Bidirectional electronic switch

Discussion in 'General Electronics Chat' started by hazim, Apr 28, 2010.

  1. hazim

    Thread Starter Active Member

    Jan 3, 2008
    419
    13
    Hello,

    I want to make an electronic controlled bidirectional switch, similar to a push button switch in operation, it is to be controlled (to connect) with a pulse. I want it small and simple.

    I thought about using two similar transistors and connect them in parallel but in reverse direction (C&E - E&C - B&B), or an NPN and it's complement PNP parallel but not reversed (I tried using 2n3904 and 2n3906). Whenever a pulse comes to the base, the transistors should be connecting in both directions. I also tried to use bc548, bc109 and bc337 with bc327.. they didn't worked properly.. Any idea what transistors I should use?? I'm not caring about the current through this switch.. 20mA is good.

    I attached this picture for more clarification
    [​IMG]

    Thanks,
    Hazim
     
  2. rjenkins

    AAC Fanatic!

    Nov 6, 2005
    1,015
    69
    The thing with using two transistors in opposition it that the bases then need driving with opposite polarities from two separate supplies (each positive with respect to it's emitter, and the emitters are connected to the two different switch terminals).

    There are specific devices to do this, eg. a 'bosfet' relay:
    http://www.irf.com/product-info/datasheets/data/pvr13n.pdf

    That uses two FETs with seperate, optically isolated gate feeds.
     
    hazim likes this.
  3. rjenkins

    AAC Fanatic!

    Nov 6, 2005
    1,015
    69
    Another thought, if you are not bothered by a couple of volts drop when the switch is 'on', is simply connect to a single transistor using a bridge rectifier.

    The polarity is fixed at the DC side, and the AC side will be near a short circuit when the transistor is on - just a couple of diode voltage drops.
     
  4. hazim

    Thread Starter Active Member

    Jan 3, 2008
    419
    13
    Thanks rjenkins. I can't figure out how the connection should be between the transistor and the bridge rectifier. The supply voltage of the main circuit is 5VDC so I think your second idea doesn't work?

    I'm looking to do it with a simple circuit if possible. I can't get any devices other than than what I have, and also the area on the PCB where I can put this switch circuit is too small and constricted to even less than an 8-pin dip IC place, this is why I'm looking toward doing it with 2 or 3 transistors..
     
  5. Ron H

    AAC Fanatic!

    Apr 14, 2005
    7,050
    656
    What are actually connected to terminals 1 and 2, in terms of voltages and impedances?
     
  6. hazim

    Thread Starter Active Member

    Jan 3, 2008
    419
    13
    I just want to connect them in place of the momentary push button switch in a circuit similar to this one:
    [​IMG]
     
  7. Ron H

    AAC Fanatic!

    Apr 14, 2005
    7,050
    656
    What does your control signal look like? Where does it come from?
     
  8. naizmenic

    New Member

    Apr 27, 2010
    12
    0
    if that's all, then logic controlling your switch actually could control LEDs itself without NE555
     
  9. hazim

    Thread Starter Active Member

    Jan 3, 2008
    419
    13
    It is a pulse of about 3.5V.. Its width can be adjusted, it's coming from the output of a 555 timer in monostable operation with Vcc=5V or 6V.
     
  10. hazim

    Thread Starter Active Member

    Jan 3, 2008
    419
    13
    I just want to use this circuit above to control something else... I want to use it as it is (latching switch? I'm not sure about the name). No that's not all :)
     
  11. naizmenic

    New Member

    Apr 27, 2010
    12
    0
    If you understand how this work then you'll see it is unnecessary. Whatever your controlling signal for switch is, it could be (easy) powerful enough to drive LEDs without 555 and without switch.

    If you do this only for practice then:

    1. you should define how your controlling signal will trigger the switch (on every rising, or every falling edge or maybe you want high level corresponds to switch "turn on" state and low level correspond to "turn off" state)

    If you want the simple solution assume high level of your signal means turn ON and low level means turn off. For such realization you can use (in general) mosfets (without diode) use both N channel and connect source from one to drain to other. Those two will be your terminals. Connect gates together and it will be input for your controlling signal.

    edit:
    if you want to use such switch in other design take care on current through it. In this design current is limited by resistors R3 and R4. Also if your controling signal could fly (has tri-state) put pull-down resistor to gates to keep switch off.
     
    Last edited: Apr 28, 2010
  12. hazim

    Thread Starter Active Member

    Jan 3, 2008
    419
    13
    Don't care about the LEDs, my aim is not to turn LEDs on and off :)
    Your suggestion for using two mosfets is similar to what I said in the first post (see the attachment there). You suggested using mosfets instead of BJTs. I think this may be a solution but the reply to my suggestion (on the first page) is applicable with your suggestion too I see.
     
  13. naizmenic

    New Member

    Apr 27, 2010
    12
    0
    In your first post you are using DIFFERENT types of components. That is NOT what do you ask.

    By your first schematic if your controlling signal is low, only P channel would be open and only if positive side is on terminal 2. But if positive side is on terminal 1 such "switch" doesn't work. Same thing is if controlling signal is high. Only N channel is open and only if positive side is on terminal 1. And that is not what real switch does.

    Get it?

    Schematic with 555 works like this:
    Suppose your RED led lit. It means output from 555 is low and capacitor is discharged through R3. By pressing a button it's right terminal is then connected to low level and left terminal to positive level. Capacitor will charge through R4 while switch is closed and it will decrease voltage on input of 555 which in turn will toggle its output to high level. That will turn on green led (off course red will be turned off). Now by pressing switch positive voltage is on it's right side etc.

    Your first schematic can not allow that by unified controlling signal (means controlling logic is same for all cases and not depends on polarity on terminals)

    edit:

    I forgot to say, if you prefer design from you first schematic to allow it works you'll need to separate gates and use one inverter to connect them. Then it will works. So your controlling signal will be directly connected to let say N channel and inverting signal to P channel. Then high level on it will open both sides and low level will close borh sides and that is what you want.
     
    Last edited: Apr 28, 2010
  14. hazim

    Thread Starter Active Member

    Jan 3, 2008
    419
    13
    I tried mosfets as naizmenic suggested but it didn't worked.
    Any other idea?
     
  15. naizmenic

    New Member

    Apr 27, 2010
    12
    0
    Than you made some mistake. Can you provide exactly schematic what did you do?
     
  16. hazim

    Thread Starter Active Member

    Jan 3, 2008
    419
    13
    Here is the schematic..

    [​IMG]
     
  17. rjenkins

    AAC Fanatic!

    Nov 6, 2005
    1,015
    69
    As I mentioned earlier, when using two devices on opposite polarities you need two control signals of opposite polarities, as each device needs it's own base or gate supply relative to it's own emitter or source.

    Connecting the two bases or gates to the same signal guarantees it will not work.

    If using FETs, the 'On' gate control voltage must be several volts more positive that the highest level of the controlled voltage, otherwise the FET will not stay 'on'. Likewise, the gate Off voltage must be lower than the most negative voltage of the controlled voltage.

    The gate voltage polarities are with respect to the source of the FET that has it's drain positive with respect to it's source *at that instant*, so in absolute terms positive for Q1 and negative for Q2 as shown in your diagram.

    If using bipolar transistors, you need either high value base feed resistors or two completely floating base supplies, as the base-emitter diodes of the transistors effectively cross-connect the control side with the output side.
     
  18. hazim

    Thread Starter Active Member

    Jan 3, 2008
    419
    13
    Yes you are right. I said that to naizmenic in post #12 but he insist on his view.
    Ok. Then what to do? isn't there any simple way to do it??!
     
  19. tomkonikkara

    New Member

    Mar 24, 2010
    1
    0
    Quad analogue switch CD4066 or CD4016 is useful for this?
     
  20. hazim

    Thread Starter Active Member

    Jan 3, 2008
    419
    13
    I have CD066 IC, I can't find from the datasheet if it does function as I want or not. There should be a small 4 pin element that functions the same as a magnetic relay, but in a much smaller scale.. a 200mA relay IC for example. A 'standard' relay can holds 10 or 15A..

    I'm still thinking and waiting for a solution :confused:
     
Loading...