LED_sensor_1.bas (729Bytes)
LED_sensor_2.bas (603Bytes)
This walkthrough describes how to use an LED connected directly to a digital input as a light sensor instead of using an analog input and an amplifying transistor.
I saw this video some time ago, and I was wondering how it was done. There was some discussion about it and then, after a short while, it was quiet. So, when I saw that video again a couple of days ago, I started searching the web and ended up with a very simple demo setup where an 8mm plate with an LED stuck through a hole, functions as a optical button.
Basicly an LED can be used as a capacitator. When you reverse the V+ and Gnd connections, the LED will not light up, but the circuit will hold a bit of charge. The time it takes to get rid of this charge decreases when the LED receives more light. So when its dark it takes longer to discharge than when the LED is in lighter conditions and this time difference can easily be measured with a picaxe.
There was some debate (see comments) as to which component holds the actual charge (the LED, a MOSFET-like circuit inside the picaxe or everything in between the V+ and the LED) but for this to work, it doesn't really matter. It's probably not the LED itself as the time it takes for the charge to fade away gets shorter when you add more LEDs in parallel. If the charge was actually held in the LED you'd expect the capacitance of the circuit to increase with more LEDs and the discharge time to increase as well.
This picture is taken from the pdf in the references.
The LED has both + and - connected to pins on the picaxe. When the LED is ON (a) current flows from an output pin that is set high to an input/output pin that is set low. To charge the circuit, both pins are reversed so that the LED is connected the "wrong" way (b). In the final stage (c), the picaxe pin that is connected to the -(negative) of the LED is configured as an input pin.
To do this with a picaxe 08M you can use the following setup:
This is the setup I did on the breadboard as shown in the video. The red LED is connected to pins 1 and 2 and the green indicator LED is connected to pin 4.
The attached program lights up the red LED for 1 second and then reverses the polarity and discharges the LED in a loop. If it takes more than 80 loops to discharge the LED (i.e. pin 2 voltage drops from high to low) we light up the indicator LED. The code for taking the reading / discharge stage is as simple as this (from Led_Sensor_1.bas):
'stage 3 : Measure darkness => discharge
input triggerpin
do
inc darkness 'count how long it takes
loop while trigger=1 ' for the voltage on the trigger pin to drop below logical 1 / high
Yes it is as simple as that. No transistors needed, just a small loop.
update: to reduce the flicker in the LED as the program waits for a complete discharge, you can also use the following code (from the attached Led_Sensor_2.bas)
'stage 3 : Measure darkness => discharge
input triggerpin
pause 6
readadc10 triggerpin, darkness
Of course, this will only work on a picaxe where you can configure the same pin to switch between being an output and an ADC. If I'm correct thats only the 08M and the X2 series.
I hope this will benefit someone, my head is still buzzing with ideas of fun stuff to do with this.
references:
- Nice howto http://www.merl.com/papers/docs/TR2003-35.pdf
- Wikipedia article: http://en.wikipedia.org/wiki/LEDs_as_Photodiode_Light_Sensors
- makezine article: http://blog.makezine.com/archive/2006/06/led_touch_sensor.html
TODO:
- Make it discharge faster. You can see the LED turning of for a bit when it takes a reading.
- Testing in differtent lighting conditions (intensity and color)
