Improved Simple PI - all BJT front end for greater sensistivity

[Beenthereuk]

No, it is true because the resistance is real (e.g. CONSTANT AND NOT A FUNCTION OF FREQUENCY) and the impedance is extracted from the imaginary part of the value. The coil's DC resistance is not affecting the measurements here, the coil's reactance (which is also imaginary and a function of frequency) will have an effect, which is why you have to measure at the lowest frequency, because XL rises with F and the greater it is, the more it interferes with the capacitive reactance. I would also use a larger blocking capacitor so that the ratio of parasitic capacitance to blocking capacitance is greater and the error smaller.

It will be very hard to measure small parasitic capacitances with this method but it should be OK for larger values. The lowest frequency meter will give the best value.

[Beenthereuk]

Also, if the value approximates the blocking capacitor value at low frequencies, record it then measure the blocking capacitor on its own and subtract the two values for the parasitic value. It all depends on the accuracy of the meter, the frequency used and the size of the parasitics.

[Carl-NC]

While this trick generally works for series caps, I suspect using it to measure a coil's parasitic C will depend greatly on the specific meter design. With my BK 879B, it definitely doesn't work. I get 465nF using a 470nF series cap, and 9.8uF for a 10uF series cap, with a coil that is probably 300pF or so.

If you run your coil in a PI circuit and remove the damping R, what is the resulting frequency of the flyback ringing?

[Beenthereuk]

That is so, it cannot resolve small parasitic values because they are lost in the reactance of the coil. I've played about with this software that I found recently (works like Microwave Office - without the license fee!) and it is clear that for it to be accurate it is best to measure at two very separate frequencies (see below soon). Funny thing is, it worked well for my large coil - I've yet to measure the frequency of the ring on it.

[Beenthereuk]

OK, from simulation, the only way to extract both parameters accurately (other than solving for each at resonance) is to extract L at f << fres and C at f >> fres.
Therefore it follows that an LCR meter capable of measuring 2 very different frequencies (high and low) could extract even the parasitics accurately. But a normal meter is likely to only approximate it given a large coil.

Notice how the real part of the circuit, Z11real = 0.8 Ohm, is constant and does not vary with frequency - it plays no part in any of this.

[Waikiki_Sweep]

Even if that method of using LC meter and big capacitor to measure capacitance of coil may have some error I think it is good for relative comparison of coils.
For example you can change shape of winding (flat spiral, basket) or spacing between wires, insulation, shielding
and spacer between shield and coil and every time check how it changing coil capacitance using that fast and efficient method.

I see error of measuring is rising with higher frequency but with low frequency LC meters results are close. So check the frequency and use LC meters which operating on less than 1kH.