Using transistors as a switch for HV source.

Discussion in 'The Projects Forum' started by drevilz, Mar 26, 2014.

  1. drevilz

    Thread Starter New Member

    Feb 13, 2013
    Hi everyone,

    I am working on a high voltage (0 to ~150 volts and low current < 1 ma) circuit that has a DC, square wave, triangle wave, and sawtooth wave source from an op-amp.

    I tried to implement the switch (picture attached) but I keep on getting leakage current when my switch is 'off' or my base voltage is '0' volts. The leakage current is greater with a square wave input and almost non existent with a DC input. I am assuming it is due to the parasitic capacitance.

    Is there another transistor scheme that I can use to prevent this leakage current from going through? Or even a digital input IC switch that I can use? I want to stay away from relays if possible.

    Thanks a bunch!
    Last edited: Mar 26, 2014
  2. crutschow


    Mar 14, 2008
    Try adding a 10kΩ resistor between the base and emitter of the PNP. What is the PNP transistor that you are using?

    What is the load you show with the (?) mark? You need some load for the transistor to work properly.
  3. #12


    Nov 30, 2010
    All bipolar transistors leak a little bit, usually in the nanoamp range, but they get worse when they get hot. Look at page 3 of this datasheet, Icbo at 40 volts is 50 na. It's worse than that in the final result because that is "amps, collector to base", and that current can get amplified by the transistor. You just have to get used to providing a path for that current to escape with a resistor from base to emitter (if the driver doesn't provide that path).

    Your point, marked I/O, might apply a ground path for the base of the npn transistor when it is, "low". If it doesn't, you have to provide the path or the Icbo will get amplified by the npn, then amplified again by the pnp, and suddenly you have trouble.

    Just something to consider, especially at high voltages or hot places. A lot of times you can get away without these resistors, like in 5 volt logic operations. So, now you know.
  4. drevilz

    Thread Starter New Member

    Feb 13, 2013
    Hi Crutschow,

    Thank you for the reply. The '?' is my load. I got the picture from the internet. However, my circuit is similar with the exception of the base resistor of the pnp. My base resistor is 100k ohms and my load is a resistor (~200k) and capacitor (200 pF) in parallel (actually it's skin tissue with varying resistances).

    The pnp transistor I am using is a high voltage general purpose transistor (ksa 1381, (BJT) - Single/KSA1381ESTU-datasheetz.html) and the npn is a general purpose high voltage transistor as well.

    A triangle wave has less leakage current flowing through then the square wave.

    Thanks again!
  5. drevilz

    Thread Starter New Member

    Feb 13, 2013
    Hi #12,

    My transistor is not getting hot and the voltage across my load is virtually 0 volts when the source (Vcc) is a dc source. However, with a square wave source the voltage across my load is about 50% of my source when The switch is 'off'.

    The reason I use the npn is to float the base of the pnp to turn the transistor off. Since my voltage is high (~150 volts) turning the pnp off became a hassle to do so I decided to float the base to turn it off and ground the base to turn it on.

    Last edited: Mar 26, 2014
  6. AnalogKid

    AAC Fanatic!

    Aug 1, 2013
    What you have shown is not a linear amplifier, so no matter what you drive the NPN base with (triangle, sine, etc.), the PNP collector is going to look more like a square wave than anything else.