SMPS Topology recommendation ?

Discussion in 'The Projects Forum' started by jsm09a, Jul 8, 2009.

  1. jsm09a

    Thread Starter New Member

    Jul 8, 2009
    I am interested in thoughts on appropriate topologies for a challenging power supply project. Requirements:

    - Wide Input voltage range (8.5-65 VDC)
    - Mulitple outputs (3.3V, 5, +/-12 and 40+)
    - Primary/Secondary isolation
    - Good cross-regulation amongst outputs
    - Low power: 2.5 watts typical, 8 watts max
    - High efficiency - 90% for 10-35 VIN, 85% elsewhere
    - Minimal conducted emissions (e.g. DO-160)
    - Minimal physical size/volume
    - 50+ millisecond hold-up time at max load

    Coupled inductor SEPIC feeding isolation transformer comes close, but secondary voltage/current stress is extreme at limits of VIN range.

    Would prefer symmetric excitation of isolation transformer to reduce stresses in secondary (and transformer volume), but conventional, voltage fed topologies (push-pull, bridge, half-bridge) have significant input ripple that would require significant filtering.

    Two stage approaches (e.g. boost to a high voltage storage cap followed by push-pull) would seem to have difficulties meeting the efficiency requirements (especially with sagging in the push-pull during hold-up time).

    Seems like the converter will need to run in both DCM and CCM, calling into question most of the conventional current-fed, full-bridge converter topologies.

    Can anyone suggest a symmetric, current-fed topology that can handle the wide DC input voltage range ?

    Thanks !
  2. millwood


    looks like a boost controller / regulator running in fly-back would be the way to go for you.
  3. jsm09a

    Thread Starter New Member

    Jul 8, 2009
    Sorry, not sure I follow the "running in fly-back" part. Are you suggesting a standard flyback topology driven by a PWM chip intended for boost circuits (flyback topology would probably not meet DO160 conducted emission without significant filtering), or that the output of a boost converter is then fed-into a flyback circuit (of some type) ? ... or perhaps, the boost converter includes a coupled inductor that also operates in flyback mode (which brings it pretty close to the existing SEPIC, I believe).