1
\$\begingroup\$

If I understand correctly it would flatten the hysteresis curve. So it would allow to store more energy and it would also decrease the permeability of the core. Would it have another effect or will some type of compensation be needed in the system?

\$\endgroup\$
0
\$\begingroup\$

Just thinking this through, without having done it: it would cut the inductance per turn roughly in half. So to store the same energy you'd have to increase the current by \$\sqrt 2\$, or you'd have to increase the number of turns by \$\sqrt 2\$. Actually bringing the core up to the same value of H as before (and thus using that "extra" energy storage in the gap) would require roughly doubling the current, or you'd have to double the number of turns.

Assuming that the coil winding volume is already full, just doubling the number of turns would increase the winding resistance by a factor of four. Similarly, increasing by \$\sqrt 2\$ would double the winding resistance.

So you'd have a definite trade-off between the energy lost to switching and/or coil resistance vs. the energy the coil could hold in each cycle. I suspect that for any given size of core, and assuming copper wire, there's an optimum that depends on the switching frequency, and the capabilities of the switches you have available.

\$\endgroup\$
  • \$\begingroup\$ Also leakage would increase, requiring higher power snubber and resulting in lower efficiency \$\endgroup\$ – JonRB Aug 25 at 16:53
  • 1
    \$\begingroup\$ Not to mention fringing effects from the larger gap, which causes eddy current losses in the windings and any surrounding conductors. \$\endgroup\$ – John D Aug 25 at 17:42
  • \$\begingroup\$ Increasing the air gap has almost no effect on the leakage inductance. Try on a transformer to measure the leakage inductance with and without the core, you'll see a leakage term almost unchanged. It was demonstrated some time ago already. See Transformer modeling and design for leakage control a paper published by Hsu, Middlebrook and Cuk in 1975 I think. \$\endgroup\$ – Verbal Kint Nov 3 at 10:10

Your Answer

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

Not the answer you're looking for? Browse other questions tagged or ask your own question.