An LVDT generates its usable signal from the difference between two AC voltages induced in identical secondary windings that are wired in series opposition, a topology that cancels equal-and-opposite secondary voltages at the mechanical null and yields a linear differential output as a ferromagnetic core moves off-center [S1][S3].
The primary winding is energized by a constant-amplitude AC source typically between 1 kHz and 10 kHz, and a soft-iron core sliding inside the bore channels more flux into the secondary it approaches while starving the opposite secondary [S4][S2]. The net signal at the two output leads is therefore the arithmetic difference of the two secondary EMFs, with magnitude proportional to displacement and phase telling the direction [S5].
The Series-Opposing Winding Topology
The two secondary windings on an LVDT are physically identical and symmetrically placed on either side of the primary, but their lead dress is reversed so that, when wired end-to-end, the voltages induced in them add with opposite sign [S3][S4]. Electronics Tutorials states the configuration directly: the two secondary coils are connected in a series-opposing arrangement and are electrically 180 degrees out of phase with each other, which is what gives the device the "Differential" part of its name [S4]. Transtek adds that the result is "AC voltages 180 degrees out of phase with each other" appearing at the two ends of the series string [S3].
Because the secondaries share a common return point at the center tap, the signal accessible to downstream signal conditioner electronics is the difference E1 - E2, where E1 and E2 are the EMFs induced in windings S1 and S2 respectively [S1][S5]. NewTek's tutorial describes the same idea in plain language: the effective voltage and LVDT output is the difference between each secondary, and the value of (S1 - S2) and (S2 - S1) becomes a linear function of core position [S1]. The output is therefore a single AC waveform whose amplitude and phase both carry position information, a property that downstream signal calibrator circuits or phase-sensitive demodulators can resolve into a unipolar or bipolar DC level [S2][S5].
Why the Null Voltage Is Zero at Center
When the ferromagnetic core sits at the geometric center of the coil stack, the flux from the primary couples equally into both secondaries, so E1 and E2 are equal in magnitude and 180 degrees apart in phase, and their difference is zero [S4][S1]. ScienceDirect's formal treatment shows this directly: with the core central, Ka equals Kb, Va equals Vb, and Vo equals Va minus Vb equals zero [S5]. Transtek notes that in real parts a small residual voltage, called null voltage or residual voltage, remains because of winding asymmetry, leakage reactance, and harmonics, but on a well-trimmed device it is a small fraction of full scale [S3].
That zero-output reference is the engineering feature that gives the LVDT its name and its usefulness: the null point is mechanically and electrically defined, repeatable, and stable over temperature because it depends on a symmetry of geometry rather than on absolute magnetic constants [S2][S4]. The general-purpose LVDT body in stainless-steel housing that TE Connectivity describes is built specifically to hold that symmetry, with the coils wound on a one-piece glass-reinforced polymer form, encapsulated against moisture, and wrapped in a high-permeability magnetic shield before being secured in the housing [S2]. The measured span on commercial devices is broad: TE Connectivity quotes a usable range from a few millionths of an inch up to ±30 inches (about ±0.762 m), while ScienceDirect's overview lists typical industrial spans of ±100 μm to ±100 mm with typical inaccuracy near ±0.5 percent of full scale [S2][S5].
Off-Center Behavior and the (E1 - E2) Transfer Function
Moving the core toward secondary S1 increases the mutual inductance between the primary and S1 and reduces it between the primary and S2, so E1 rises and E2 falls, and the differential output becomes (E1 - E2) with one phase [S1][S2]. Moving the core toward S2 reverses the roles, and the differential becomes (E2 - S1), which NewTek notes is a linear function of core position in the working range [S1]. ScienceDirect expresses the same relationship as Vo equals (K1 - K2) sin(omega t minus phi), where K1 and K2 are coupling coefficients that depend on how much of each secondary the core overlaps [S5].
The output is therefore approximately linear over a reasonable range of travel, expressed as Vo equals C times x, where x is displacement and C is a device-specific constant that combines the excitation amplitude, the winding turns ratio, and the core coupling geometry [S5]. NewTek reinforces the same point: as the core moves away from the null point in either direction, the voltage outputs (S1 - S2) and (S2 - S1) increase proportionally, and the typical quoted sensitivity numbers, such as 0.1 V per millimeter of travel used in a worked academic example, are useful as design values for signal isolator front-end scaling [S1][S7]. The waveform at the output is still a sinewave at the primary excitation frequency, not a DC level, which is why AC-excited LVDTs almost always feed a demodulator and oscillator pair rather than a direct analog-to-digital converter [S2][S4].
Direction Information: Magnitude Plus Phase
Series opposition gives the LVDT something an ordinary single-secondary transformer cannot: a built-in way to tell which side of null the core is on, encoded as a 180-degree phase flip rather than as a sign of voltage [S3][S5]. ScienceDirect states the rule explicitly: for equal-magnitude displacements of plus x and minus x away from null, the magnitude of Vo is the same in both cases, and the only information about the direction of movement is contained in the phase of the output voltage, which differs between the two cases by 180 degrees [S5]. That is why the LVDT needs a phase-sensitive demodulator downstream and not just a rectifier.
Real implementations almost always sit behind a dedicated signal conditioner that provides the oscillator, the synchronous demodulator, and an output stage that delivers ±10 V DC or 4-20 mA to the rest of the control system, which is the most common form in which end users actually meet the device [S2][S4]. The phase relationship is also why two LVDTs cannot be paralleled on the same excitation without phase discipline, and why the four (sometimes five) wires that come out of an LVDT body must be kept consistent with the OEM pinout: a swap of the two secondary leads reverses the sign convention of the whole measurement chain [S3][S6]. RealPars notes that, as both secondary coils are typically wired in series, the sensor has four wires or four connections in its connection socket, which is the practical footprint of the series-opposing topology [S6].
Operating Limits, Failure Modes, and When Series Opposition Is the Wrong Choice
Series opposition gives clean nulling, infinite resolution, and inherent phase-direction encoding, but the same topology also defines the operating envelope. Excitation must be a stable sinewave in the 1 kHz to 10 kHz band; below that, transformer coupling weakens and the signal drops, and above it, eddy-current losses in the core and shield heat the device and shift its calibration [S4][S2]. Temperature range on commercial LVDTs extends from about -150 degrees Celsius to roughly 537 degrees Celsius for special-purpose builds, which is why the same physics is used in cryogenic, nuclear, and turbine-valve position feedback [S1][S2].
Common failure modes that show up in the field trace back to the series-opposing design. An open secondary winding drives the output to a large, saturated, single-ended voltage because one side of the differencing pair is missing, which is the dominant symptom in field troubleshooting and a useful built-in diagnostic for a signal repeater or alarm card [S3][S4]. Excessive core travel into the nonlinear end region makes the (E1 - E2) curve bow, which is why datasheets specify a linear range narrower than the full mechanical stroke; exceeding it does not break the part but does break the proportionality [S1][S5]. DC-operated LVDTs, often called DC LVDTs, hide the same series-opposing winding pair behind an internal oscillator and demodulator, so the user only sees a DC output and loses the ability to inspect raw phase, which is the right tradeoff for cabinet-mount installations but the wrong one for high-temperature or radiation-hardened service where the bare AC LVDT and a remote conditioner outperform any integrated DC package [S1][S2].
For related process-control work that depends on clean analog position feedback, see the analysis of smart positioner diagnostics that cut unplanned valve outages, which uses an LVDT-class position sensor inside the positioner; the tubular linear motor versus screw-driven actuator decision for a comparison of where an LVDT is and is not the right feedback element; and the motor protection relay trip-wiring discussion for the trip-amplifier side of the same analog signal chain. A reasonable next check is to compare a half-bridge and full-bridge LVDT configuration on the same signal conditioner, since a half-bridge uses a single secondary plus an external reference winding and gives up the perfect null that the series-opposing pair provides, which is the tradeoff that decides the part number on most new valve-position and ram-position bills of material.