Measuring leakage inductance of flyback transformer

We all know that in say an offline flyback transformer, we measure the leakage inductance by shorting the secondary and measuring the inductance at the primary.

However, the leakage inductance is typically very small, and simple cheap inductance meters (like the LCR40, below) cannot get an accurate value for leakage inductance like this.

There is another, better way to measure the leakage inductance…that is shown on the top of page 4 of the following……

http://citeseerx.ist.psu.edu/viewdoc...=rep1&type=pdf

[Doc title = "Analysis and Design of a Single-Phase Tapped-Coupled-Inductor Boost DC–DC Converter"]

Inductance factor k = {L(aid) – L(opp)} / [4 * SQRT(L1L2)]

Then get leakage inductance from L(leak) = L1 (1-k)

So why does no-one measure leakage inductance like this?

LCR40 inductance meter

https://uk.farnell.com/peak/lcr40/at...st=lcr%20meter

Re: Measuring leakage inductance of flyback transformer

Measuring open circuit impedances will be much more susceptible to parasitic capacitances, I think.

And accurate LCR meters aren't expensive anymore. I use this and it works fine for hundreds of nanohenries and up. Good enough that I would confidently use it to bin 1% tolerance parts.

Re: Measuring leakage inductance of flyback transformer

Best way is the resonance test, choose a good quality, hi Q, cap, say 470nF 400V ( but can be any good quality cap )

solder in parallel with the terminals to be measured ( for Lpri-leak across the pri with the sec's shorted )

put a 4k7 resistor in series with this combo

hook up to a sig gen, with a scope on the sig gen o/p and a 100x probe across the LC parallel ckt

vary the freq until you see the resonant rise across the LC and it is in phase with the sig gen volts out

Ckt now at resonance - you know C, and F, hence L = 1/ (( 2.pi.F)^2 . C )

p.s. this is a very accurate test, and can be done at any freq of interest by choosing C

Also if you measure the 3dB points either side of resonance ( Vpk x 0.707 ) you can measure the Q accurately ( Q = Fr /( Fhi - Flow) ) - but bump the R up to 47k and make sure the Q of the cap is > 600 ....

happy measuring ...

p.s. once you know Lpri, by the same method ( unshort the sec's ) you can then take the freq way up until you get the SRF, then you can calc the capacitance of the pri easily at you know Lpri & the SRF ...