digital switch that conducts unpowered?

Discussion in 'The Projects Forum' started by brumac57, Apr 10, 2008.

  1. brumac57

    Thread Starter Member

    Apr 10, 2008
    18
    0
    Hi all,

    I'm building a battery powered (1 AA and 5v step up) remote control for a radio. The Press To Talk (PTT) is controlled by an AVR driving a MAX323. Pulling the PTT line above 20k ohms turns it off, below 2k ohms to ground turns it on. The MAX323 just makes or breaks the PTT line to ground.

    I want to add a fail safe PTT directly wired to one of the switch buttons on the remote so that if the battery is dead or removed the PTT can still be operated. The proposed button is also used for other functions (when the battery is good) so I have to disable the hardwired PTT to prevent activation.

    I tried using a MAX325 that has one NC (normally closed) and one NO line, intending to open the NC line when powered and hoping that it would close when unpowered. But it was open when unpowered.

    A NC reed switch would work but would draw too much power.

    What I need is a simple circuit that when powered disconnects (or blocks by changing to high resistance) the hardwired PTT and when unpowered allows it to operate. Preferably requires _very_ low current to operate.

    The PTT current being switched is tiny.


    Thanks in advance
     
  2. scubasteve_911

    Senior Member

    Dec 27, 2007
    1,202
    1
    I'm not really sure if I understand the problem, since it's 230am here and I should be sleeping.. Could you use a simple optoisolator? Maybe run the PTT signal to the top of the phototransistor through an appropriate resistor, then when power is normal you can put enough current through the LED to pull down the output?

    Maybe I don't understand the problem..

    Steve
     
  3. brumac57

    Thread Starter Member

    Apr 10, 2008
    18
    0
    Thanks
    I'm not sure the opto idea would do what I want.
    Basically I want some sort of switch circuit that conducts (or has low resistance) when no power is applied to it (the control side of the switch )
    and
    it doesn't conduct (or has high resistance) when power is applied to it

    sort of the _reverse_ of a normal transistor switch eg. when you apply power to the base the transistor conducts
    no power it doesn't

    I need it the other way

    cheers
    Bruce
     
  4. nanovate

    Distinguished Member

    May 7, 2007
    665
    1
    Try looking at a JFET or depletion mode MOSFETs.

    Supertex and Vishay make them
     
  5. brumac57

    Thread Starter Member

    Apr 10, 2008
    18
    0
    Hey thanks!

    I'm more of a firmware/software guy so I'm a bit out of my depth with this stuff.

    Any pointers/suggestions to how I might use a JFET etc in a practical circuit.

    Also, as I explained in my first post when the battery in the controller is dead or removed there will be NO power available to the switching circuit, will the JFET etc still "work" and allow flow?

    That was the beauty of the NC reed switch idea, because it allowed current flow when there was no power, but it draws too much power to keep it open for long periods.

    Thanks!
     
  6. rwmoekoe

    Active Member

    Mar 1, 2007
    172
    0
    I believe there's no such thing as a normally closed (when unpowered) digital switch.
    But at least you could dig up more on the relay thing. try finding the smallest tiniest relay switch there is, and / or wind yourself a very high impedance magnetic coil just enough to disengage the relay. I guess maybe you could get it done at a minimum current of 1 to 10mA quiescent.

    to get it lower, it is possible, but the initial current to move the relay would still be high enough. this could be done with a capacitor in parallel with the quiescent resistor.

    vcc
    |
    |
    relay
    |
    |------------
    | |
    resistor capacitor
    | |
    |------------
    |
    transistor
    |
    |
    ground
     
  7. brumac57

    Thread Starter Member

    Apr 10, 2008
    18
    0
    Thanks rwmoekoe,
    I have previously got an NO reed relay down to a couple of ma activation (for a different project) but even that is too much for long term (days) use. One idea I had, but need to test, is opening the reed relay as soon as the AVR sees the button switch has been pressed, that way I could save power by disconnecting the hardwired PTT only as needed, hopefully that would be quick enough that the PTT on the radio doesn't respond (otherwise you would get a click everytime the alternate use button was pressed) I could only determine this by experimentation, also I haven't managed to source a NO reed switch, my usual supplier Farnell here in Australia doesn't have them, anyone have an idea for a supplier?

    Interesting that there is no NC non-powered digital switch!

    Are JFET or depletion mode MOSFETs still a possibility? - anyone?

    Thanks to you all!
     
  8. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    Well, JFETs will conduct until the gate is pulled several volts negative of the source.
    I was just fiddling with a 2N4416 JFET in a sim, with the Vgs at 0v, and Rds was about 33 Ohms. With Vgs @ -6v, Rds was roughly 167 megohms.

    A 2N4416 can carry >150mA at up to 30v.

    The problem would be the necessity of providing a source of -6v to turn the JFET off. Current requirements for the -6v would be around 22pA (yes, pico Amperes).
     
  9. nanovate

    Distinguished Member

    May 7, 2007
    665
    1
    There are also p-ch jfets that have a positive cut-off voltage.

    2N5461
     
  10. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    That's an idea. I didn't have a 2N5461 in my (limited) pspice library, but I did have a 2N5460. If the jfet is used as a current sink, with Vgs=0 it conducts, but in order to fully turn it off, the gate needed to be > 3.1v above the Vdd of the load. Got similar results if it's the current source, but then Vgs needed to be roughly 3.3v - or about 3.3v above the supply voltage.

    Maybe someone else has another P-ch jfet to model?
     
  11. rwmoekoe

    Active Member

    Mar 1, 2007
    172
    0
    so there is a normally closed digital switch :)... it sure is a good idea.
     
  12. nanovate

    Distinguished Member

    May 7, 2007
    665
    1
    One thing to note is that the current handling of a JFET is much lower relative to other devices like MOSFETs and BJTs.
     
  13. brumac57

    Thread Starter Member

    Apr 10, 2008
    18
    0
    Hi all,
    I just got up (7am here in Oz) to find that you had been discussing the problem, thanks.

    Sounds good - 22pA is the sort of draw I'm looking for, that's almost below the natural discharge rate of the battery?! also the AVR gets down around that in sleep mode so I can handle that sort of drain easily

    the only source of power I have on board is the AA 1.5v or maybe an A123 3v followed by the 5v on the other side of the step up.


    getting the -V would be a problem? or add to the complexity could I somehow get it out of my current power setup ?

    I worked out that the reed relay I need is either a Form B or Form C, which Farnell does actually carry.

    The problem with the relay is the power consumption - especially if they have to be driven all the time the battery is good.

    Cheers
    Bruce
     
  14. brumac57

    Thread Starter Member

    Apr 10, 2008
    18
    0
    "One thing to note is that the current handling of a JFET is much lower relative to other devices like MOSFETs and BJTs."

    The current is no problem , I just checked the PTT by using my microameter to short the ground line to the PTT line - a big fat Zero, it just needs to see the low resistance to ground.

    the -V and the 3V over supply requirements seem like a problem. I checked the voltage across the PTT line
    2.9V (the radio runs on 2 AAs) does that help at all?

    Cheers
    Bruce
     
  15. John Luciani

    Active Member

    Apr 3, 2007
    477
    0
    As suggested the depletion mode MOSFET will work.

    CP Clare makes a Form B solid state relay (normally closed). I am not sure of the price.

    http://www.clare.com/Products/SSR.htm

    (* jcl *)
     
Loading...