Connect with us

Simple circuit, stupid problem

Discussion in 'Electronic Basics' started by Eriswerks, Aug 21, 2005.

Scroll to continue with content
  1. Eriswerks

    Eriswerks Guest

    I'm trying to build a two transistor LED flasher, as shown at
    http://www.reprise.com/host/circuits/transistor_flasher.asp. It should
    be perfectly simple, but I must be making some mistake that I'm just
    not seeing yet.

    When I build this circuit, both LEDs light continuously rather than
    flashing. I've tried varying the values of the resistors and caps that
    should control the timing, but it doesn't make a difference. I can
    actually pluck the capacitors right out of the circuit and both LEDs
    are still lit. I've built this thing a dozen times now from several
    different people's instructions, and I get the same result every time.

    The transistors I have on hand are a little mixed... I've tried
    2N3904s, 3906s, and 4424, any of which ought to work as they're all
    NPN.

    Does anyone have any suggestions as to what stupid mistake I'm making
    here? I am starting to get sick of looking at this thing. Thanks in
    advance..!
     
  2. Andrew Holme

    Andrew Holme Guest

    Have you reversed the base and collector leads? That would light both LEDs.

    Failing that, with the capacitors temporarily removed -
    1. Try grounding one of the bases - LED should go out
    2. Try removing a 39k resistor - LED should go out
    3. Try placing wet fingers across where the 39k was - LED should light
     
  3. John Larkin

    John Larkin Guest


    This circuit has two stable states: "oscillating" is one, and "both
    transistors saturated" is the other.

    If the transistors have high beta, saturation is more likely. In this
    example, if betas are above about 80, the thing can lock up, and lots
    of transistors are hotter than that. Bigger base resistors might help,
    but that has other problems.

    If you return the tops of the 39Ks to their own transistor's
    collectors, rather than to V+, that will prevent saturation lockup.

    Or maybe something else is wrong. If you briefly short one base to
    ground, then turn it loose, it should start oscillating. If not, look
    elsewhere.

    John
     
  4. john jardine

    john jardine Guest

    Offhand it looks like the Leds should be flashing about 4 times a second.
    Ensure the 10uF capacitors have their "+" terminals connected to the
    transistor collectors. If not the caps will 'leak' and keep both transistors
    and Leds on.
    regards
    john
     
  5. Bob Eldred

    Bob Eldred Guest

    No Mistake, you have found one of the stable states of this circuit, both
    trnasitors on. This can happen when things are closely balanced. This
    circuit can be improved by adding two diodes that stop the posibility of
    both transistors coming on together.

    Disconnect the ends of the 39K resistors connected to +V on the NPN circuit.
    Connect these together. Connect the anode of a signal diode, 1N4148 or
    equiv, on the collector of each transistor. Connect the cathodes together
    and to the common point of the two 39K base resistors.

    This circuit provides base bias to the transitors by rectifying the square
    wave on the collectors of the transistors. They can't latch on because a
    both on state will rob the bias. The circuit starts because a linear
    amplification region is traversed as the bias developes. This allows the
    switching action to increase over several cycles until full operation is
    attained.

    The circuit will work equally well on both the NPN and the PNP versions of
    the circuit but the diode polarity will change. All astable multivibrators
    should have this or a similar circuit to insure operation. Check the
    literature.

    Be sure you are using the right polartiy transistors. A 3906 is a PNP part.
    Bob
     
  6. John Fields

    John Fields Guest

    ---
    You're not making any mistakes, it's a stupid circuit.

    Try this:


    +V>----+------------------------+
    | |
    [470] [470]
    |A |A
    [LED] [LED]
    | |
    +-[+10µF]--+ +-[10µF+]-+
    | \ / |
    | X |
    | / \ |
    C / \ C
    NPN B-------+ +------B NPN
    E | | E
    | [39K] [39K] |
    | | | |
    GND>---+---------+-----+--------+

    If it won't start, change one of the 39k's to 47k
     
  7. John Larkin

    John Larkin Guest


    But that circuit has a stable non-oscillating state: both transistors
    off.

    John
     
  8. John Fields

    John Fields Guest

    ---
    I think the 47k will take away that stability, but the circuit's got
    more problems than that, since I forgot the cross-coupling
    resistors.

    Here's one from GE on abse:
     
  9. Jasen Betts

    Jasen Betts Guest

    try running it off closer to 12v than 5V

    you could also try putting 1K resistors in parallel with the LEDs.


    here's my LED flasher circuit, drawn sideway's because it's easier to draw
    diodes -|>|-- and resistors --[XXX]- running that way.

    //
    +3v--+--|>|---.
    | | OV
    |--[1K]--+----+--C E----+--
    | | |
    |--[10K]-+----|----B |
    | | | |
    | 47uF_L_ _L_+ |
    | +~T~ ~T~47uF |
    | | | |
    |--[10K]-)----+----B |
    | | |
    |--[1K]--+-------C E-----+
    | |
    | // | for the transistors practically
    +--|>|---' any small signal NPN should work

    last time I used some scrounged from an old VCR

    if the leds get stuck on try increasing the 10K resistor to 20K

    use a fixed pitch font like courier if it looks garbled.

    Bye.
    Jasen
     
  10. John Fields

    John Fields Guest

     
  11. The problem is that you used tabs in your text (0x09) and these are
    not necessarily translated the same on various readers, regardless of
    the fixed spacing of the font. When relying upon fixed spaced fonts,
    I'd recommend avoiding tabs entirely.

    Jon
     
  12. ehsjr

    ehsjr Guest

    What limits the current through the LEDs?
    Ed
     
  13. Kitchen Man

    Kitchen Man Guest

    Similarly, I like to avoid tabs when writing code. That way,
    indentations always line up neatly, regardless of the user's source
    code interface program.
     
  14. I insist on doing the same thing, as well, with what I write. There
    was a time in my earlier days when the tab actually saved enough space
    in source files that it really made a difference in one's life. That
    day is long since gone. I use a little detab facility I wrote many
    years ago when I'm faced with code written under some fixed tab stop
    mentality. I can list out the tab stops or let it default, but it
    neatly strips the tabs from the source.

    Jon
     
  15. Dan Akers

    Dan Akers Guest

    Eriswerks wrote;
    "I'm trying to build a two transistor LED flasher, as shown at
    http://www.reprise.com/host/circuits/transistor_flasher.asp. It should
    be perfectly simple, but I must be making some mistake that I'm just not
    seeing yet.
    When I build this circuit, both LEDs light continuously rather than
    flashing. I've tried varying the values of the resistors and caps that
    should control the timing, but it doesn't make a difference."
    ______________________________________
    Re;
    This is one of the first circuits I ever built back when I was kid in
    the early 70's and have used it many times sense. Your inquiry prompted
    me to put this one together on a breadboard. I built it exactly per the
    NPN schematic with a 10V supply. It worked great with a running freq or
    about 4.2 Hz. I tried several methods of reproducing your
    non-oscillating result and I couldn't get it to NOT oscillate. I tried
    varying the supply voltage from as little to 3V to 12V with no effect
    except for an increased freq and dimmer LEDs with lower voltage; it does
    stop oscillating at about 3V. I tried very slowly increasing the supply
    voltage from 0V in an attempt to "gently" charge the caps and maybe
    "fool" the circuit into the non-oscillating state; no dice. I tried
    switching the supply voltage off and on in an attempt to achieve your
    stable state; again no luck. I tried swapping transistors out to no
    avail. I tried two different LEDs and it still kept going. I tried
    varying values of caps and got differing freqs and with different cap
    values, I got a proportional duty cycle of the transistors.
    So maybe you've got a wiring problem? Did you pay particular attention
    to the tie and NON tie points on the schematic? Make sure that the base
    of each NPN transistor is tied to V+ via the 39K resistor and that each
    transistor's base is tied to the opposite transistors collector via a
    10uF cap. The schematic may be confusing; THERE IS NO DIRECT TIE
    BETWEEN THE BASES OF THE TRANSISTORS NOR IS THERE A TIE BETWEEN A
    TRANSISTOR'S COLLECTOR AND IT'S BASE VIA ONE OF THE CAPS. The tie
    points are schematically represented by a heavy "dot" (see where 39K
    resistors tie to bases of transistors); lines that simply cross are not
    connected (see "cross over" of the bases in the middle of the schematic;
    no connection there). Good luck. I hope that helps...

    -Dan Akers
     
  16. Dan Akers

    Dan Akers Guest

    John wrote;
    "If you return the tops of the 39Ks to their own transistor's
    collectors, rather than to V+, that will prevent saturation lockup.
    _____________________________________
    Re;
    The circuit still works in this configuration, however, the LED's do not
    fully extinguish during their respective off cycles.

    -Dan Akers
     
  17. What does "high beta" mean?
     
  18. John Fields

    John Fields Guest

     
  19. Jasen Betts

    Jasen Betts Guest

    AIUI "beta" is another (archaic?) name for "Hfe",

    So, high gain.


    Bye.
    Jasen
     
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

-