Posted by SillyOldDuffer on 20/11/2018 12:56:56:
Not sure how it works. I believe it puts an oscillator and a known capacitor across the unknown inductance, measures the resulting frequency, and calculates the inductance from that
Nothing wrong with that, it's a simple way of measuring inductance. It will certainly be using AC, otherwise the frequency will be zero!
Even quite small gaps cause the inductance to fall significantly. For items like pot cores that come with ready ground gaps for power applications the gaps are on the order of a few tens of thou at most. It's why contactors chatter when there are small specks of dirt between the poles; even a few thou can have an effect.
The effect of DC on an inductor opens a can of worms. The simple answer is that the inductance is irrelevant given a DC current. The SoD original tests show that, after some time the current settles down to a DC level determined purely by the applied voltage and resistance of the coil. Inductance only comes into play when the current is changing.
A mixture of DC and AC currents is rather more complicated. The effect on the inductance, and the AC current, will depend upon the shape of the B-H curve for the core material, as the DC current will bias the operating point. There are two other possible effects of a DC current. One, the DC current may magnetise the core material to some extent. It's how permanent magnets are made – expose them to a strong magnetic field. To demagnetise you expose the material to an alternating magnetic field, with no DC bias. Two, at higher DC currents a magnetic field is produced that can saturate the magnetic core, ie, an increase in current does not produce a proportional increase in magnetic flux density. Effectively the inductance falls rapidly, although not quite to zero.
Small loose wound air cored inductors are pretty close to an ideal inductor. Once you start introducing cores all bets are off, and the inductor becomes far from ideal. Of the passive components inductors are the most difficult to understand and design, and often require on the bench refinement – aka trial and error!
I bet that lot will put people to sleep more effectively than counting sheep!
Andrew