Selecting DC Current Sensors for Accurate 0 to 90 Hz

The panel shows current during a change, then the indication collapses toward zero even though DC still flows. Or the reading follows 50/60 Hz correctly but becomes unstable, nonlinear, or absent near 0 Hz. Start here: identify whether the sensor is a passive current transformer or an actively excited magnetic transducer. A passive CT cannot produce a continuous DC measurement.

Stop applying the wrong fixes

Several familiar actions can change the symptom without correcting the measurement principle.

• Changing the burden resistor: This changes secondary voltage, loading, gain, and AC error. It cannot make steady DC induce a secondary voltage. An excessive burden can also drive the core closer to saturation.
• Adding turns to an ordinary secondary: More turns change the AC transformation ratio and induced voltage. They do not create the changing flux required for continuous DC output.
• Calibrating after the indication disappears: Calibration cannot recover information that the magnetic circuit is no longer transferring. A DC-biased or saturated core produces a nonlinear response, so a fixed scale factor is not a cure.
• Adding a reset winding without modulation: A tertiary winding can move or reset core flux, but periodic resetting alone does not yield an accurate DC value. The electronics must apply controlled excitation, observe the core response, demodulate it, and relate that response to primary current.
• Using a Rogowski coil by itself: A Rogowski coil responds to changing current. It can extend AC bandwidth in a hybrid instrument, but it cannot report the zero-frequency component.
• Assuming a 50/60 Hz CT remains accurate down to 0 Hz: Passing a power-frequency test proves only operation at that test frequency and amplitude. It says nothing about DC response.

Also separate sensor capability from display behavior. Filtering, scaling, rectification, or a slow panel update can hide a working sensor signal, but those settings cannot turn a passive CT into a DC sensor.

Find the real cause in the magnetic circuit

An ordinary CT operates by transformer action. Its secondary voltage follows changing core flux, described by e = N·dΦ/dt. When primary DC first rises, the change creates a transient secondary signal. Once the current becomes steady, dΦ/dt approaches zero and the induced signal disappears.

The DC ampere-turns still bias the core. Flux moves away from its centered operating point and can push the material into saturation. Once saturated, incremental permeability falls and the CT stops reproducing even the changing part of the current accurately. The result can be lost gain, clipped peaks, phase error, slow recovery, or different readings for positive and negative current.

Low-frequency AC creates a related problem. For a given secondary voltage and number of turns, required flux rises as frequency falls. The core therefore needs progressively more flux capacity as the signal approaches 0 Hz. At exactly DC, there is no alternating volt-second balance to return the flux through a symmetric cycle.

That is why a product described as using a normal-looking CT core from DC through 90 Hz must contain another operating mechanism. If electronics inject an oscillating secondary signal and measure how primary current changes the resulting magnetic response, the assembly is an active magnetic current sensor, not a passive CT measurement channel.

Match the symptom to the cause

Choose a sensing method that includes zero frequency

Select the principle before selecting the package or conductor aperture.

• Hall-effect sensing: A Hall element measures magnetic field at DC and AC. The core normally needs a gap for the element. This is common in DC clamp instruments, including instruments used across ranges from hundreds of milliamps to 200 A. Account for sensitivity drift, zero offset, and any required de-gaussing or zeroing process.
• Fluxgate sensing: Drive a magnetic core into alternating saturation and detect the asymmetry caused by the measured DC field. Fluxgates provide high sensitivity and true DC response. Commercial implementations can extend from DC into the megahertz range, but bandwidth, offset, excitation feedthrough, and recovery remain product-specific.
• Transductor or magnetic-amplifier sensing: Apply a known oscillating excitation to a saturable magnetic element and measure how primary ampere-turns alter the driven response. This matches the described method in which electronics excite the secondary of an earth-leakage CT and infer current from its effect on the carrier. Accuracy depends on the complete excited-core circuit, not on the CT alone.
• Orthogonal fluxgate sensing: Arrange excitation and sensing fields orthogonally. One winding is driven alternately into saturation; changes observed in the other reveal the DC component. US Patent No. 6,885,183 describes this approach together with a Rogowski coil for the AC component.
• Commutating windings: Alternate the effective primary and secondary winding directions so the core does not accumulate one-direction flux. US Patent Application No. 20120063055 describes measurement of DC and AC through this commutation method. Treat any bandwidth claim as a property of the implemented device, switching circuit, and signal processing.
• Optical current sensing: Optical methods can cover DC through 90 Hz without wire-core saturation. They require optical conversion and amplifier electronics, so evaluate noise, offset, auxiliary power, and output-interface behavior as part of the sensor.
• Hybrid sensing: Combine a DC-capable Hall or fluxgate channel with a CT or Rogowski channel for AC. The crossover and summing networks must preserve gain and phase through their overlap region.

Define the measurement before selecting hardware

Do not buy against the phrase 0-90 Hz alone. Write down the test conditions that determine whether the instrument works.

• Specify the minimum and maximum current, including DC offset plus AC peak.
• State whether the required value is instantaneous current, average DC, RMS, leakage current, or separate AC and DC components.
• Set allowable zero offset, gain error, nonlinearity, noise, and drift. Read the applicable values from the candidate datasheet; no universal figures apply across these methods.
• Define performance at 0 Hz, at the lowest nonzero test frequency, at normal power frequency, and at 90 Hz.
• Record required isolation, conductor aperture, output type, auxiliary supply, and startup behavior.
• Include the largest DC bias that may coexist with the measured AC signal. This is the condition most likely to expose saturation in a passive magnetic channel.
• Determine whether the product requires zeroing or de-gaussing and whether that operation can occur while installed.

For low-amplitude current, offset and noise often set the useful lower limit before full-scale accuracy does. Compare the smallest required signal with the specified zero error and output noise. If the datasheet omits those quantities, request them from the manufacturer before qualification.

Test the complete channel from DC to 90 Hz

1. Identify the architecture. Trace the conductor, core, windings, electronics, auxiliary power, and output. A powered module with carrier excitation or synchronous detection is not equivalent to a bare CT connected to a burden.
2. Check zero. Remove primary current, allow the specified startup process to complete, and record the raw output and displayed value. Perform only the manufacturer-defined zeroing or de-gaussing action.
3. Apply steady DC. Use a known current within the sensor and test-source ratings. Hold it constant. A valid DC channel must retain a stable indication after switching transients disappear.
4. Reverse polarity. Apply equal positive and negative currents. Compare magnitude, zero crossing, and residual offset after returning to zero.
5. Sweep frequency. Test the lowest required nonzero frequency, intermediate points, power frequency if relevant, and 90 Hz. Keep the commanded current amplitude controlled and compare gain and phase against a reference instrument suited to each test condition.
6. Add DC bias to AC. Repeat the frequency checks with the maximum required DC component present. Watch for clipped output, gain reduction, asymmetry, or delayed recovery.
7. Inspect carrier artifacts. On an actively excited sensor, view the raw and filtered outputs. Separate excitation ripple from actual current variation and check that downstream sampling does not alias the carrier into the measurement band.
8. Challenge recovery. Apply the largest permitted current condition, return to zero, and repeat the low-current test. A shifted zero or reduced gain points to remanence, saturation, or recovery limits.

Verify the result at the panel and the sensor

Prove the signal at three points: the primary reference, the sensor output, and the panel value. If the sensor output is correct but the panel is wrong, inspect input type, scaling, polarity, filtering, grounding, and update behavior. That is a signal-chain fault, not magnetic saturation.

Record DC gain separately from AC gain. For a hybrid sensor, inspect the crossover region for a bump, dip, or phase discontinuity. For an excited-core sensor, repeat zero and low-current checks after the unit has reached its normal operating condition; carrier amplitude and demodulator offset can affect the result.

A successful test shows a sustained DC indication, correct polarity, acceptable zero return, and specified response through 90 Hz with the required DC bias present. Passing a single 50/60 Hz point does not qualify the channel.

FAQ