Leakage inductance / General Science and Electronics / Forums

Forums

4hv.org :: Forums :: General Science and Electronics

« Previous topic | Next topic »

Leakage inductance

Move Thread

LAN_403

nrhoades

Tue Mar 13 2012, 06:53PM

Registered Member #610 Joined: Wed Mar 28 2007, 09:44PM
Location: Middletown, RI
Posts: 110

Impedance power matching: An engineer today told me that, for a known load, leakage inductance in a transformer will overheat the transformer core. I briefly thought about this and I don't agree.

Leakage inductance simply represents flux that is not linked to any other isolated windings. It is equivalent to adding voltage divider (1-k)*Lp : k*Lp, and then (1-k)*Lp in series with an ideal transformer with an impedance ratio of Ls/Lp.

The alternative is a tightly, bifilar wound 1:1 transformer (k=1) with in inductor Lm in series with the primary. This leads to adding a voltage divider Lm : Ls to an ideal transformer with an impedance ratio of 1:1.

The input impedance seen by the ideal transformer of both cases will limit the current at the frequency of interest. Therefore, in both cases, the current may be properly limited from creating excess magnetic fields in the core that will cause a lot of heating, or even saturation.

The only advantage is that it is easier to design by using Lm, and Lm can probably be made larger than any leakage inductance can as to provide higher matching ratios.

Am I missing something?

Steve Conner

Tue Mar 13 2012, 08:33PM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

Yes you are, it's not current that overheats transformer cores, it's voltage.

If the leakage flux is concentrated in some part of the core due to the winding geometry, it could make that part hotter than the rest. Although, you would hope that in a good transformer design the leakage path would contain a big air gap, limiting the leakage flux density.

Also, leakage inductance can be deliberately resonated to get very high voltages. These correspond to lots of volts per turn, which means lots of core heating.

nrhoades

Tue Mar 13 2012, 11:35PM

Registered Member #610 Joined: Wed Mar 28 2007, 09:44PM
Location: Middletown, RI
Posts: 110

Toroids don't have air gaps though. If you had poorly coupled and physically separated primary and secondary windings on a toroid, I would imagine that any flux that gets into the toroid will make it to the secondary, via the very low magnetic path due to the geometry and permeability of the toroid. I can't see that much flux jumping through the center of the toroid and missing the "other side" completely.

But with an E-core with tight and physically separated windings, a higher ratio of "core leakage" to "winding leakage" exists due to less flux linking through the loose air in the primary winding to more flux jumping across the center of the core. Then I could see the need to put a gap on one side.

The peak field in gauss is approximated by B = V/(4.44*f*A*N) if V is the peak of a sinusoid. This is when I plug V = jwLI into the B and L equations. I can see that voltage has the dependence.

Is there, then, for impedance matching, any advantages for using low leakage 1:1 transformers with a separate ballast inductance on the primary over step-up/down transformers with leakage inductance?

Steve Conner

Wed Mar 14 2012, 08:20AM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

I'm not sure of the theory, but in practice, when Steve Ward et al. were building SLR power supplies, they found that they got better results using a separate air-cored inductor for the resonant tank circuit. I think Steve tried using the leakage inductance of the transformer and something bad happened.

radiotech

Thu Mar 15 2012, 12:29AM

Registered Member #2463 Joined: Wed Nov 11 2009, 03:49AM
Location:
Posts: 1546

As to the heating of the core, the eddy current and hysteresis losses are responsible for
the iron losses. If the transformer is designed around a certain efficiency for a constant
wattage load then the magnitude of the primary and secondary leakages would mean
a higher primary voltage, (causing higher iron losses) would be needed to furnish the
power output of the secondary circuit. Perhaps this is what your prof meant.

teravolt

Thu Mar 15 2012, 03:22AM

Registered Member #195 Joined: Fri Feb 17 2006, 08:27PM
Location: Berkeley, ca.
Posts: 1111

hi nrhoades, are you trying to do power conversion or drive IGBT gates. Some times I have made gate drivers and the core got quite hot. the type of material that it is made out of makes a big difference. I am hopping some one will enlighten us. It is my understanding that powdered iron is more for EMI suppression and ferrite which has more silicon and is better for RF work and each type of material have different operating band widths. it gets even more complicated because with temperature changes the permeability. I'm also guessing that powdered iron is more conductive and eddy and circulating currents are worse making it less efficient at higher frequencies.

hi Steve I agree I have tried making a SLR with only the leakage reactants and I have had a hell of a time making it stable and predictable.I think pert of the problem is parasitic and strange other reactants. most typical examples I've seen use a separate inductor or LC in series with the primary.

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob