17

I see negative capacitance measuring electrolytic capacitor at high frequencies (40kHz). How should I interpret that?

The device is LCR Meter Hantek 1832C. I failed to find anything about that in the manual.

chicks
  • 183
  • 3
  • 8
max5555
  • 275
  • 2
  • 8
  • Negative capacitance is impossible physically. It could be a bug in the meter's measurement algorithm (which naturally wouldn't be covered in the manual), or perhaps a calibration problem (which may be covered). – Dmitry Grigoryev Mar 07 '22 at 13:01
  • When your measurements show that the ac current lags the ac voltage (for one single frequency) the device under test behaves either (a) as a (lossy) inductor or (b) as a (lossy) negative capacitor. The decision between (a) and (b) can be made when the frequency is varied. A negative capacitance can be created by active circuits only (impedance converters, NIC or GIC). – LvW Mar 08 '22 at 15:28
  • Negative capacitance would have a positive imaginary reactance INVERSELY proportional to frequency. Thus it is NOT the same as inductance. – richard1941 Mar 13 '22 at 18:52

4 Answers4

16

I see negative capacitance measuring electrolytic capacitor at high frequencies (40kHz). How should I interpret that?

Negative capacitance can be regarded as inductance so, at high-ish frequencies it appears you might be measuring the effective series inductance (ESL) of the electrolytic capacitor. Strictly speaking you'll be measuring the inductive reactance minus the capacitive reactance and this is then converted to negative capacitance when the inductive reactance is a higher value than the capacitive reactance. Example: -

  • If the ESL is 1 μH, at 40 kHz it has a reactance of 0.251 Ω.
  • If the capacitance is (say) 100 μF, it has a reactance of 0.04 Ω at 40 kHz
  • Given that they are both in series, the net reactance is 0.251 Ω - 0.04 Ω = 0.211 Ω.
  • This might be displayed as minus 18.9 μF.
  • Or, on a more sophisticated meter, it might be displayed as 0.84 μH
Andy aka
  • 434,556
  • 28
  • 351
  • 777
  • 6
    I disagree that negative capacitance is inductance. – Scott Seidman Mar 06 '22 at 15:39
  • @ScottSeidman OK, maybe i should say that negative capacitance can be regarded as inductance. I've made that change BTW. – Andy aka Mar 06 '22 at 15:41
  • 7
    Different transforms. -1/sC vs sC. Two different things. A pole vs a zero – Scott Seidman Mar 06 '22 at 15:43
  • @ScottSeidman are you still disagreeing with my altered words? – Andy aka Mar 06 '22 at 16:43
  • 6
    Yes. A negative capacitance isn't an inductance, and CD shouldn't be regarded as such. – Scott Seidman Mar 06 '22 at 16:56
  • @ScottSeidman I'm not trying to say that a negative capacitance is under all circumstances equated to a positive inductance; The question asked "How should I interpret that?" and I'm giving an interpretation of what a displayed value of negative capacitance means in these circumstances. RTFQ. – Andy aka Mar 06 '22 at 17:42
  • 6
    Gentlemen, There's no contradiction. @ScottSeidman 's negative capacitance would for ex. get charged from a DC voltage through a resistor with exponentially ever increasing current. Free energy makers and apocalypse preachers maybe could use the idea, but positive inductors are useless as replacements for true negative capacitors. A not so perfectly constructed C-meter which operates in a fixed frequency may show an inductor or effectively at that frequency inductive more complex circuits as negative capacitance. The questioner should test the measurement with a known inductor. –  Mar 06 '22 at 19:54
  • @user287001 a negative capacitor would charge to a negative voltage exponentially for sure. – Andy aka Mar 06 '22 at 20:36
  • 18
    I think that there’s a bit of subtlety here: the measurement technique itself makes inductance appear as negative capacitance on the readout. This doesn’t mean that positive inductance and negative capacitance are equivalent in circuit analysis. It’s a measurement artifact and something to know about to properly interpret the measurement results. – Kuba hasn't forgotten Monica Mar 06 '22 at 22:46
12

If you measure reactance at a single frequency, you can't distinguish between negative capacitance and positive inductance.

Basically, the meter is converting the imaginary part of the reactance ( Im(Z) ) to 'capacitance' by using C = -1/(2 π f Im(Z)). If Im(Z) is positive (which occurs with a inductance), the meter will display a negative capacitance.

The meter could measure at different frequencies (in fact it can, but it doesn't combine the results) and combine the results it could analyze for an equivalent circuit of the component (e.g. some network of R, C, L) -- but that is v. difficult and unnecessary in a LCR meter.

jp314
  • 18,395
  • 17
  • 46
  • 2
    Can a VNA distinguish it? If so, I think it would be great to point OP into directions on how to measure his components in order to get more reliable results. – Thomas Weller Mar 07 '22 at 07:53
  • Actually just the measurement would happen very differently. Connecting a meter starts a transient which would in case of negative capacitance expand exponentially instead of decaying towards a solid state. The reactances of positive inductance and negative capacitance are not distinguishable. –  Mar 08 '22 at 23:33
  • A VNA will show impedance measurements vs. frequency. There is not a unique conversion from this to an equivalent circuit, but if you presumed a C in series with an L and R, you could certainly 'fit' this around the frequencies of interest and see the 'pure' C and L values. – jp314 Mar 11 '22 at 01:50
6

It's called "capacitance", but it's not really capacitance.

When a sinusoidal voltage, at any specific frequency, is applied to the capacitor, a sinusoidal current, at the same frequency, results.

The relationship between the current and the applied voltage has two parts -- it differs in phase by a fixed amount, and it has a magnitude that is proportional to the voltage with some factor.

It is mathematically convenient to represent these two quantities (the phase difference and the factor) by a single complex number called "impedance", denoted "Z". The derivation is a little too long for this answer, but google "negative frequencies" if you want to understand how it works. The impedance is convenient, because it allows the relationship between voltage and current to be expressed by extending Ohm's law to complex numbers: V = IZ, where voltage and current are sinusoids of a particular frequency.

An ideal resistor has a constant real impedance (meaning Z has no imaginary component) at every frequency, and so the real part of impedance is often called "resistance".

An ideal inductor or capacitor has a purely imaginary impedance (meaning Z has no real component) at every frequency. The imaginary component is called "reactance" and it is not constant. For both inductors and capacitors, reactance is inversely proportional to frequency, though, so (Imaginary part of Z)/f is often called "inductance" if it's positive, or "capacitance" if it's negative.

So your meter is just measuring Z at some specific frequency and labelling -Im(Z)/f as "capacitance". It doesn't mean you have a negative capacitor. It also doesn't mean you have an inductor. It just indicates that the phase of the current is lagging behind the voltage at that frequency.

Matt Timmermans
  • 1,358
  • 6
  • 7
  • 1
    Those meters apply a sinusoid to the component under test; measure the real and Im parts of the impedance and calculate a 'capacitance'. This just is the capacitance that would have an equivalent impedance AT THAT FREQUENCY. The meter is unable (no one can if calculations are done at just 1 frequency) to determine what precise components are there. – jp314 Mar 09 '22 at 05:18
  • Note the capitalized words in the above comment. Very important! – richard1941 Mar 13 '22 at 18:54
0

Since your capacitor is electrolytic, AC is not a good way to measure it. Electrolytics should never be reverse charged, which is what AC will do. A good way to measure the capacity is by rate of discharge though a known resistance. Hook the capacitor and a known resistor across an oscilloscope and measure the time required for the capacitor to discharge to 1/e (37%) of an initial voltage. Then do the math using the formula r = ct, where r is the time constant, c is the known resistance, and t is the capacitance.

I do not recommend testing the voltage rating unless you are sampling from a large batch; the voltage test is destructive and sometimes stinks real bad.

Also, if the voltage rating is several hundred volts, do not charge the capacitor and play catch with it as we did in high school electronics shop (1955).

richard1941
  • 632
  • 3
  • 7
  • If you wrecked the capacitor by application ofAC, you may be able ro re-form it by application of a slowly increasing voltage of correct polarity. IIRC, the ramp should take several hours and final volage held overnight. – richard1941 Mar 11 '22 at 00:05
  • AC is required to compare capacitors for their capability to damper short pulses – max5555 Mar 12 '22 at 15:27