I want to measure the reflection of light with at least 10 bit resolution. Therefor I put a phototransistor (BP103B) in series with a 10k resistor and a sensing pin in the middle. The light comes from an LED (NSPL510DS or LL-583VC2C-V1-4DA) in series with a 0.3k resistor.
Holding a sheet of wihte paper in front of the setup, the sensor value changes about 35 -- and I wanted 1000. What can I do to improve my measurement?
I could surround the phototransistor with 6 LEDs and take a more adequate resistor, though I would expect a change less than 500, and I need at least 1000. Even more light would be better, then I could use an additional diffusor to get a more homogeneous illumination. Are there better LEDs/Phototransistors?
Are there devices like CCDs/CMOS-Sensors in a digital camera, where I can change the exposure time to get along with low light?
Could I use a external A/D converter with a higher resolution? Is there a tutorial and a libary?
Do you have suggestions for a simple solution?
Change the 10k resistor to alter the range of values to whatever you want.
Note that 10bit precision is unobtainable and meaningless(*) in this context - but the dynamic
range that 10+ bits gives you is useful. In low light the phototransistor impedance increases
and you probably need 100nF on the analog pin to stablize the readings (but at expense of
slower settling time).
If accuracy is important then using a good photodiode in reverse bias mode and sensing the current
is more repeatable and less sensitive to temperature variation. That takes an opamp as a current
to voltage converter.
(*) the variation of illuminance and sensitivity with direction in typical devices is many percent. Also
reflectivity of surface depends on angle of incidence and angle of reflection, again many percent.
Many thanks, MarkT!
With a 20k resistor I get a change of about 60, maybe 100k and some more LEDs would do it. One value per second is enough for me, so the 100nF won't hurt. If I set the LED directly in front of the FT, I get the full 1023. Though, I don't know about linearity.
But the temperature variations may be a problem, so I should test your suggestion with the photodiode. Do you have a link with the circuit? I have no experience with an opamp and even don't really know what it is.
I got the "light-to-voltage" sensor TSL250R which is easy to handle and very sensitive. The output range is from 0 to 788. There are less sensitive versions (TSL251R, TSL252R). The main drawback is that I can't set the amplification on my own. I will see later, if I need a A/D-converter with a higher range.
I want to determine the amount of coating on a surface. I'm not finally sure about the coating and the surface, but the coating may be small particles (~µm) the surface could be glass with a diameter of 20mm and the coverage is somewhere between none and total.
At the moment the exact placement of the components is my main task.
Though, maybe a light-meter with adjustable sensitivity would make it easier.