Connect with us

Track another Voltage rail with fixed 5V drop @500mA question

Discussion in 'Electronic Design' started by The Hemulen, Apr 11, 2013.

Scroll to continue with content
  1. The Hemulen

    The Hemulen Guest

    I need a Voltage rail to track another rail with a fixed 5V drop between the
    two. ie if main rail is at 16V output at 11V, main rail drops to 15V output
    at 10V etc. Need to pull up to 500mA from this tracking output.

    An obvious way would be a series power NPN transistor with 5V6 zener between
    collector and base and resistor from here to ground but I'd also like the
    tracked output to have some current protection in case the output gets
    shorted.

    Is there a sneaky way to do this with standard 3 pin fixed or adjustable
    regulators that would provide both the fixed drop between the two and short
    protection?

    Thanks
     
  2. You could hang a negative three pin regulator 'upside down' off the
    top rail, but that will only sink current. (I think)

    George H.
     
  3. Guest

    Something cheap, the 6.25 could be 6.2V zener or precision shunt IC:
    Please view in a fixed-width font such as Courier.

    ..
    ..
    .. MAIN RAIL >--+-------------------+---------> Vmain
    .. | |
    .. | |
    .. | -
    .. | ^ 6.25V
    .. | 317 - ref
    .. | ----- |
    .. | | | |
    .. +--|I O|----------|-----+---> Vmain-5 V
    .. | | A | | |
    .. | ----- | |
    .. | | | |
    .. | | | |
    .. === -------------+ === C
    .. | | |
    .. | [R] |
    .. | | |
    .. | | |
    .. ----+-------------------+-----+--------
    ..
    ..
     
  4. Guest

    Main Rail o---+---------------------------------o Main Rail
    |
    +--->|--->|--->|--->|--->|---~~---o Tracking Rail
    D1 D2 D3 D4 D5 F1

    Ground o-------------------------------------o Ground

    D1-D5: 1N4001
    F1: about 1 A; fast or slow blow to taste

    F1 is the expensive part. :)

    Matt Roberds
     
  5. Guest

    It meets all the stated requirements of the original poster. If there
    are additional requirements, then the above design may not be
    appropriate.

    For lower noise:

    Main Rail o---+-------------------------------------o Main Rail
    |
    +--->|--->|--->|--->|--->|---~~---+---o Tracking Rail
    + | D1 D2 D3 D4 D5 F1 + |
    C1 === === C2
    | |
    Ground o---+---------------------------------+--o Ground

    D1-D5: 1N4001
    F1: about 1 A; fast or slow blow to taste
    C1-C2: 100 uF. 25 V or more depending on main rail voltage

    For better accuracy:

    Main Rail o---+-----------------------+----------------o Main Rail
    | |
    | |
    | |
    | +----------------+
    | ____ | | |
    | | | | | |
    +-----+---| U1 |---+-----+ / R5 / \ M1
    | + | |____| + | | \<-+======( )
    | C1 === | === C2 | / | /\_/
    | | | | | | | | |
    | | | | | +--+ | |
    | +---+-+------+ | | | |
    | | | | | |
    +-----+ | +------+ +----------------o Tracking Rail
    | | | | | |
    / | | / | |
    \ R1| | \ R3 | |
    / | ____|_____ / | |
    | | | Vcc | | | |
    +------|Ain1 Ain2|-----+ | |
    | | | Dout1|-+ | | |
    / | | U2 | | / | |
    \ R2| |___Gnd____| | \ R4 | |
    / | | | / | |
    | | | | | | |
    | | | +-------------+ |
    | | | | |
    Ground o---+---+-------+-----------+-----------+----o Ground


    C1: 100 uF. 25 V or more depending on main rail voltage
    C2: 10 uF, 6.3 V
    R1, R3: 9k
    R2, R4: 1k
    R5: 100 ohms, 3 W, wirewound pot
    U1: 7805
    U2: PIC, AVR or similar (with at least 2 analog in, 1 digital out,
    internal clock, internal ADC reference, 5 V supply)
    M1: Model airplane type servo motor

    Note: === is the mechanical linkage of R5 and M1

    Matt Roberds
     
  6. Jamie

    Jamie Guest







    MAIN buss 5..20 Volts
    +---------+------++
    | .-.
    | | |
    | | |10
    + '-'
    z |
    4.8V A |
    + |/
    +----+| 5 volt drop
    | |>
    + +------------------+
    .-. |
    | | .-.
    100k | | | |
    '-' | |30 output LOAD
    | '-'
    | |
    === ===
    GND GND
    (created by AACircuit v1.28.6 beta 04/19/05 www.tech-chat.de)

    Just an idea..

    This is simple and if you are not looking for supper precision
    this should do.
    You can use a post regulator or something like it ofcourse and
    lift the common leg with a zener.

    I guess it all depends on real-estate and your needs.

    Jamie
     
  7. Tim Williams

    Tim Williams Guest

    Pffbt! And this from the guy who approves of hFE bias in production?
    Come on, it's even got damping for those pesky emitter follower mode
    oscillations! ;-)

    Tim
     
  8. Guest


    / / / /
    / / / /
    / / / / / /
    .----. / .----. /
    | |/ / | |/ /
    '\ /' / / '\ /' / /
    ____/ | | / / /___________________/ | | / / /_____
    / / | |/ / / / | |/ / / /|
    .---' '---/ / .---' '---/ / / |
    ___| |/____________________| |/_____/ |
    '---------' '---------' | '
    Main rail Tracking rail | /
    ____________________________________________________|/

    Ground
     
  9. Nico Coesel

    Nico Coesel Guest

    A TCA0372 (high power, thermal and short circuit protected opamp), a
    shunt regulator and some resistors.
     
  10. Nico Coesel

    Nico Coesel Guest

    It doesn't. Look at the datasheet of the 1N4001 and especially the
    graph forward current versus forward voltage. And then look at what
    the temperature does to the forward voltage.
     
  11. 1.2 cents? You buying them one at a time at Starbucks?


    Best regards,
    Spehro Pefhany
     
  12. tm

    tm Guest

    LOL. Good one.
     
  13. Guest

    I liked Matt's circuit, but thought it needed illustration...
    Fred Bloggs is the ASCII King.
     
  14. Guest

    Since there's no ac-ripple to speak of as the input is also regulated, a simple current limiting resistor in series with ADJ should be enough. Since on the modern parts Iadj doesn't budge by 0.2uA typically across all combinations of input voltage and loading varied from min to max.
    Please view in a fixed-width font such as Courier.

    ..
    ..
    .. MAIN RAIL >--+-------------------+---------> Vmain
    .. | |
    .. | |
    .. | -
    .. | ^ 6.25V
    .. | 317 - ref
    .. | ----- |
    .. | | | |
    .. +--|I O|----------|-----+---> Vmain-5 V
    .. | | A | | |
    .. | ----- | |
    .. | | | |
    .. | | | |
    .. === '----[1K]-----+ === C
    .. | | |
    .. | [R] |
    .. | | |
    .. | | |
    .. ----+-------------------+-----+--------
    ..
    ..
     
  15. Guest

    The 317 doesn't enough internal resistance to protect itself against excessive current from Adj to Out with Out shorted in this application, so the 1K in series with Adj is required.
    Please view in a fixed-width font such as Courier.

    The 6.25V reference is best done with a 431 type precision shunt reference IC:

    ..
    ..
    .. MAIN RAIL >--+-------------------+---------> Vmain
    .. | |
    .. | |
    .. | -
    .. | ^ 6.25V
    .. | 317 - ref
    .. | ----- |
    .. | | | |
    .. +--|I O|----------|-----+---> Vmain-5 V
    .. | | A | | |
    .. | ----- | |
    .. | | | |
    .. | | | |
    .. === '----[1K]-----+ === C
    .. |0.1u | |
    .. | [R] |
    .. | | |
    .. | | |
    .. ----+-------------------+-----+--------
    ..
    ..
     
  16. Guest

    Key word: stated.

    Quick version: I know the voltage drop from the main rail to the
    tracking rail is going to vary. The original poster didn't say how much
    variance was acceptable.

    Long version:
    I actually did look at one data sheet (Diodes Inc) before I posted the
    first time, because I had one memory of the nominal forward voltage
    being 1.0 V and another of the nominal forward voltage being 0.6 V.
    That data sheet quotes a maximum forward voltage of 1.0 V at 1.0 A
    forward current, 25 C ambient, single phase, half wave, 60 Hz,
    resistive or inductive load. (Yes, this application does not meet
    those conditions.) So I went ahead with 5 of them in series.

    That data sheet also has a graph that shows forward voltages of about
    0.7 to 0.9 V over currents from 0.01 to 1.0 A at a _junction_
    temperature of 25 C, but this graph also has the test conditions of
    "pulse width 300 us, 2% duty cycle". I'm sure there is a reason for
    this test condition, but I don't know what the reason is, so it seems
    a little weird to me. The application that the original poster asked
    about is probably much closer to continuous than pulsed.

    I just looked at some other data sheets. Vishay's has a similar If vs
    Vf chart, with about the same curve, except their test condition is a
    1% duty cycle. Fairchild has a similar chart and curve and uses the
    2% duty cycle. ON has kind of a different chart; it doesn't have the
    pulsed test conditions, and has curves for three different Tc (case?).
    For 25 C, Vf varies from about 0.79 to 0.91 V over If 0.1 to 1 A.
    The Diodes Inc, Vishay, and Fairchild data sheets don't have any
    temperature vs forward voltage data. The ON one sort of does, in the
    three curves on the same graph mentioned above. At a constant If of
    0.5 A, Vf varies from about 0.72 to 0.86 V as Tc (case?) varies from
    150 C to 25 C.

    So yeah, Vf will vary over If and temperature. The original post did
    state a current of "up to" 0.5 A, which implies that it could be less,
    so If will probably be changing. The original post did not state a
    temperature requirement; I realize the diodes will heat up some from the
    power dissipated in them, so temperature will probably be changing.

    This means that in this circuit, Vf will change, so the total voltage
    drop over D1-D5 will change, roughly in the range of 3.5 to 4.5 V.
    There will be a small additional voltage drop from the fuse (Bussmann
    says an AGC-1/2 is 0.59 V typical at rated current), so the total
    voltage drop from the main rail to the tracking rail will vary over
    the range of roughly 4.1 to 5.1 V.

    The original post did not state a tolerance on the 'fixed 5 V drop'
    requirement. The claim could be made that "4.1 to 5.1 V doesn't meet
    the 'fixed 5 V drop' requirement", but that would mean that an
    alternative solution that provides a 4.9999 to 5.0001 V voltage drop
    also doesn't meet the 'fixed 5 V drop' requirement.

    The fanciest DMM I can find in a little bit of Googling is an 8.5-
    digit one (Agilent 3458A), and its resolution on the 10 V range is
    100 nV. So, I guess if there is a solution that provides better than
    4.9999999 to 5.0000001 V voltage drop, it would be hard to prove that
    it didn't meet the requirement.

    Data sheets referred to:

    Diodes Inc http://www.diodes.com/datasheets/ds28002.pdf
    Vishay http://www.vishay.com/docs/88503/1n4001.pdf
    Fairchild http://www.fairchildsemi.com/ds/1N/1N4001.pdf
    ON http://www.onsemi.com/pub/Collateral/1N4001-D.PDF

    Standard disclaimers apply; I don't get money or other consideration
    from any companies mentioned.

    Matt Roberds
     
  17. Guest

    Right, is definitely the way to go, especially if there's substantial variation on the main supply rail.

    This circuit looks hopeful, avoids the short circuit through the Adj, uses cheap readily available parts, and accuracy is power supply adequate. Adj currents cancel:

    Please view in a fixed-width font such as Courier.

    ..
    ..
    .. MAIN RAIL >--+-------------------+---------> Vmain
    .. | |
    .. | |
    .. | | Metal Oxide
    .. | [1.0K] 1%
    .. | 317 | 1W
    .. | ----- |
    .. | | | |
    .. +--|I O|------------------------+---> Vmain-5 V
    .. | | A | + | |
    .. | ----- | |
    .. | | 1.25 | |
    .. | | | |
    .. | | - | |
    .. === '-------------+ ===
    .. |0.1u | |0.1u
    .. | 317 ----- |
    .. | | I | |
    .. | --|A | |
    .. | | | O | |
    .. | | ----- |
    .. | | | |
    .. | | | | |
    .. | | [200] | 6.25mA |
    .. | | | | |
    .. | | | v |
    .. ----+-------------+-----+-------------+--------
    ..
    ..
    ..
     
  18. The Hemulen

    The Hemulen Guest

    Many thanks for all the replies.. Tempting though the PIC/Motor based
    solution is I think the LM1117 with 6.25V ref should do nicely.
     
  19. John S

    John S Guest

    Don't forget that you will dissipate 2.5 watts at 500mA.
     
  20. Guest

    Four units is a production quantity, right?

    I don't do hardware design for a living; I do software for a living.
    Recently it seems I have been doing various kinds of embedded software.
    When I do anything with hardware it's usually for fun. Sometimes it's
    for work, but then it often consists of providing goofy ideas to the
    real official hardware guy. Mostly this provides comic relief to the
    real official hardware guy. Once a year or so, it might help him get
    unstuck - not by using one of my goofy ideas, but by helping him come
    up with a goofy idea of his own that actually works. :)

    This all furthers my goal of getting the board in my hands sooner, so
    I can write the software, and then get on to the real fun of arguing
    with the hardware guy about whether it's his crummy board or my crummy
    code that is causing it to screw up. Sometimes we have a shootout in
    the lab at noon with a compiler and an oscilloscope. If neither of us
    turns out to be a good shot, the sheriff^Wboss comes by and talks to
    both of us. Usually the boss decides that adding more crummy code is
    cheaper than revising the crummy board. This gives me more code to
    write, frees up the hardware guy to start working on the next board,
    and frees up the boss to slide the bar ends to the left on the Gantt
    charts, so everybody is happy? Right?

    Matt Roberds
     
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

-