Quality and brans on LCR meters

[Joe Gwinn]

Well, will it read AC resistance at 10K? That alone would be real useful.

Well, I have not gone through the theory, but AC resistance should still be correct, and a lot higher than at 1 KHz (as eddy currents are much stronger at higher frequencies).

So, what about the readings it WILL give, will these be useful if just doing comparative readings, say you pop in an alnico 3 magnet instead of 5, even though its not a correct reading will it give a relative reading from changing one thing to another just for comparison purposes?

I think it will be hard to make much sense of LCR meter readings taken above resonance. Getting around this problem is one reason for development of the soundcard-driven IV instrument, which will measure complex impedance versus frequency.

[David King]

Getting around this problem is one reason for development of the soundcard-driven IV instrument, which will measure complex impedance versus frequency

Joe, What does that sentence mean in the context of what we need to own or buy?

[Mike Sulzer]

Well, I have not gone through the theory, but AC resistance should still be correct, and a lot higher than at 1 KHz (as eddy currents are much stronger at higher frequencies).

A single frequency impedance gives two quantities: impedance magnitude and phase (or the equivalent real and imaginary parts). If you have a pure reactance (L or C) with loss (equivalent to an R). Then you can determine both quantities from a single frequency measurement.

If you have three elements (L, C, and R) you cannot determine the values of all three from two numbers. Can you get any of them right?

Consider a specific case: an L with an R in series, and a C in parallel with this combination. If the C is negligible, clearly we can determine R and L: the current through the series combination is related in a simple way to the voltage across it. However, if the C is significant, it alters both the magnitude and phase of the total impedance.

The impedance of the R L combination is sL + R. That of the C is 1/(sC).
The parallel combination is 1/(sC + 1/(sL + R)) = (sL + R)/(sC + 1). This is the ratio of two impedances. The resulting impedance is the ratio of the two magnitides and the difference of the angles. Both the magnitude and angle of sC + 1 are nearly arbitrary, and so there is no way to determine either L or R.

[Possum]

...

If it will read AC resistance that is highly useful to me, I tend to use that more than inductance readings especially in buckers.

[Mike Sulzer]

Sorry, I did that algebra above wrong last night. I will try to fix it sometime today.

[Mike Sulzer]

The last paragraph in my post above with the algebra should read:

The impedance of the R L combination is sL + R. That of the C is 1/(sC).
The parallel combination is 1/(sC + 1/(sL + R)) = (sL + R)/(s2RC + sRC + 1). At each frequency f, where s = 2*pi*f*j, this impedance is the ratio of two complex numbers. The resulting impedance is the ratio of the two magnitudes and the difference of the angles of the complex numbers. Both the magnitude and angle of the denominator vary with frequency; so this is complicated. Above resonance the denominator shifts the phase to indicate a capacitance. But the ac resistance we want to measure is the losses in an inductor. I do not think that the meter can sort this out and get it right.

[Joe Gwinn]

Joe, What does that sentence mean in the context of what we need to own or buy?

Nothing. The soundcard VI instrument is a research project at present, and a slow-moving one at that. It is not in any stores just yet.

[Joe Gwinn]

The last paragraph in my post above with the algebra should read:

The impedance of the R L combination is sL + R. That of the C is 1/(sC).
The parallel combination is 1/(sC + 1/(sL + R)) = (sL + R)/(s2RC + sRC + 1). At each frequency f, where s = 2*pi*f*j, this impedance is the ratio of two complex numbers. The resulting impedance is the ratio of the two magnitudes and the difference of the angles of the complex numbers. Both the magnitude and angle of the denominator vary with frequency; so this is complicated. Above resonance the denominator shifts the phase to indicate a capacitance. But the ac resistance we want to measure is the losses in an inductor. I do not think that the meter can sort this out and get it right.

No, it can't. The meter will nonetheless report a inductance or capacitance plus a AC resistance, according to what equivalent circuit the user has by pressing buttons told the meter to assume.

Now AC resistance is a conservation-of-energy thing, and so ought to be the same regardless of the L or C, above and below resonance. Unless the meter becomes totally perplexed, but if the user tells the correct lie about the equivalent circuit perhaps the perplexity can be resolved. At least the meter will be happy. Perhaps Possum too.

One thing to worry about is that if the apparent Q is too low, even the Extech will be unable to measure accurately. Yesterday, I tried to measure the self-inductance of a 500 ohm carbon-film resistor, and the Extech claimed 50 microhenries. Umm, no. That's a lot for a tiny air-core coil with perhaps 5 turns. It will be in the nanohenries. There is no doubt a standard impedance bridge able to make this measurement, but that's too much like work for me.

[Mike Sulzer]

Now AC resistance is a conservation-of-energy thing, and so ought to be the same regardless of the L or C, above and below resonance. Unless the meter becomes totally perplexed, but if the user tells the correct lie about the equivalent circuit perhaps the perplexity can be resolved. At least the meter will be happy. Perhaps Possum too.

Perhaps it can be resolved. But if we really have a resonant circuit, consisting of a C in parallel with an L and R in series, how does the measured phase shift relate to the energy loss? For example, there is a frequency close to or at resonance (depending on how you define resonance in a low Q circuit) where the current and voltage are in phase. If you set the meter on "R" you could get a measurement in this way (if your meter measures Rs at other than dc). Set on "L" or "C", it would not give a valid reading. I suspect going somewhat above resonance and setting the meter on "C" will not give a valid reading either.

One thing to worry about is that if the apparent Q is too low, even the Extech will be unable to measure accurately. Yesterday, I tried to measure the self-inductance of a 500 ohm carbon-film resistor, and the Extech claimed 50 microhenries. Umm, no. That's a lot for a tiny air-core coil with perhaps 5 turns. It will be in the nanohenries. There is no doubt a standard impedance bridge able to make this measurement, but that's too much like work for me.

That is a good example. So when Possum puts a piece of steel near a pickup and the inductance reading goes down, is this also an unexpected error?