[NEED]Color Sensor Circuit Diagram

Discussion in 'The Projects Forum' started by xKael, Aug 17, 2011.

  1. xKael

    Thread Starter New Member

    Jun 30, 2010
    24
    0
    Hi, I have a color sensor project and I need a color sensor circuit that can sense 10 colors. Specifically the 10 colors that can be seen on a resistor(Bla, Bro, Red, Ora, Yel, Gre, Blu, Vio, Gra, Whi).

    Please make it simple. Thank you for those who will help! :D


    Also, I showed my professor a circuit(which can be found here: http://hobbyelectron.blogspot.com/2011/01/color-sensor-circuit-diagram.html)
    He declined it because it didn't have LDRs yet, the description says it has. And, 10 colors need to be 'sensed', the circuit only senses 8.

    Thanks for those who'll help! :D
     
  2. jj_alukkas

    Well-Known Member

    Jan 8, 2009
    751
    5
  3. Kermit2

    AAC Fanatic!

    Feb 5, 2010
    3,791
    945
    You're kidding right?
     
  4. xKael

    Thread Starter New Member

    Jun 30, 2010
    24
    0
    Nope, not really. What I mean is, keep it as simple as possible.

    @jj_alukkas
    My professor already know that circuit and told us that it's not what we needed. Umm, we need a color sensor that 'converts' a color to voltage. Something like that.
     
  5. Kermit2

    AAC Fanatic!

    Feb 5, 2010
    3,791
    945
    Start by showing us what you have researched and considered trying. What approaches do you think would be workable? Is there a problem you can foresee with any aspects of this project?

    We would love to help, but don't rely on us to 'DO IT' for you.
     
  6. jj_alukkas

    Well-Known Member

    Jan 8, 2009
    751
    5
    exactly, as kermit said.. we dont do it, just help. Once thing straight.. If you need to use a colour sensor, you need to use a microcontroller.
     
  7. xKael

    Thread Starter New Member

    Jun 30, 2010
    24
    0
    Well, part of what I want to know is if the circuit from the website I posted on the 1st post right for a color sensor. If it is, then how should an LDR or LDRs be put on the circuit.

    Thanks for the help anyway.
     
  8. MrChips

    Moderator

    Oct 2, 2009
    12,446
    3,362
    Before you get into circuitry, tell us what you know about the physics of colors.
     
  9. eceblr2011

    New Member

    Jul 8, 2011
    6
    0
    The color sensor implementation is really not straight forward. I had implemented color sensor which used to detect three colors. You have to use good quality of color filters, LDRs or photodiodes, very low offset opamps and voltage references. I had used the photodiodes, the output of the photodiode changes with the wavelength. So one can differentiate between different wavelenghts using the same photodiode. The biaising of the photodiode is also important. Once the circuit is built up, extra effort is required to fine tune the triggering levels for outputting the results for each color. Ambient light conditions will play a major role in getting correct output. It is better to use some good mechanical fixtures which will guide the light to the color filters.
     
  10. xKael

    Thread Starter New Member

    Jun 30, 2010
    24
    0
    And so, I found these:
    [​IMG]

    [​IMG]


    My question is, which of the two circuits is a better color sensor. Also, my teacher told us that being LDR-based, the output may be unstable. What can I do to make it more stable?


    *NOTE: My goal is for a circuit that will sense a color and convert it to current/voltage or something.
     
    Last edited: Aug 18, 2011
  11. MrChips

    Moderator

    Oct 2, 2009
    12,446
    3,362
    Neither will work.
     
  12. xKael

    Thread Starter New Member

    Jun 30, 2010
    24
    0
    Can you.. um.. explain.. um.. why? :)
     
  13. MrChips

    Moderator

    Oct 2, 2009
    12,446
    3,362
    Now that I have your attention, firstly, explain what you are attempting to do. Be specific. Are you trying to determine the color of patches say at least 1x1 inch square or are you trying to read the color codes off a resistor?
     
    Last edited: Aug 18, 2011
  14. wayneh

    Expert

    Sep 9, 2010
    12,127
    3,048
    Neither circuit has an output that depends on which color is "seen". The second one sort of implies it has LDRs that can discriminate, whether because they are different internally or because they are placed behind filters. That's a start. But then their outputs are paralleled together and the spectral information lost. Instead, each should get it's own op-amp. You would then get 3 voltage readings, and possibly that would give you enough resolution to discriminate between your colors.

    You still have mechanical issues of how to focus the light into the LDRS, how to illuminate the test area, and so on. You might consider a fourth unfiltered LDR with a window comparator to ensure that sample illumination is within a certain range where you can get accuracy with the other 3.
     
  15. xKael

    Thread Starter New Member

    Jun 30, 2010
    24
    0
    Um, this is the explanation that my teacher told us.

    We are to make a simple color sensor circuit. A circuit that, somehow, converts color into voltage/current. You put a 2x2 colored card in front of the 3 LEDs and light will bounce back. The intensity of the light that bounces back depends on the color card. The LDR will absorb the light that will bounce back and this will determine its resistance. And so, based on this resistance, the output voltage of the circuit will be determined. Thus, for every color, there is a different voltage output.

    The next part is a Signal Conditioning Circuit, and then an ADC and then a BCD to BCD Converter and then a 7-Segment display but these are not important for now. What's important for now is the color sensor part.

    Anyway, my teacher told us that using LDRs(which were used on past projects by past students) is okay but it won't be very stable. So, is there any way to improve the circuit so it would be more stable?

    Hopefully, you get me this time. :)
     
  16. Kermit2

    AAC Fanatic!

    Feb 5, 2010
    3,791
    945
    A dedicated light source for the circuit. One which can be considered to always be the same EVERY TIME the circuit is operated.

    An enclosed area for the card to be inserted and illuminated by your dedicated light source. This would exclude ALL external light from effecting the circuits operation.

    These two things would be a good step toward making the device more stable in operation and therefore more reliable in output.



    A multiple illumination of the card by several different colors of light in sequence would also increase the devices ability to discriminate between a larger number of colors (tints and shades). With just a single white light it is very possible for complimentary colors to produce an equal voltage output. Doing it this way would require the circuit to have 'memory' to which it can compare readings. All of this would be possible with a mid range micro.
     
  17. xKael

    Thread Starter New Member

    Jun 30, 2010
    24
    0
    Last edited: Aug 18, 2011
  18. MrChips

    Moderator

    Oct 2, 2009
    12,446
    3,362
    The answer for the second and final time is NO.
    There is no need to repeat a post. We can read previous posts.

    BTW, if you expect us to answer your questions, you need to answer ours.
    And we don't do your homework for you.
     
    Last edited: Aug 18, 2011
  19. xKael

    Thread Starter New Member

    Jun 30, 2010
    24
    0
    Oh, okay then. Thanks for the replies anyway.

    So, I guess this won't work also..
    [​IMG]
     
    Last edited: Aug 18, 2011
  20. wayneh

    Expert

    Sep 9, 2010
    12,127
    3,048
    How in the world would you expect a siren to tell you anything about what color something is?

    I like the idea of sliding the test panel into an otherwise dark chamber with consistent, controlled illumination. Like a spectrophotometer. That would increase the chance of success a lot, to eliminate ambient light. Variable, controlled illumination would also increase resolution but adds complexity. You could do it manually, just have switches and run through every possible combination of 3 or 4 LEDs, taking measurements as you go.
     
Loading...