LCR Meter

What an LCR Meter Measures

An LCR meter applies a known sinusoidal voltage (or current) to the device under test, measures the resulting current (or voltage), and from their ratio and phase relationship computes the complex impedance. Impedance is written in rectangular form as Z = R + jX, where R is the resistive part and X is the reactance, or in polar form as a magnitude |Z| and a phase angle theta. For an inductor the reactance is XL = 2 pi f L, and for a capacitor it is XC = 1 / (2 pi f C), so the same component shows a different impedance at every test frequency. This frequency dependence is the reason an LCR reading is meaningless unless the test frequency is quoted alongside it.

From the single measured impedance the instrument derives an entire family of parameters: the admittance Y = G + jB (the reciprocal of impedance, used for high-impedance parts), the primary value the operator selected (Cs, Cp, Ls, Lp, or R), and a secondary value describing loss. The two standard loss figures are the dissipation factor D, also called tan delta, defined as D = R / X, and the quality factor Q, defined as Q = 1 / D. Capacitors are usually graded by D, inductors and resonators by Q. The same loss expressed in ohms is the equivalent series resistance, ESR = D / (2 pi f C), which is the parameter that governs ripple-current heating in power-supply capacitors.

It is important to separate the LCR meter from two neighbors that look similar. A handheld component tester or the capacitance range on a digital multimeter typically applies a DC charge ramp and reports only capacitance with coarse accuracy, with no frequency control and no loss term. A micro-ohmmeter measures DC resistance only. The LCR meter is distinguished by its AC excitation at controlled frequency and level, its phase-sensitive detection, and its ability to resolve the reactive and resistive parts of impedance independently, which is what lets it report D, Q, and ESR.

The technique has a long industrial lineage. Early impedance measurement used manually balanced AC bridges built on the Wheatstone, Maxwell, Hay, or Schering circuits, in which an operator adjusted a calibrated standard until a null detector read zero. General Radio (later GenRad) was a dominant maker of these bridges from the 1920s onward and later introduced digital impedance meters. The manual null was eventually replaced by the auto-balancing bridge, in which an operational amplifier holds a virtual ground and automatically forces the feedback current to cancel the current through the unknown, achieving balance electronically and continuously. That topology underlies essentially every precision benchtop LCR meter sold today.

Four engineering questions decide whether an LCR reading can be trusted: at what frequency and level was it taken, was the correct series or parallel model selected, was a four-terminal Kelvin connection used, and was open and short compensation performed. A meter with 0.05 percent basic accuracy will still produce a wrong answer if any of these is neglected, which is why the chapters below give as much attention to method as to the instrument itself.

Instrument Classes and Types

LCR instruments split into four practical classes by form factor and duty, and conflating them is the most common procurement error. A handheld unit, a bench precision meter, a production meter with a handler interface, and a frequency-sweeping impedance analyzer share the same physics but differ by an order of magnitude in price, accuracy, and frequency reach. The table below summarizes the classes before the text examines each.

Handheld LCR meters apply a small set of fixed test frequencies, commonly 100 Hz, 120 Hz, 1 kHz, 10 kHz, and 100 kHz, and trade accuracy for portability and battery operation. They are the right tool for verifying an SMD value at a repair bench or sorting a parts bin, where 1 percent uncertainty is acceptable. Their fixed signal level and limited compensation make them unsuitable for grading components against a tight datasheet tolerance or for measuring ESR below roughly 0.1 ohm.

Benchtop precision meters are the reference class. They offer continuously selectable frequency, programmable signal level, internal and external DC bias, four-terminal-pair Kelvin connection, and full open, short, and load compensation. Basic accuracy of 0.05 percent is typical, with a guaranteed impedance range spanning milliohm to roughly 100 megohm. The Keysight E4980A (20 Hz to 2 MHz) and Hioki IM3536 (DC and 4 Hz to 8 MHz) define this segment, and both add an equivalent-circuit analysis function that fits a component to a model.

Production LCR meters prioritize measurement speed and a parallel handler (binning) interface over the widest frequency range. A measurement time of about 1 ms in fast mode lets a pick-and-place handler sort hundreds of parts per minute into pass and reject bins against comparator limits, with the handler I/O signalling the result electrically. These instruments still carry 0.05 percent accuracy, but the buying decision turns on the I/O connector, the comparator and binning logic, and contact-check (to catch a part that is not seated).

Impedance analyzers extend the LCR meter by sweeping frequency, level, or DC bias and plotting the result as a curve rather than a spot value. The Wayne Kerr 6500B series, for example, sweeps from 20 Hz to between 5 MHz and 120 MHz depending on model, with a frequency-set accuracy of 0.005 percent, which is what design engineers need to find a capacitor's self-resonant frequency or an inductor's Q peak. Above roughly 10 MHz the auto-balancing bridge gives way to RF current-voltage and network-analysis methods, the subject of the next chapter.

Impedance Measurement Methods

The Keysight Impedance Measurement Handbook catalogs six measurement methods, and no single method covers the whole frequency span from millihertz to gigahertz. Selecting the right method (or, more practically, the right instrument built around it) is a function of the test frequency and the impedance range. The table below maps the methods to their working envelope before the text explains the two that matter most for component LCR work.

The auto-balancing bridge is the workhorse of the LCR meter. An operational amplifier drives a feedback current through a precision range resistor so that the low side of the device under test is held at a virtual ground; balance is reached when the feedback current exactly equals and opposes the current through the unknown. Measuring the test voltage and the feedback current then yields impedance directly. This method delivers wide frequency coverage from low frequency up into the high-frequency region while maintaining high accuracy across an impedance range from milliohm to roughly 100 megohm, which is why it is used in nearly every precision benchtop meter. Its practical ceiling is around 110 MHz; beyond that, stray reactance in the bridge dominates.

The RF current-voltage method takes over above the auto-balancing bridge's ceiling. It measures the voltage across and current through the device using precisely matched RF transformers and a coaxial test port, giving high accuracy and a broad impedance range at frequencies into the gigahertz, which the Keysight E4982A reaches to 3 GHz. The resonant or Q-meter method tunes a known reference to resonance with the unknown and is historically the most accurate way to measure very high Q inductors, at the cost of having to retune for every measurement. The network-analysis method infers impedance from a reflection coefficient and is most accurate near the system characteristic impedance, typically 50 ohm, degrading for very high or very low impedance.

For the component engineer the practical takeaway is narrow: if the parts are passives tested at or below a few megahertz, an auto-balancing-bridge LCR meter is correct and most economical. Only when the test frequency must exceed roughly 10 MHz, as for RF inductors, multilayer ceramic capacitors near self-resonance, or matching networks, does the RF I-V or network method, and therefore an RF-class instrument, become necessary. Buying gigahertz reach for a 1 kHz electrolytic-capacitor line is wasted capital.

Test Frequency, Signal, and Standards

An impedance value is reproducible only when the test conditions are reproducible, and the two conditions that change the answer most are the test frequency and the test signal level. Because reactance scales with frequency, a 100 microfarad capacitor reads a different impedance at 120 Hz than at 1 kHz, and a ferrite inductor or a Class II ceramic capacitor additionally changes value with the applied voltage or current. The conventions below come from component standards and datasheet practice, and a measurement quoted without them cannot be checked.

Test frequency must match the application or the standard. IEC 62391-1, the standard for electric double-layer (super) capacitors, specifies that AC equivalent series resistance is measured at 1 kHz, while aluminum electrolytic capacitance and dissipation factor are conventionally taken at 120 Hz (or 100 Hz in 50 Hz regions). Ceramic and film capacitors are graded at 1 kHz, with high-frequency behavior checked at 100 kHz or above. Quoting an electrolytic ESR at 1 kHz against a datasheet figure specified at 120 Hz is a frequent source of false rejects at incoming inspection.

Test signal level is the second hidden variable. Class II ceramic capacitors built on ferroelectric dielectrics lose capacitance under both AC drive and DC bias, so two labs measuring the same X7R part at different levels will legitimately disagree. Ferrite-cored and powdered-iron inductors lose inductance as the core approaches saturation, so inductor value and Q must be quoted at a stated current. Precision meters let the operator program the AC level (for example from about 5 mV to 5 V on Hioki instruments, and up to a 20 Vrms option on the Keysight E4980A) and superimpose an internal or external DC bias.

DC bias deserves explicit attention because it is where many selection mistakes hide. Characterizing a power-supply ceramic capacitor at its working DC voltage often reveals only a fraction of the rated capacitance remains, and a power inductor must be biased at its operating DC current to show its in-circuit value. Internal bias on a bench meter is modest (the GW Instek LCR-6000 supplies plus or minus 2.5 V internally; the Keysight E4980A offers a 40 V DC bias option), so high-voltage or high-current bias requires an external bias source or a dedicated DC bias unit such as the Wayne Kerr current bias accessories.

Two governing standards anchor this chapter. IEC 60384-1 is the generic specification for fixed capacitors for use in electronic equipment and defines the test methods that downstream sectional specifications (IEC 60384-4 for electrolytics, and others) refer to. IEC 62391-1 covers fixed electric double-layer capacitors and defines capacitance, ESR, and leakage-current test methods plus the four application classes (memory backup, energy, power, instantaneous power). Citing the exact clause used keeps a supplier dispute factual.

Key Specification Parameters

An LCR meter datasheet lists dozens of lines, but only a handful drive the selection and the validity of the reading: measurement parameters, frequency range, basic accuracy, signal level and DC bias, measurement speed, compensation, the equivalent-circuit model, and the connection method. Each is decoded below, with the series-versus-parallel model treated as a full subsection because it causes more disputed readings than any other single setting.

Measurement parameters. A capable meter reports a primary and a secondary parameter chosen from Cs, Cp, Ls, Lp, Rs, Rp, |Z|, theta, |Y|, D, Q, ESR, and (on DC-capable models) Rdc. The primary parameter is the value of interest; the secondary describes loss. Confirm the meter can output the exact pair the inspection plan calls for, for example Cs with D for an electrolytic, or Lp with Q for an RF choke.

Basic accuracy is the floor figure, valid only under stated conditions (a mid-band frequency, a specific signal level, and a defined impedance window). Real accuracy degrades toward the extremes of frequency and impedance and is given by a formula or a chart in the datasheet, not by the single headline number. Benchtop instruments quote 0.05 percent basic; handheld units quote 0.1 to 1 percent. Always read the accuracy as a function of where your part actually sits, not as the best-case value.

Series versus parallel equivalent circuit. Every real component is modeled as a reactance with an associated loss resistance, but the meter must be told whether that resistance is in series (Rs) or parallel (Rp) with the reactance, because the two yield different numbers for the same part. The accepted rule is impedance-based:

• Low impedance, |Z| about 10 ohm or less (large electrolytics, low-value inductors): use the series model, reporting Cs or Ls with Rs / ESR. The series resistance dominates the small reactance.
• High impedance, |Z| about 10 kilohm or greater (small-value capacitors, high-value inductors): use the parallel model, reporting Cp or Lp with Rp. Parallel leakage dominates the large reactance.
• Between 10 ohm and 10 kilohm: follow the component datasheet or the governing standard; either model is defensible only if stated.

Choosing the wrong model is the usual reason two correctly calibrated meters disagree on the same capacitor. The mismatch is largest when D is high; for a near-ideal low-loss part the two models converge.

Compensation is a specification because its absence is a defect. Open compensation captures the fixture's parallel stray admittance with no part fitted; short compensation captures its series residual impedance with a clean short bar; optional load compensation references a known standard for the tightest accuracy. Compensation is valid only for the cable length, fixture, and frequency in use and must be redone when any change. Confirm the meter supports all-frequency (list) compensation if you sweep.

Connection and fixtures. Precision meters use a four-terminal-pair (4TP) connection split into current and voltage paths (Hc, Hp, Lp, Lc), realized as a Kelvin clip lead for through-hole parts or a guarded SMD fixture for chips. The 4TP connection excludes lead resistance at low impedance and cancels lead magnetic coupling at high impedance, and it is what makes accurate milliohm and high-megohm readings possible. Cable length is a calibrated parameter (commonly 0 m, 1 m, 2 m, or 4 m); using a length the meter is not compensated for invalidates the accuracy spec.

Selection Decision Factors

To turn the preceding chapters into a specific model, work through the ordered sequence below. Most selection errors are not a single wrong line item but a decision made at the wrong level, for example choosing frequency reach before defining the parts and the duty. These steps double as an RFQ template.

1. Define the parts and the parameter pair: list the component families to be tested (electrolytic, MLCC, film, power inductor, supercapacitor) and the exact primary plus secondary parameter each requires (Cs+D, Lp+Q, ESR, Rdc). This dictates everything downstream.
2. Set the required test frequencies and levels: derive them from the datasheets or governing standard (120 Hz, 1 kHz, 100 kHz are the common anchors). The needed maximum frequency, not marketing reach, sizes the instrument; few component lines need beyond 1 MHz.
3. Fix the accuracy class: distinguish field triage (0.1 to 1 percent handheld), production grading and bench characterization (0.05 to 0.1 percent), and reference-lab work (request the accuracy chart at your actual impedance and frequency, not the headline figure).
4. Specify DC bias needs: confirm the internal bias voltage and current, and whether an external bias source or DC bias unit is required for biased MLCC or power-inductor characterization. Internal bias is typically modest (single-digit volts to tens of volts).
5. Choose the connection and fixtures: four-terminal-pair Kelvin lead is mandatory below roughly 1 ohm; select the matching SMD test fixture, through-hole fixture, or component-handler contacts, and note the cable length the accuracy is compensated for.
6. Match the duty to the form factor: production lines require a fast measurement mode (about 1 ms) and a parallel handler interface with comparator binning and contact-check; R&D requires list sweep, equivalent-circuit analysis, and remote control over LAN, USB, or GPIB.
7. Confirm interface and data path: LAN, USB, GPIB, and RS-232C for logging; SCPI command support for automation; and a handler I/O connector type that matches the existing line. Verify it integrates with the test rack alongside the data logger and controller.
8. Total cost of ownership: purchase price plus calibration interval and laboratory cost, fixture and Kelvin-lead spares, and the cost of false rejects from an under-specified meter. A cheaper meter that cannot bias an MLCC at its working voltage can reject good parts and idle a line.

A frequently overlooked dimension is serviceability and traceability: the calibration interval (commonly 12 months), the availability of an ISO/IEC 17025 accredited calibration laboratory and Kelvin-fixture spares in your region, firmware and SCPI command stability, and how long the maker supports the model. Keysight, Hioki, Wayne Kerr, GW Instek, B&K Precision, and Chroma all maintain regional calibration and support, which makes them defensible choices for a multi-year production line. An instrument that cannot be recalibrated locally becomes a stranded asset regardless of its original accuracy.