An IEPE accelerometer or pressure sensor embeds the charge-conversion stage, typically a MOSFET-input charge-to-voltage amplifier with a capacitive feedback network, inside the transducer housing and superimposes its output on a 2–20 mA constant-current bias fed through the same coaxial pair [S4][S6].
The acronym stands for Integrated Electronics Piezo-Electric; the same architecture is sold under the trade names ICP and Piezotron, all of which differ from a charge-output (PE) sensor in that they need a current source rather than a low-noise high-impedance cable and an external charge amplifier [S1][S3].
What the sensor actually contains: JFET or MOSFET front end and a DC bias
The piezo element in an IEPE device still produces charge in picocoulombs per unit of measurand, the same physical mechanism as a PE sensor, but a high-input-impedance JFET or MOSFET stage sits in the same hermetic package and converts that charge to a voltage referenced to the case ground [S4][S6].
Following the front end, a bias network sets a quiescent DC voltage at the amplifier output, typically 8–14 V on a 24 V compliance, so the AC vibration or pressure signal rides on top of a known DC level that the readout can AC-couple out [S3][S6]. This eliminates the need for the low-noise high-impedance cable that charge-output systems mandate [S1].
Two-wire operation: current source, same pair carries power and signal
An IEPE-compatible DAQ card or signal conditioner sources 2–20 mA (commonly 4 mA) through the center conductor and shield of a standard coaxial cable; that same pair returns the amplified, biased voltage to the instrument, where a coupling capacitor strips the DC bias and feeds the AC signal to the ADC [S3][S5][S6].
Because the cable now sees a low-impedance voltage-mode signal rather than a high-impedance charge node, run lengths of hundreds of metres are practical without the triboelectric and capacitive pickup problems of PE cabling [S3][S8]. Kistler's application note makes the same point: IEPE needs a standard two-wire cable plus an IEPE coupler for power and decoupling, while PE requires a low-noise high-impedance cable plus a separate charge amplifier stage [S1].
Comparison: charge-output (PE) versus IEPE on the four decision axes
The two architectures are not interchangeable. A charge-output (PE) sensor needs no supply current and can operate at temperatures above the 120–150°C electronics limit of typical IEPE devices, so PE is the only option on hot turbine and engine stages [S1][S3]. An IEPE sensor, in return, gives a low-impedance voltage output, fixed sensitivity in mV/g or mV/psi, direct connection to a wide range of DAQ modules, and cable runs that scale to plant-wide monitoring without exotic cabling [S1][S3].
The bandwidth and noise floor also differ in practice: a charge amplifier's gain is set only by its feedback capacitor, so cable and source capacitance drop out of the transfer function and quasi-static response is preserved, whereas an IEPE front end adds a high-pass corner that blocks very low frequencies and prevents quasi-static measurement [S1][S9]. For typical machinery vibration, the IEPE high-pass corner sits well below the band of interest, but for processes that need DC or sub-Hz pressure readout, PE remains the correct tool [S1].
Why the current source matters: bias detection and open-sensor diagnostics
Because the same two wires carry bias current and signal, the readout stage can monitor the DC bias voltage at the input as a health check. A normal sensor sits near its nominal bias, a shorted cable pulls it near 0 V, and an open element lets it rail toward the compliance voltage, all visible on the DAQ without any extra wiring [S3][S5].
That same constant-current source is what allows the inductive sensor and IEPE ecosystems to coexist on a plant floor: an IEPE channel is just a 4 mA loop with a coupling cap, and the bias is what differentiates it from a generic voltage input. NI's DAQmx documentation groups IEPE and charge-mode conditioning under the same Sound and Vibration hardware family precisely because the front-end topology, current excitation, and AC-coupled readout are shared [S5].
Limitations, failure modes, and the cable trap that still bites people
IEPE's internal electronics impose a ceiling of roughly 120–150°C on commercial accelerometers and limit sustained operation in engine, turbine, and combustion-pressure service, which is exactly where PE sensors remain specified [S3]. At cryogenic and very high temperatures, only the passive piezo element survives, so the IEPE form factor is simply not available [S1].
Long cable runs can still be derailed by low-quality coaxial cable: high loop resistance reduces the headroom available for the DC bias, and excessive cable capacitance rolls off the high end of the bandwidth. The classic field failure is a technician substituting a long, thin two-conductor cable for a real 50 Ω coax, watching the bias sag, and concluding the sensor is bad when the wiring is the problem [S3][S6]. Pair IEPE parts only with the cable and constant-current source the manufacturer rates, or move the displacement sensor and vibration chain to a charge system if the environment forces it.
Selection rules: when to pick IEPE and when to stay with charge-mode
Specify IEPE for machinery-health monitoring, modal analysis, and most plant-floor pressure pulsation work where the temperature stays under about 120°C and the readout is a standard DAQ, and specify PE (charge) when the measurement is quasi-static, when the sensor sits in a hot section, or when one physical sensor must cover multiple measuring ranges selected at the external charge amplifier [S1]. The same divider is used in calibration labs: a flow sensor on a hydrostatic rig with a PE charge amplifier gives adjustable range, while a portable IEPE accelerometer gets you out of the lab faster with only a coaxial cable and a DAQ [S3][S5].
A related decision point is signal conditioning topology. The TI front-end reference design shows that the same op-amp and JFET pair used in an external charge amplifier can be repackaged inside the sensor with a series diode network and bias resistor, which is the entire hardware difference between a charge-mode accelerometer and an IEPE accelerometer at the schematic level [S6]. Understanding that equivalence is what lets a capacitive sensor or IEPE channel be redesigned into a custom probe without re-architecting the DAQ.
Trackable signals: the IEPE standard continues to ship at 24 V compliance and 4 mA bias on essentially every modern Sound and Vibration DAQ; the practical next data point is the per-channel cost gap closing as more MEMS-class IEPE devices hit the market, plus incremental improvement in the upper temperature limit, which remains the dominant reason to still spec a charge-output sensor on hot-path machinery.
Background reading: 135° Seismic Hook Bending on a Rebar Bender: Angle, Mandrel, and Machine Gates.