Power Zones in Battery-powered Device - Switch power rail or ground?

Discussion in 'The Projects Forum' started by Phobbes, Feb 26, 2010.

  1. Phobbes

    Thread Starter New Member

    Feb 26, 2010
    Hey Forum,

    I'm working on a low-power design to run off of a AAA battery. Not all parts of the circuit have to be on all the time, so I'd like to turn off the parts that I'm not using to save power.

    This is somewhat complicated by the fact that I have both 5.5V and 3.3V power rails running off of boost converters powered by the battery.

    My microcontroller is running off of the 3.3V line, and I need it to switch off some 5.5V-powered sensors.

    The initial design(not designed by me) had a PFET on each sensor's 5.5V supply line, intended to be able to cut power by bringing the gate high. The designer overlooked that we were applying a 3.3V logic signal, so even when the gate is "high" there's still a 2.2V V_gs, so the FET never turns off.

    I have two possible solutions that I'm looking for feedback on:

    1) Use an NFET & pullup resistor to create basically a 3.3V-to-5.5V level-shifting inverter to drive the PFET

    2) Use an NFET to switch the ground terminal on the sensor instead of switching the 5.5V

    Option 2 is simpler and I can't think of a reason why it wouldn't work, but it seems a counter-intuitive to be switching ground instead of the positive supply. Is there a common best-practice for creating individually-switchable power zones?

  2. SgtWookie


    Jul 17, 2007
    If you switch the ground to the sensor, won't the output float to +5.5v when ground is disconnected? That may not be very good for the sensors or the analog inputs on your uC, if you're using it to monitor the sensor outputs. If the uC's inputs are protected by clamping diodes, you may wind up with a current path from +5.5v through the sensor's output to the +3.3v supply via the clamping diodes.
  3. Phobbes

    Thread Starter New Member

    Feb 26, 2010
    The outputs of the sensors are AC-coupled, so there won't be a DC path from 5.5V to 3.3, but that's a good point, I hadn't really been considering what would happen to the sensor output.