Short-duration power dissipation in MOSFET/SCR

Discussion in 'General Electronics Chat' started by Veracohr, Jan 28, 2013.

  1. Veracohr

    Thread Starter Well-Known Member

    Jan 3, 2011
    I'm trying to choose a MOSFET or SCR to use as a switch to discharge a capacitor for a relaxation oscillator. I've got a question regarding the power dissipation requirements.

    In the simulation I'm doing (just for testing this specific concern, not a full design), I get a peak power dissipation of about 1.4W. The full discharge phase is about 5μs but only about 2μs of that are above 700mW.

    Not all datasheets show the same information, but some of them show higher max values for current and power over short durations. For instance, NTR4170N gives these ratings:

    - .48W (steady state)
    - .82W (t<30s)
    - 1.25W (t<10s)
    - Continuous drain current 3.9A (t<10s)
    - Pulsed drain current 8A (t=10μs)

    And 2N6027 gives this:

    - 300mW (only Pd rating given)
    - DC forward anode current 150mA
    - Repetitive peak forward current 2A (20μs pulse width)
    - Non-repetitive peak forward current 5A (10μs pulse width)

    Clearly these things can handle more current and/or power at the shorter time periods I'm aiming for, but the datasheets don't ever seem to give both current and power capability in the same time range. In the first example above, I assume it can handle even more power at 10μs than it can at 10s, but it doesn't say. And if I take the maximum power and maximum pulsed current ratings, I get a voltage of 156mV. I assume they're typically operated at higher Vds than that and thus the 10μs power capability should be higher. In the second I assume it can handle more than 300mW over shorter times, but it only gives the one rating.

    How would I go about choosing one?
  2. kubeek

    AAC Fanatic!

    Sep 20, 2005
    I´d say everything depends on the lenth of the off period and the total energy dissipated during the on period. Then the size of the transistor and it´s thermal capacity could tell you if it will survive or not.
  3. crutschow


    Mar 14, 2008
    If you multiply the pulse current times the maximum ON resistance of the transistor at that current, then you will have an estimate of the pulse power it can handle.
  4. Veracohr

    Thread Starter Well-Known Member

    Jan 3, 2011
    Oh yeah, sometimes I forget power is a time-referenced measurement. So if I have a roughly triangle shaped pulse of 4.5μs that peaks at 1.5W (calculated as Vds*Id), I get a total power of 3.375μW? That sounds awfully low. The off period would be 235μs-36ms.

    Not I^2 * Ron?