Constant Current cource

Discussion in 'The Projects Forum' started by pratik21, Aug 8, 2014.

  1. pratik21

    Thread Starter New Member

    Aug 8, 2014
    1
    0
    Hello,
    I am creating a constant current source for LED to use it as a constant light source for Biosciences experiments.
    I have designed a circuit in which there is a Voltage Reference IC 5025I(Texas Instruments) which will provide me with a constant voltage of 2.5 V.
    This 2.5 V will be converted into a constant 25 mA current with the help of a high precision V-I converter XTR111(Texas Instruments) which will be further connected to an LED for constant light source.
    I have attached my schematic.
    The problem I am facing is that there is no output current or voltage to load.
    The XTR111 IC is heating a lot!
    The ref5025 IC gives proper output when XTR111 is not connected, but just when XTR111 is connected, values of output current,voltage become zero
    The Power supplies are as follows:
    24V supply for XTR111
    5V supply for voltage regulator of XTR111
    3.3V supply for REF5025I
     
  2. mitko89

    Member

    Sep 20, 2012
    123
    19
    Did you check your soldering? The schematic design seems right to me. Try changing the transistors with MMBT2907 and BC807.
     
    Last edited: Aug 8, 2014
  3. Dodgydave

    Distinguished Member

    Jun 22, 2012
    4,974
    744
    why don't you use an Lm317 as a constant current source instead,a 47R resistor will give 26mA.
     
  4. mcgyvr

    AAC Fanatic!

    Oct 15, 2009
    4,769
    969
    All that when just a single resistor will do the same thing?
     
  5. bertus

    Administrator

    Apr 5, 2008
    15,646
    2,345
    Hello,

    As said, when the powersupply is regulated, a resistor will be enough to supply the current for the led.
    The value of the resistor will be (Vsupply - Vled) / Iled.

    Bertus
     
  6. ErnieM

    AAC Fanatic!

    Apr 24, 2011
    7,386
    1,605
    Are you sure about Q1? It looks to me that it will turn the output full on in the event of an overcurrent condition.

    I believe you desire a completely opposite functionality.

    Edit: nevermind... the pass element is a P channel so it will indeed turn off.
     
    Last edited: Aug 8, 2014
  7. ErnieM

    AAC Fanatic!

    Apr 24, 2011
    7,386
    1,605
    A simple resistor cannot guarantee a constant current thru the LED even though the supply voltage is constant as the LED voltage itself is not a constant.


    Also, as far as I know a constant LED current will not guarantee a constant LED intensity.
     
  8. mcgyvr

    AAC Fanatic!

    Oct 15, 2009
    4,769
    969
    correct.. The OP needs to let us know or decide how "accurate/consistent" this light source needs to be.
     
  9. ErnieM

    AAC Fanatic!

    Apr 24, 2011
    7,386
    1,605
    I once needed a damn accurate, or should I say damn constant pair of LEDs for a blood monitor used in kidney dialysis. The basic concept was to illuminate a tube containing the waste water with alternating red and blue light. As blood will attenuate the blue signal more than the red the device was surprisingly sensitive to small amounts of blood in water.

    There is a whole write up on how much blood is too much; the basic problem being in dialysis a patient’s blood is forced against a membrane that passes just the impurities to be flushed out. However, should the membrane be compromised there is a short circuit of blood out of the patient and literally(!) down the drain.

    Nothing keeps you on your on your design skills toes like the thought of your safety device not working thus killing another human being.

    While the intensities of the LEDs were set with an initial calibration step I did have data indicating they may not stay that way with temperature changes, either from self heating or from a portable machine going from a cold car to a warm room.

    My boss using his limited skill set would not allow any exploration for temperature effects unless the customer explicitly demanded it, and I was forced to only report on operation at room temperature. Yeah, let it run over a weekend then a week and it still detected the correct amount of blood, but left me very unsatisfied.

    Making the calibration samples was amusing. I used a commercially available bovine hemoglobin sample, which is the fancy way of saying I had a jar of chemically pure dried cow blood. We rented a precision scale for a week and I measured out 100 samples, each a fraction of a gram, and stored them in brown glass vials with screw on lids.

    Where does one obtain 100 brown glass vials with screw on lids? Well at the local “head shop,” of course. There are two close by to where I live. Cheap too, something like 5 or 10 cents a piece.

    Bottom line, LEDs do vary with temperature
     
  10. bertus

    Administrator

    Apr 5, 2008
    15,646
    2,345
  11. wayneh

    Expert

    Sep 9, 2010
    12,094
    3,033
    But didn't your design inherently, by looking at the ratio of colors, mitigate much of the problem? I guess if one color moves much differently than the other color, then you would need to adjust the ratio target to compensate for temperature. But as long as they react somewhat similarly, the ratio would still stay in a narrow range that is small compared to all the other experimental variances. Just a guess.
     
Loading...