I am designing a SEPIC converter, powered by 3.7 V Li-Ion cell. Frequency 500kHz, Transformer 10uH per side, 1:1 ratio. Ferrite material #87. Output to be variable by adjusting duty cycle.

It has good efficiency over 80% at 7W. But when I push it to the design goal of 10W, efficiency drops too much. Tested with a load of 10 ohms and 39 ohms. When duty cycle adjusted to increase output using either load, the the transformer core and MOSFET heat up quickly going over 7W. Looking at the MOSFET drain with a current transformer, I was expecting to see a saturating core with current spiking, or rising drain-source voltage. But no. Observing the drain current wave on the scope, the inductive charge ramp rises as expected with increased output current. As the voltage to the load is increased by increasing the duty cycle, the inductor current ramp flattens out, and even the slope reverses! What is the cause?Normal inductor current ramp at 7W VS flattening of inductor current at 10W

  • \$\begingroup\$ Your drawing describes a flyback transformer rather than a SEPIC. A SEPIC doesn't have a coupled transformer. You should provide the gate driver wave-form as well as the waveform between the mosfet and the transformer. It may be helpful to see the output side of the transformer as well. Without that info, it'd be hard to determine what the issue is. \$\endgroup\$ – horta Jan 13 '15 at 15:48
  • 1
    \$\begingroup\$ Actually it is a coupled-inductor SEPIC- The two inductors in a SEPIC are often wound on the same core: eetimes.com/document.asp?doc_id=1279541 \$\endgroup\$ – John D Jan 13 '15 at 15:53
  • \$\begingroup\$ @JohnD Ah gotcha, always learn something new here. \$\endgroup\$ – horta Jan 13 '15 at 15:58
  • \$\begingroup\$ If your drain current isn't ramping it seems like you're not getting enough voltage across the primary inductance. Too much ESR? \$\endgroup\$ – John D Jan 13 '15 at 16:18
  • \$\begingroup\$ Too many unknowns to answer this. But, coupled inductor with N87 core, if there is no gap, will saturate. N87 can't take any DC bias without a gap. Measuring current with current xfmr where you are will not do, since there is DC current there to saturate that as well, so measurement is suspect. Also if you could trust measurement, it would not tell you much about inductor currents since both input and output currents are seen by FET drain. Numbers don't add up, duty cycle 50% is Vin=Vout and 3.7V across 10 Ohms is not 7W. \$\endgroup\$ – gsills Jan 15 '15 at 4:25

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.