usmanator Posted December 1, 2015 Report Share Posted December 1, 2015 Hello,I'm currently completing a project for my uni course and am struggling.I'm trying to use photodiodes to detect and IDENTIFY light sources - basically I need to design a system that can output what the light source is e.g. sunlight, artificial light, torch, candle light etc. I will use the output from an Op-Amp to feed into a processor which will output to a display (7-segment) showing what the light source is.I have a very rudimentary circuit I am using which involves using photodiodes connected to Op-Amps to amplify any current that passes through them. I'm struggling with there being enough difference between different light sources to successfully identify what the light source is. Measuring the voltage coming out from the various Op-Amps just shows the resistor being being used to provide the gain is what's causing results to differ from the next. I've tried using various different types of photodiodes too but they just provide a voltage regardless of what the light source is too.Is there anything anyone knows that could help me in any way? Much appreciated. Quote Link to comment Share on other sites More sharing options...
audioguru Posted December 3, 2015 Report Share Posted December 3, 2015 Your photodiodes will produce an output current if they have no bias voltage because they perform like a tiny solar cell. Then a photodiode directly feeds the (-) input of an opamp. The opamp will have an output voltage in the dark caused by its input offset voltage and current.Another way is to reverse-bias the photodiode Then it leaks a tiny dark current and a current dependant on the intensity of the light.Since the photodiode simply produces an output level dependant on the brightness of the light then you need to guess what is the source. Quote Link to comment Share on other sites More sharing options...
usmanator Posted December 6, 2015 Author Report Share Posted December 6, 2015 Yes I currently have the photodiode reverse-biased into an Op-Amp with a 1M feedback resistor (changed it to 500k to see different responses). As you say the current is amplified and voltage given at the output.During my various tests I did notice that regardless of what light source I used (normal bulbs, sunlight, flame light etc), the voltage would ONLY vary depending on how close the light source i.e. how bright it was. This is not what I'm looking for.So I began to think at a more basic level - light is compromised of different levels of RGB (or CYMK). If I had filters in between the photodiodes and light source - then I could use the voltages given from them (sent to a processor) and using an algorithm, make a likely guess as to what the light source is?Apart from using actual filters in front of the photodiodes - what about the "TAOS TCS3200D" component? Only thing is I'm unsure of is what the S0 and S1 pins do - what is meant by "output scaling frequency"? Quote Link to comment Share on other sites More sharing options...
audioguru Posted December 7, 2015 Report Share Posted December 7, 2015 Will you teach your processor that the intensity of the 3 colors and white from the sensor are a "..." light source at a certain distance with a certain brightness?I think your processor will be frequently confused because for example the sun and a close candle both produce fire colors.The frequency scaling has something to do with the resolution of the outputs. Quote Link to comment Share on other sites More sharing options...
usmanator Posted December 7, 2015 Author Report Share Posted December 7, 2015 Will you teach your processor that the intensity of the 3 colors and white from the sensor are a "..." light source at a certain distance with a certain brightness?I think your processor will be frequently confused because for example the sun and a close candle both produce fire colors.The frequency scaling has something to do with the resolution of the outputs.Hello,Unfortunately I don't think I'll be able to do that - as you rightly say the processor would be frequently confused. Is there anything yourself (or anyone reading) could suggest? Apart from using colour filters to "likely guess" what the light source is - I don't have much else to go forward with?Looking again at a basic level - I can't think of any other way that a light source could be identified apart from identifying what colours it primarily comprises of? Quote Link to comment Share on other sites More sharing options...
audioguru Posted December 9, 2015 Report Share Posted December 9, 2015 Anything that is hot produces infrared. The sun produces infrared, visible light and ultra-violet.A candle is hot so it also produces infrared but at a level much lower than the sun. I doubt a candle produces ultra-violet.Measure what an artificial light produces. Quote Link to comment Share on other sites More sharing options...
usmanator Posted December 11, 2015 Author Report Share Posted December 11, 2015 Based on what you've said here has got me thinking.I have purchased a "TCS3200D" component. Basically I'm thinking of using this to measure a light source. There are 4 filters (Red, Green, Blue, Clear) above 16 stacked photodiodes. Each filter will give different responses when exposed to different light sources. The ouput would feed into a processor which would make an appropriate "guess" as to the light source. I'm also thinking of having photodiodes separate which are suceptible to IR and UV, and also to provide a redundancy to give greater accuracy to results. I haven't got the components yet - but think it could work? Quote Link to comment Share on other sites More sharing options...
audioguru Posted December 11, 2015 Report Share Posted December 11, 2015 I think the results will frequently be wrong. It might say that a nearby candle is sunlight. Quote Link to comment Share on other sites More sharing options...
login
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.