Parallel resistor circuit

Discussion in 'The Projects Forum' started by susannah, Feb 27, 2014.

  1. susannah

    Thread Starter Active Member

    Feb 14, 2010
    87
    0
    I have 72 LEDs which can be on or off and software in my computer needs to know which ones are on and which ones are off.

    I was thinking putting LDR top of each LED and then using parallel resistors. Or serial?? This way i only need to measure one wire, right?

    I don't have access to the electronics or wires of these 72 LEDs.

    Anybody know where i could find circuit plans like this. Thanks
     
  2. ericgibbs

    Senior Member

    Jan 29, 2010
    2,499
    380
    hi,
    How do you plan to 'read/connect' the LED's/LDR's states into your PC.?

    E
     
    Last edited: Feb 27, 2014
  3. wayneh

    Expert

    Sep 9, 2010
    12,088
    3,027
    Are the LEDs moving or fixed? Another approach would be to use a camera and software that can look at each of 72 assigned regions of the image.
     
  4. sirch2

    Well-Known Member

    Jan 21, 2013
    1,008
    351
    When you say computer do you mean a PC? If so as Wayneh suggests, go for a cheap webcam.
     
  5. susannah

    Thread Starter Active Member

    Feb 14, 2010
    87
    0
  6. ericgibbs

    Senior Member

    Jan 29, 2010
    2,499
    380
    Hi,
    Are you thinking of trying to detect the change in resistance value of a series resistor chain in order to determine which LED's/LDR's are light or dark.?

    E
     
  7. susannah

    Thread Starter Active Member

    Feb 14, 2010
    87
    0
    Yes. Thanks
     
  8. ericgibbs

    Senior Member

    Jan 29, 2010
    2,499
    380
    hi,
    Have you calculated the possible permutations of 72 LED's.:eek:

    Its not a viable plan, consider the option suggested by 'wayneh', a simple web cam with on screen X, Y image point recognition.

    E.
     
  9. wayneh

    Expert

    Sep 9, 2010
    12,088
    3,027
    Of course there is! Choosing the right one might be a challenge, though. How quickly do you need to make repeat measurements? Every day, every millisecond? Do you have any programming experience, or favorite language that you might prefer to get this done?
     
  10. susannah

    Thread Starter Active Member

    Feb 14, 2010
    87
    0
    I need repeat about once a day only.

    I can program in AppleScript and little Unix. CLI app would be nice.

    Permutations???
     
  11. wayneh

    Expert

    Sep 9, 2010
    12,088
    3,027
    Let's start with the obvious. How about grabbing an image and posting it on the internet for a human to judge which LEDs are on or off? In 24hrs somebody is likely to come by.

    Applescript or Unix won't help you much (although either might be useful for automating a frame capture from the video camera). You need something that can read and process an image file. Unfortunately I don't know what that would be. It would be "trivial" for someone that knows what they're doing. Sooo much easier than facial recognition, for instance.

    I'd start by getting Apple's free developer tools, and then using the developers forum as a place to get some helpful advice on which direction to go. This is a bit of a steep learning curve, depending on your comfort level.
     
    susannah likes this.
  12. susannah

    Thread Starter Active Member

    Feb 14, 2010
    87
    0
    Thanks.

    I will probably use camera, but i still would like to learn how to do this thing for 5 LEDs: "detect the change in resistance value of a series resistor chain in order to determine which LED's/LDR's are light or dark".

    There sure is somewhere in the net circuit plans like this and calculators what size resistors to use??
     
  13. sirch2

    Well-Known Member

    Jan 21, 2013
    1,008
    351
    I know nothing about Macs but I assume there is some sort of scheduler (cron?) in the OS and an app that can save a webcam frame to an image file? I guess the best place to ask would be a Mac forum but most common programming languages can manipulate image files and would allow you to extract the colour value of each pixel (there are also libraries like OpenCV).
     
    susannah likes this.
  14. ericgibbs

    Senior Member

    Jan 29, 2010
    2,499
    380
    Hi S,

    A major problem with the parallel LDR's is that the Light/Dark resistance will not be Zero/Infinite, but they will have indeterminate resistance value depending upon the LED light intensity and the possible ambient light.

    The LDR's will also have 'typical' Light/Dark resistance that will be different from LDR to LDR.

    Even for only 5, LED/LDR's you will have to select from a batch quantity, 5 that have matched light/dark resistance values.

    I would not consider that method.

    E.
     
    susannah likes this.
  15. wayneh

    Expert

    Sep 9, 2010
    12,088
    3,027
    Yes, getting the screen grab can almost certainly be accomplished with existing software. Getting an image analyzer application to do the analysis might also be possible. There are a lot of image analyzer apps used in science, to read chromatography gels and all sorts of other things. Your application seems a lot like detecting red eye (if the LEDs are red?). Last resort should be writing your own program, but it might be needed. Absolutely you would want to start with something that is as close as possible already, so that you don't have to reinvent the wheel.
     
    susannah likes this.
  16. wayneh

    Expert

    Sep 9, 2010
    12,088
    3,027
    Would using phototransistors for detectors help solve these problems? I'm imagining them as much more on and off.
     
  17. ericgibbs

    Senior Member

    Jan 29, 2010
    2,499
    380
    hi wayneh,
    If the series PT where used in reverse bias mode, the PT light current would be shunting the current flowing in the parallel resistor, I would consider that detecting a such a small PT current change would be very difficult to detect and to determine which PT's were conducting, even for a 5 chain.

    If used in forward bias mode and assuming a 1V Vfwd for each PT, would mean at least > 5V would be required across a chain of 5.
    The voltage generated across the forward biased PT, would change the total current flowing in the parallel resistor chain, but again determining which PT is On/Off IMHO would be close to impossible.

    I would go with your web cam screen image method, if I was doing this project.

    Eric
     
    susannah likes this.
Loading...