Fast Power Switching

Discussion in 'General Electronics Chat' started by skov, Sep 30, 2013.

  1. skov

    Thread Starter New Member

    Sep 30, 2013
    Hi All,

    I'm currently building a circuit for an experiment I'm working on.

    All it needs to do is switch a load on and off. The only caveat is that it needs to do it quite fast!

    The input to my circuit is a TTL squarewave.
    The load is a small laser requiring up to 1A at 12V.

    When I say it needs to be fast, the requirement is that the rising and falling edges on the output need to be as short as possible, ideally less than 5ns.
    The total switching time (input changing to output changing) isn't too important and I can tolerate resonably long switching delays as long as the switching edges are nice and short.

    I'm currently using an n-channel power MOSFET, driven by a MOSFET driver IC, to switch the laser on and off.
    The fastest edge time I've managed using this circuit is ~15ns, and I can only get a fast switch-off edge by sticking a big resistor in parrallel with the laser which seems a little inelegant!

    So... am I heading along the right lines using a MOSFET as a low-side driver? or is there a better way of doing it?

  2. wayneh


    Sep 9, 2010
    I think you're using the right approach but perhaps you need a different driver and/or a MOSFET with a small gate charge and a lower gate inductance. Can you supply a photo of your build? Fast operation requires attention to design, eg. short traces with low inductance. The experts here may be able to give you some pointers.

    Can you 'scope your edges? I wonder if you're seeing inductive ringing.
  3. ronv

    AAC Fanatic!

    Nov 12, 2008
    How are you limiting the current?
    Do you have any specs on the diode?
    Does it run at a fixed frequency or random pulses?
  4. skov

    Thread Starter New Member

    Sep 30, 2013
    Thanks guys. I think I'll probably persevere with what I've got.

    The input signal is a fixed frequency square wave.

    The circuit is currently built on veroboard, which isn't ideal. I did however put some thought into the layout to keep the tracks as short as possible.
    The MOSFET is also bigger than it needs to be.

    I'll be laying out a proper PCB for it shortly, just wondered if there was a better approach before I did.
  5. crutschow


    Mar 14, 2008
    Use a ground plane on the PCB, keep all traces short carrying high frequency signals, and don't use a MOSFET bigger than you need. Select one with the lowest input capacitance. If the line is more than a few inches, connect to the Laser diode used twisted pair wire.