Probably too late to be of any use, but I am building an M&M sorter. I am using 3 LEDs 1 red, 1 blue, and 1 green and a photo-resistor

(My design is inspired by http://www.philohome.com/sensors/colorsensor.htm)

The three LEDs are connected to a pin on a PIC the photo resistor is connected to a 5V source and an ADC port.

I flash the LEDs and I get an ADC reading (actually I get a bunch of them and average)

Then I compare with sample values.

I am able to get almost 100% correct based on the 6 M&M colors.

(woo-hooo first post by me)

 

that is interesting, but just with 1 LDR???

the only way I can see this circuit pull off is if each light ray has a different frequency (which I think each one does), and colour is determined by frequency.

__________________
-=: The best low-priced components to troubleshoot with are the speaker and the LED :=-

 

Different colours of light ARE different frequencies, although it's not really relevent to the thread?.

What actually makes something look a particular colour depends on what colours are absorbed, and what are reflected, by the object - in this case, it's very easy to see that by switching the LED colours you will find some colours reflect more then others, and can decide which M&M it is.

__________________
PIC programmer software, and PIC Tutorials at:
http://www.winpicprog.co.uk

 

Hmmmm. Is there any difference (and does it matter) between something that reflects purple, which is both red and blue being reflected, and something that reflects violet, which has a frequency higher than red or blue?
Why do purple and violet objects look the same colour when they are completely different?

Shine a red LED then a blue LED on a purple object and an LDR will pickup about an equal amount of reflection from each. Our eyes would also.
Then shine the red and blue LEDs on a violet object and the LDR will pickup almost no reflection from the red and a small amount from the blue, and I think the same with our eyes.
Therefore shouldn't we be blind to violet like our sensor?

Maybe there is no such thing as a violet object.
Shine red and blue LEDs at the same time and you see purple or violet light. Then use a violet filter on a white light (not a "white" LED) and you also see purple or violet. How come?? :lol:

__________________
Uncle $crooge

 

I'm not really sure what you're on about Audioguru?, purple and violet are just slightly different shades of the same colour.

The three primary colours are Red, Blue and Green, you can mix these to get all the visible colours you want (as the monitor you're watching does!) - purple and violet are just Red and Blue, perhaps in slightly different ratios?.

Visible coloured surfaces (paint etc.) work differently though, Red, Blue and Green are the primary colours for LIGHT - this is an addative process, paint is a subtractive process, so the primary colours are different.

So your coloured suface (purple or violet) doesn't 'reflect' purple, it ABSORBS green - leaving the purple arrearance you see. So what you see is the light that hasn't been absorbed - which is where the subtractive part comes in!.

__________________
PIC programmer software, and PIC Tutorials at:
http://www.winpicprog.co.uk

 

Hi Nigel,
Thanks, I understand coloured objects now. A purple object is a primary paint colour and absosbs green light, reflecting red and blue to our eyes. :lol:

But I am still confused by purple or violet light, that has its own high frequency without any red light and hardly any blue light to stimulate our eyes. :?: :?:

__________________
Uncle $crooge

 

As I said, violet is just a particular shade of purple, and includes red and blue light - once you get above the visible range it's a different story, ultra-violet can't be mixed from red and blue. But any of the visible spectrum can be created from the three primary colours.

In fact, purple and violet are often used interchangeably, some people call the resistor code for '7' purple (I do), and others call it violet (some even call it magenta!).

Incidently, a standard TV colour bar signal is:

White - Red, Blue and Green.
Yellow - Red and Green.
Cyan - Blue and Green.
Green - Green.
Purple - Red and Blue.
Red - Red.
Blue - Blue
Black - None.

__________________
PIC programmer software, and PIC Tutorials at:
http://www.winpicprog.co.uk

 

You cod try it whith LEDs audioguru get an red,greed and blue LED and focus them in the same spot on paper then try thurning them on in difrent combinations and you will get difrent colors.

This only simulates the color.Each color has its own freqency.The flourescent light bulb tricks you this way too it aculy emits red, green, blue and a litle ultraviolet.At night thurn on an flurescent light blub and watch a CD.You will only see the colors i menthond before.The UV light is not suposed to be emited.This escapes.Flourecent lights have mercury steam in them and this makes UV light wen ionized.This UV light makes the red ,green and blue phospor on the glass glow.

Thats how the worlds first white LED was made.It containd an red,green and blue LED chip that made it apear white.After that the true white LED was invented that used an single chip that made white light.

Almost evryting that works whith color is based on the red,green and blue.TVs,PC monitors,color LCDs,cameras,our eyes...

If you have an strong magnefing glass you can see those Red,Green and Blue pixels on the TV.(You will hardly see it on your PC monitor since there too small)

__________________
Il give you shocking experience.

 

Years ago I made gadgets with red, green and blue LEDs dimming and brightening slowly, at slightly different speeds. When they shine on my white ceiling at night, they make every combination of colours.

That's correct, and that's my problem. Red and blue light stimulate the red and blue receptors in our eyes. Then our brains tell us the colour is purple. But purple light also has its own frequency that is much higher than red and blue, and our eyes don't have a receptor for it. So how can we see the high frequency of purple?

I don't think so. White LEDs are a blue one with a see-through yellowish phosphor.

Attached Images

white_leds.png (24.5 KB, 295 views)

__________________
Uncle $crooge

 

The old ones used phosphor too.

But now the chip it sef makes white light.


Your eyes can see a bit out of the spectrum so that the blue sensistive cells will also pick up purple.Above we cant see thats the UV spectrum and we cant see the IR spectrum.Mixing red and blue will only make an purple collor but not UV.

UV LEDs emit a lot of visible purple light.But If you put it on paper in the dark it will look a lot brigher and you cod mistake it for an blue LED (since paper is converting the UV in to blue that we can see)The UV black light tubes make almost no visible light but make a lot of UV.You can see those in bars and discos.If the table is coverd by a wite cloth is makes a LOT of visible light.

Dont even think about looking in that UV LED (I tryed it once and its painful and blinding)


It wod be a prety good joke if you get an scary UV reactive mask mask and paint so you wod install an UV bulb and you cod realy scare the guts out of a frend.So it wod look like you are glowing.

__________________
Il give you shocking experience.

 

Hello there again. Well my GP2D120 sensor has arrived and i tested it. As we discussed in this thread, i was looking for something to distinguish between two colours. ( or even black or white :roll: )

In the following pictures you see a simple GUI i made on the PC to be able to observe my results ( and control my robots ). Anyway, on the right you can see a blue bar. That blue bar is ROUGHLY a percantage of ~~~ 0 and 3.3 volts (which actually represent the distance). I should also tell you that GP2D120 has filtering to minimize color distorsion of the distance measurement. when i placed the testing object REALLY close to the sensor 1 ~ 1.5 cm away i got the following results. The first one shows the result when i placed a black object, and the second shows the result when i placed a white object.

That means i will be able to do what i initially wanted (distinguish objects) when i place my robot really close to the object.

Well the results are the following:

Attached Images

	black.jpg (26.8 KB, 262 views)
	white.jpg (35.1 KB, 262 views)

 

hey i want to design a circuit that differentiates between different shades of the same color say for eg. light green and dark green... can someone pls help me out?

 

I find it interesting that one has a hard time convincing the uninitiated that a color CRT does not generate BLACK and that the screen on a TV that's off is as black as it gets. It's amazing how contrast can fool the eye.

Dean

__________________
Dean Huster, Electronics Curmudgeon
Contributing Editor emeritus, "Q & A", of the former "Poptronics" magazine (formerly "Popular Electronics" and "Electronics Now" magazines).

R.I.P.

 

Actually there is a big difference between true violet and purple. The former is a spectural colour and the former is just a mixture of two different wavelengths of light.

If you look at an ideal 410nm light source like a laser it will appear violet.

If you looked at a red and blue beams of light shining on top of each other; the blue being much brighter than the red. You'll see the same colour.

Physically the two light sources are totally different. One is a mixture of red and blue and the other is a pure wavelength.

The reason we see shorter wavelengths as more redish colours is due to the frequency response of the different cone cells in our eyes.

http://en.wikipedia.org/wiki/Purple#..._versus_violet

__________________
I also post at the following sites:
http://www.stop-microsoft.org http://www.heated-debates.com
Screen name: Aloone_Jonez

 

I dont know if it is the same one but here is an M&M sorter ive seen:
http://blog.makezine.com/archive/200...mm_sorter.html

Although the link to the actual projects down at the moment the blurb does say:


"Identifying color is easy with the new Texas Advanced Optical Systems TCS230 frequency to color sensor. This is a high-sensitivity low-noise light-to-voltage optical converter that incorporates on board blue, green, and red optical filters. The sensor combines a photodiode and an amplifier on a single monolithic CMOS integrated circuit with a color filter over the photodiode. What's really neat about this sensor is that it provides standard RGB color values with 90% accuracy."