Home / News / How to Choose a Current Sensor: Fluxgate, Closed-Loop, Open-Loop and ASIC

How to Choose a Current Sensor: Fluxgate, Closed-Loop, Open-Loop and ASIC

2026-11-02Selection Guide10 min read
Current sensor selection: comparison of four technology routes

Start With Three Questions, Not With the Range Table

Given a current sensor datasheet, the most common mistake is to start filtering by range. Range is only an entry condition; three other questions decide whether a part actually works.

  • What current are you measuring: DC, AC, or a complex waveform with a DC component? Do you need bidirectional measurement?
  • Where does it mount: busbar, cable, or PCB? What window size and installation space are available?
  • Who consumes the signal: an ADC, a protection circuit, or a PLC analogue module? Voltage output or current output?

Answer those three and the technology route is basically settled. What remains is shortlisting models — for that we maintain a complete parameter set covering 51 series and 180 models in six categories, filterable by range, accuracy, bandwidth and temperature.

Four Technology Routes and Where They Differ

RoutePrinciple and traitsTypical rangeTypical accuracyRepresentative seriesBest for
FluxgateZero-flux principle; strong on DC and small signals5 mA–700 A±0.3%~±1%FR2V, FR7V, FR8V, FR1CMicro current, insulation monitoring, high-accuracy DC
Closed-loop HallMagnetic balance; high bandwidth and linearity6–5000 A±0.1%~±1%CR8A, CM9A, CM1A, CR1V, CN2ADrives, servo, high-power supplies, SST main path
Open-loop HallDirect amplification of the Hall output; low cost, compact50–1500 A±1%HS1V, HK1V, HK4V, AS1VGeneral monitoring, energy metering
ASIC open-loopHall plus signal conditioning integrated; single-supply analogue output10–800 A±0.5%~±2%AN1V, AN3V, AN4VA, AN4VB, AN4VCOn-board integration, automotive and volume products

In one line: closed-loop for accuracy and bandwidth, fluxgate for small signals and DC, ASIC open-loop for cost and size, open-loop Hall for general monitoring.

Note that the four routes are not substitutes for one another. A typical machine uses closed-loop for control sampling on the main power path, ASIC or open-loop for branch condition monitoring, and a separate fluxgate for insulation monitoring.

Shortlist by Range

RangeRecommended routeRepresentative seriesTypical application
5 mA–0.3 AFluxgateFR2V, FR7V, FR8VLeakage current, insulation monitoring, micro current
6–50 AClosed-loop (small window / PCB)CR1V, CN1ALow-voltage side sensing, on-board sampling
25–100 APCB closed-loop / ASICCN2A, AN3V, AN4VAOn-board power sensing, inside modules
50–300 AClosed-loop / ASIC open-loopCR1A, AN1V, AN4VC, AS1VIndustrial drives, UPS, control cabinets
200–1500 AClosed-loopCM1A, CM2A, CM3A, CM9AHigh-power supplies, PCS
500–5000 AClosed-loop (large window / bolt)CR8A, CM5ABusbar current, high-power modules, electrolysis and metallurgy

The range table is only the start. Within one range you still compare window size, supply, output form and operating temperature — those four usually decide whether the part physically fits and wires up cleanly.

Mounting: PCB, Window or Bolted Busbar

PCB integration

CN1A and CN2A are PCB plug-in closed-loop designs. CN2A PB01, for example, covers 25–100 A in four ranges with 200 kHz bandwidth, ±0.5% accuracy, ±12…15 V supply and -40~85 °C operation, suited to on-board power current sensing. On the ASIC route, AN4VA comes in SOIC-8 with 3.3 V or 5 V single supply and -40~125 °C operation for volume products.

Window type

Window types are chosen by bore diameter. CR8A has a Ø120 mm window for busbars and multi-core cables; CR1V targets 6–25 A with PCB mounting and a small window, 3.3 V or 5 V single supply and ratiometric output for direct ADC sampling.

Bolted mounting

For high current, M6 bolt mounting (as on CR8A) clamps directly onto the busbar, giving better mechanical strength and contact reliability than threading cable through, and simplifying later maintenance.

Rule of thumb: current level and insulation requirement decide PCB vs. window; volume and space decide ASIC vs. closed-loop; maintenance access decides whether to bolt. Settle those three before reading datasheets.

Reading the Datasheet: Accuracy, Bandwidth, Output and Temperature

  • Accuracy: check whether it is room-temperature or full-temperature, and whether offset and drift are included. The gap is large in equipment with significant internal heating.
  • Bandwidth: choose for the highest frequency component you need to keep, not the maximum available. 100 kHz is the usual dividing line; below 50 kHz needs justification.
  • Response time: mandatory for protection loops. Applications needing under 2 μs and those tolerating tens of μs call for entirely different parts.
  • Supply: single 3.3 V / 5 V eases system integration; ±12…15 V and ±15…24 V are common in industrial closed-loop designs.
  • Output: voltage output (for example 2.5±0.625 V ratiometric) suits ADCs; current output needs a burden resistor; 4–20 mA suits long-distance transmission.
  • Temperature: select against measured extremes at the mounting point, and assess positions near power devices separately.

A concrete example: CR1V PB02 covers 6–25 A on a 3.3 V single supply with 1.65±0.625 V ratiometric output, ±0.4% accuracy, 200 kHz bandwidth and -40~85 °C operation — a good fit for direct ADC sampling on the low-voltage side. At the same range, if noise immunity and transmission distance matter more, consider a current output or a 4–20 mA solution.

Reminder: datasheet accuracy is normally quoted at room temperature, rated input and steady state. Treating it as a full-condition figure is the most common cause of "measurement drift" appearing late in a project.

Do Not Forget Voltage and Leakage Current

A current sensor is rarely used alone. Voltage measurement feeds power calculation and bus supervision, while leakage current monitoring is about insulation and safety.

  • Voltage: VN4A covers 500 V, 750 V, 1500 V and 2000 V (±0.5%); VN1V offers a 600 V version and a 1 mA signal version; VN4V targets 48 V systems (5 V single supply, 2.5±2 V output); VN4VA PB00 outputs 4–20 mA for direct PLC connection.
  • Leakage current: CSMD1, TR6A, FR1D and FR5D are used for residual current and insulation monitoring modules, forming part of the safety chain in PV, charging and storage systems.

Within one system, plan the output form and interface of the current, voltage and leakage current measurement chains together. It reduces complexity in the downstream acquisition circuit and gives a single reference basis for calibration.

FAQ

Is closed-loop always better than open-loop? On accuracy and bandwidth, yes; on cost and size, no. Closed-loop needs a compensation winding and a higher supply requirement, so it costs more. Where you only need monitoring and the signal does not feed control, an open-loop device is usually more economical and smaller.

Can I just use a current transformer for AC? For line-frequency AC you can, but a CT cannot measure DC components and has limited bandwidth. For VFD outputs, PCS and SST — anywhere waveforms contain DC or are non-sinusoidal — use closed-loop Hall or fluxgate instead.

The spec says ±0.5% — how much error will I really see? You must add temperature drift, offset, supply variation and mounting position on top. Confirm the full-temperature specification and test conditions, and ideally calibrate once under real operating conditions rather than relying on the nominal figure.

How do I narrow down 51 series quickly? Filter in four steps: range, then mounting, then output form, then temperature range. That usually converges to two or three series, after which you compare the specific model datasheets item by item.

Related reading

FAQ

Is closed-loop always better than open-loop?

On accuracy and bandwidth, yes; on cost and size, no. Closed-loop needs a compensation winding and a higher supply requirement, so it costs more. Where you only need monitoring and the signal does not feed control, an open-loop device is usually more economical and smaller.

Can I just use a current transformer for AC?

For line-frequency AC you can, but a CT cannot measure DC components and has limited bandwidth. For VFD outputs, PCS and SST — anywhere waveforms contain DC or are non-sinusoidal — use closed-loop Hall or fluxgate instead.

The spec says ±0.5% — how much error will I really see?

You must add temperature drift, offset, supply variation and mounting position on top. Confirm the full-temperature specification and test conditions, and ideally calibrate once under real operating conditions rather than relying on the nominal figure.

How do I narrow down 51 series quickly?

Filter in four steps: range, then mounting, then output form, then temperature range. That usually converges to two or three series, after which you compare the specific model datasheets item by item.

Need a similar solution?

Our engineers survey site conditions and design the solution; configuration and results are based on on-site measurement.