Engineering Guide
Integrating an RCD sensor on a charger PCB: wiring, self-test and common mistakes
Practical layout and firmware notes from our datasheets: power supply, signal interface, self-test timing, placement near relays and the mistakes we see most often in first prototypes.
A residual current sensor is only as good as its integration. The same part that passes every test on the bench can give nuisance trips or miss a fault if it is powered from a noisy rail, placed next to a relay, or never calibrated. This guide collects the practical points from EVsense datasheets, together with the mistakes we most often see in first prototypes. It applies to packaged sensors such as the 2R4Z and 6R2Z, and in most points to chip + core solutions such as the CT1836 with the ILM2112TQC IC.
The interface in brief
Packaged EVsense Type A + DC 6mA sensors use a simple interface:
| Signal | Direction | Function |
|---|---|---|
| VCC | Input | 5 V supply (4.8 V min, 5.2 V max) |
| GND | — | Ground |
| TRIP | Output | Digital leakage fault output (high = 4.2 V min, low = 0.6 V max) |
| CLK&CHK | Input | Starts zero calibration and self-test |
Some models add more pins. The 2R4Z also brings out voltage transformer and current transformer terminals, and the 6R2Z brings out its current transformer. The chip solutions (CT1836, CT1719, CT2728 with the ILM2112TQC) also have configuration pins CF0, CF1 and CF2, which must be set to match the core model, and test points T1 to T4, which must be reserved on the PCB.
Power supply: clean 5 V matters
DC residual current detection measures very small signals: the smooth DC trip threshold is 3 mA min, 4.5 mA typical and 6 mA max. Supply noise can therefore affect accuracy. The datasheets recommend:
- generating VCC with a dedicated 5 V LDO regulator,
- keeping ripple below 30 mV, and
- staying within the absolute maximum rating (−0.3 V to 6 V).
Do not feed the sensor directly from a switching regulator output that also supplies relays, displays or communication modules. Place the LDO and its decoupling capacitors close to the sensor.
Start-up and self-test sequence
Getting the timing right is the single most important firmware task. The sequence from the datasheets is:
- Power-on blanking. After power is applied, the sensor has an 800 ms start-up delay during which it does not respond or process signals. Do not read TRIP during this time.
- Wait. The timing diagram recommends waiting at least 1 s after power-on is complete (t1) before starting calibration.
- Trigger zero calibration and self-test. Drive CLK&CHK for 0.6 s to 1.2 s (t2). The sensor calculates its zero point during power-on and self-test.
- Zero-clearing. Allow 380 ms to 450 ms (t3) for the sensor to complete zero-clearing.
- Self-test. The sensor injects a test signal and drives TRIP; this takes 200 ms to 400 ms (t4). When TRIP returns to low level, the self-test has passed and normal residual current detection begins.
Run this sequence once at power-on and again before each charging session.
During calibration and zero-clearing, the charger’s main relay should be closed only after the TRIP signal sequence has completed. If load current flows while the sensor sets its zero point, the calibration will be wrong.
Your firmware should also treat a missing self-test response as a fault. If TRIP does not go high and then low as expected, do not start the session, and report a protection fault to the user or back office.
Placement on the board and in the enclosure
- Install the sensor at the output side of the charger. The datasheets specify placement at the output terminal of the charging pile, so that the residual current of the vehicle circuit is measured.
- Keep it more than 3 cm from the relay. Relay coils and contacts create magnetic fields that can disturb the sensor. Keep the minimum distance and avoid routing relay coil current under the sensor.
- Route all current-carrying conductors through the core. The sensor measures the difference between outgoing and returning current. Line and neutral (all three phases plus neutral in a three-phase charger) must pass through the aperture together. The protective earth conductor must not pass through.
- Keep conductors centred and fixed. Use cable ties, clips or the pre-wired sensors (such as 4R2Z, CT1719 or CT2728) so that the conductor position does not change after assembly.
Signal routing to the controller
- Keep the TRIP and CLK&CHK traces short and away from high-current and switching nodes.
- If your MCU uses 3.3 V I/O, check the input levels. The ILM2112TQC CHK input is a 5 V TTL-level input, and the datasheets suggest a transistor stage for 3.3 V to 5 V conversion. TRIP high level is at least 4.2 V, so use a divider or a 5 V-tolerant input on the MCU side.
- Treat TRIP as a safety signal: the controller should open the main relay immediately on a TRIP and log the event. Never filter it in software for longer than your target standard allows.
Common mistakes in first prototypes
- Skipping the pre-session self-test. The sensor is only calibrated at power-on, and offset drifts over weeks of operation without a power cycle.
- Closing the relay during calibration. Load current is present while the zero point is set, causing nuisance trips or reduced sensitivity.
- Noisy 5 V supply. VCC is shared with relays or a buck converter, and ripple exceeds 30 mV.
- Sensor too close to the relay. Mounting it within 3 cm of the relay for a compact layout leads to false TRIP signals when the relay switches.
- Earth conductor through the core. The PE wire is routed through the aperture with the phase conductors, which hides real earth faults.
- Not all conductors through the core. In a three-phase design, neutral is routed outside the sensor, so normal load current appears as residual current.
- Reading TRIP during the 800 ms start-up window. The firmware interprets start-up behaviour as a fault.
- No test points on chip solutions. T1 to T4 are left out of the PCB, so production testing and debugging become difficult.
Validate before production
After integration, test the complete charger, not just the sensor: smooth DC, AC and pulsating DC residual currents at your target thresholds, with the relay closed and the charger under load. EVsense has in-house Type B leakage current testing capability and can review your test results during sample evaluation.
If you are still choosing a model, read how to choose an RCD sensor for a wallbox. For the physics behind the self-test and zero calibration, see how 6mA DC residual current detection works. To get reference circuits and samples for your design, send us your project details.
Frequently asked questions
How often should the RCD sensor self-test run?
EVsense datasheets recommend a self-test and zero calibration through the CLK&CHK input once at power-on and again before each charging session.
Can I close the main relay while the sensor is calibrating?
No. During calibration and zero-clearing, the charger's main relay should only be closed after the TRIP signal sequence has completed, so that the sensor sets its zero point with no load current flowing.
How should I power the RCD sensor?
Use a 5 V supply (4.8 V to 5.2 V). The datasheets recommend generating it with a 5 V LDO regulator with ripple below 30 mV, because supply noise can affect the accuracy of DC residual current measurement.
Can a 3.3 V microcontroller drive the CHK input?
The CHK input of the ILM2112TQC chip solutions is a 5 V TTL-level input. Because charger MCUs often use 3.3 V I/O, the datasheets suggest a transistor stage to convert 3.3 V to 5 V.
Need help choosing a sensor?
Send your charger platform, target standard, rated current and PCB space. Our engineers reply with a model recommendation, datasheets and sample options.