs (Light Dependent Resistors), Photodiodes, Phototransistors, etc.
For the sake of speed, I choose a phototransistor, the BPW40. Long time in the market, is cheap and readil available.
I've seen different ideas and schematics on the net. Most are using a photodiode or phototransistor as a detection device,
in a very simple setup like these ones:
- 
The measuring
To do the measuring, several ideas have come out from different people, the most common is feeding the line coming from the
photoelement to the audio input of a normal PC. (That's the reason to put a capacitor in the output in the left schematic, also the power supply is 1.5 volt).
In this case, using an audio program like Audacity, one can visualize the false "audio" signal and measure
on the screen the width of the pulse generated onto the photoelement, and henceforth the shutter speed.
The other option is to draw a digital signal either to a microcontroller that would do the measuring or some people having access to a storage oscilloscope can connect the setup to the oscilloscope input
and watch directly the signal onto the oscilloscope screen. It happens to happen that I own a cheap USB storage oscilloscope (made by PO Labs, called "the POscope", a great and cheap tool (around 160€).
The problem with this approach is that the shutter speed meter is not an authonomous solution and you have to have the PC running and a lot of wires along the workbench.
I would rather prefer to face the use of a microcontroller.
I spent yesterday the whole day at home because an unimportant health problem,
and I was testing the "proof of concept" of the shutter speed tester based on a microcontroller.
Basically what I'm trying to do is feeding two input binary ports of a microcontroller
with the shutter photodiode signal and a CMOS clock of 32768 Hz.
When the photodiode gets light, it turns on driving the port bit to "0".
At that point I start counting the pulses on the other pin until the Shutter signal comes back to 1. Fairly simple.
To prove the concept, I took my old Basic Stamp-I, a stamp size microcontroller board that contains a PIC16C56 microcontroller from Microchip Technologies. This microcontroller has a small Basic language interpreter inside. I don't expect too much performance from such a setup, but at least to be enough to cope counting clock cycles from a 32768 Hz quartz crystal oscillator.
Besides the Basic Stamp, I've built a small circuit based on a CD4060 chip that has an clock generator inside and a divider chain that allows for dividing the main clock upto 16384 times. Nice to be used as s real time clock generator, because if using a 32768 Hz quartz crystal as in our case, we could obtain a 2 Hz clock signal (0.5 seconds), base for most clock designs.
The base for this RTC is a CD4060 CMOS binary counter, that also has an oscillator inside.
Building it is quite simple even for a novice. I use perf board, and previously I design the layout with a nice German CAD tool called LochMaster, that is a program to design circuit layouts over stripboard o padboard. You can even test continuity, so once the design is ready, the success probabilities are very high.
Here you have the design (top view)
and here the bottom side:
And this is the device, piggy-backed on top of the STAMP-I
From the theory to practice there is a big deviation. I expected a pulse count of 16384
to match the 0,5 seconds shutter speed and I was getting only 191.
So I'm doing something the wrong way...
Credits
The Shutter Speed Tester (Soundcard) came from these two places: DIY explanation
and from Schematic
I found here the idea of using a cheap laser pointer as a light source.
Here there is also another person who also thought on using the storage oscilloscope to control the shutter speed
and a fine way of mounting the light source.
