Connect with us

Constant Current Source again!

Discussion in 'Electronic Design' started by GraemeC, Sep 1, 2006.

Scroll to continue with content
  1. GraemeC

    GraemeC Guest

    Hi people

    I've posted on this before and I've never really got it resolved.

    I need to create an adjustable constant current drive that can supply
    between 150-300mA to drive an experimental LED. I only have a 5v
    supply to play with.

    I think there is some confusion out there with regards the term
    "constant current source"

    If you look at wikipedia "constant current source" and look on the
    discussion page you will see a heated debate going on.

    What I mean is a source that will provide the same current regardless
    of the supply voltage and load resistance(obviously within reason). I
    would expect to change the value of the current by changing ratios of

    I guess have two questions. Does anyone have a design that meets my
    spec for a constant current source and could it be modified to meet my
    design criteria.

    I know we can do this!

  2. Guest

    Yes, but you haven't told us what they are (speed, cost, stability,
    accuracy, or any other considerations you have).

    James Arthur
  3. Genome

    Genome Guest

    Seems pretty simple to me. Then again you are posting from Google and have a
    Gmail account...... and also mention the Wiki......

    Then you say....

    "I've posted on this before and I've never really got it resolved."

    So perhaps you are looking for something 'more' than a constant current

    As you say.....

    "I need to create an adjustable constant current drive that can supply
    between 150-300mA to drive an experimental LED. I only have a 5v
    supply to play with."

    Again... seems quite simple to me..... so I am left thinking you are hiding
    something and perhaps you are doing it deliberately!!!!!! I mean.....

    Can we expect that later on you will inform us that this has to be supplied
    from the auxilliary winding of a multiphase multisource power supply pack
    for use in an SLICBM and..... whoopsy, you meant to ask for one of those as
    well. Plus it turns out that the experimental LED is part of the on board
    high energy anti-anti-missile tracking system and you also want to pulse the
    laser whilst tracking on target with the facility for multi targeting as the
    MIRVS go down whilst the mother delivery system remains in orbit?

    Or are you just going to ask for a programmable current source and change
    your mind as time goes by for a really nice long thread?

  4. The simplest may be good enough:

    A PNP transistor, emitter hooked to a 2 ohm resistor, other end of
    resistor to a 10 ohm pot, 2Watt, other end of pot and wiper hooked to

    base hooked to the cathode of a 1N4007 diode, anode hooked to another
    1N4007 cathode. anode of second diode hooked to +5. Base also has a
    330 ohm resistor to ground.

    collector is your + side of the current source, ground is the - side.
    With the pot at maximum resistance, you'll get 100mA; when at lowest
    resistance, you'll get 300mA.

    This should hold the current steady within 7.2% from zero to 3.14159
  5. How close to a constant current does it have to be? A 2N3055 will do it;
    put 3.3 ohms 2 W in series with the emitter and vary the base voltage
    with a 330 ohm resistor and a 100 ohm pot in series across your supply
    (330 ohms to +5 V). The current will stay pretty constant at the value
    you set, if you put a big heat sink on the transistor. The 3.3 ohms
    provides negative feedback to keep the current more constant than it
    would otherwise be.

    An LED is a very low resistance device, so even though the 2N3055 isn't
    a brilliant choice for high output impedance, it makes a very simple
    circuit. You can go on to refinements later, if necessary. A power FET
    is probably the next step, but you only have 5 V to play with, so it
    needs care.
  6. Guest

    The OP hasn't posted what sort of LED he's using so we don't know
    what kind of compliance he needs.

    Still, substitute an IR LED for the two 1n4007s and you'll get more
    headroom, plus temperature compensation.

    James Arthur
  7. Guest

    Ok, that was uncalled for.
  8. Genome

    Genome Guest

    In what way was it uncalled for?

  9. Almost all of what DNA contributes is uncalled-for, but when you get to
    SMPS and, as recently demonstrated, proximity effect, the slate is wiped
  10. What is the LED? Or, what are its simplified characteristics? A
    simple LED model uses a voltage and a resistance in an equation
    something like this:

    V(diode) = 1.55V + 2 ohms * I(diode)

    Or somesuch. Do you know what voltage is required for your LED, when
    driven at 150mA and at 300mA? Or can you at least say what voltage
    will be required at 300mA? How much less than 5V will cover it?
    (Stating the LED technology, at a minimum, would help.) Any
    precision/noise, drift over temperature and time, and accuracy specs?
    What about variation over supply voltage? How much supply voltage
    variation will there be to tolerate?

    How will you vary the current? Manual knob? Computer control?
    I don't think there is _any_ confusion about what a current source is.
    It would help to know what you are trying to do, in detail. You had
    an opportunity to clarify these things earlier and didn't take that
    chance, as far as I'm aware.

  11. What is the peak current the LED can handle and at what duty cycle?
  12. Fred Bloggs

    Fred Bloggs Guest

    Transcendentally compliant CS- too funny...
  13. Fred Bloggs

    Fred Bloggs Guest

    A simple prototype, made of simple and readily available parts, with
    maximum compliance, would look something like this:
    View in a fixed-width font such as Courier.
  14. PeteS

    PeteS Guest

    Especially as it's supposed to be driving a load with a forward
    characteristic proportional to e ;)


  15. Fred Bloggs

    Fred Bloggs Guest

    Maybe juice that thing up a bit:
    View in a fixed-width font such as Courier.

  16. The slate may be clean, but not Genome.

    Service to my country? Been there, Done that, and I've got my DD214 to
    prove it.
    Member of DAV #85.

    Michael A. Terrell
    Central Florida
  17. We only have his word for that, of course, and other things he writes
    aren't exactly credible.

  18. And the pictures of his cooking don't look edible. ;-)

    Service to my country? Been there, Done that, and I've got my DD214 to
    prove it.
    Member of DAV #85.

    Michael A. Terrell
    Central Florida
  19. Ban

    Ban Guest

    the thermal behaviour is not very promising, since Hfe increases with T, as
    does Vbe decrease and also the Vf of the diode, all leading to an increase
    in Ic. I tried a simple compensation and it looked nice in the simulation:

    |470R...2k2 |
    .-. |
    | | |
    | |R1 LED V /_
    '-' - \
    | Pot 1k |
    +-------. |
    | | |
    V | |
    - D1 .-. |/
    | | |<---| BD139
    | | | |>
    V D2 '-' |
    - | .-.
    | | | |2R2
    | V D3 | |
    | - '-'
    | | |
    .-. .-. ===
    | | | | GND
    | |100 | |100
    '-' '-'
    | |
    === ===
    (created by AACircuit v1.28 beta 10/06/04

    R1 has to be chosen empirically according to the Hfe of the transistor. The
    range is from 20 to 330mA. The 3 diodes have to be mounted in thermal
    contact with the transistor.
  20. GraemeC

    GraemeC Guest

    Thanks for all help everyone

Ask a Question
Want to reply to this thread or ask your own question?
You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.
Electronics Point Logo
Continue to site
Quote of the day