When makers talk about relay coding, they usually mean the simple digitalWrite(RELAY_PIN, HIGH) command used to switch a load. But in professional electromechanical design, relay coding is the critical bridge between microsecond microcontroller logic and the millisecond physical realities of moving metal armatures. If your firmware ignores the physics of the relay coil and the thermodynamics of the contacts, you will end up with welded contacts, fried GPIO pins, and erratic system resets.
This guide translates physical relay specifications into actionable firmware logic and wiring practices, ensuring your MCU survives the inductive kickback and your contacts survive the inrush current.
Decoding the Spec Sheet: Coil vs. Contact Ratings
A relay is effectively two separate circuits sharing a magnetic core: the coil side (your low-voltage DC control) and the contact side (your high-voltage or high-current load). Understanding which rating governs your specific application is the first step in writing robust relay control code.
| Model (Manufacturer) | Coil Voltage / Current | Coil Resistance | Resistive Contact Rating | Max Breaking Capacity |
|---|---|---|---|---|
| Omron G5V-2-DC5 | 5V DC / 70mA | 71 Ω | 2A @ 30VDC / 0.5A @ 125VAC | 60W / 125VA |
| Songle SRD-05VDC-SL-C | 5V DC / 71mA | 70 Ω | 10A @ 250VAC / 10A @ 30VDC | 2500VA / 300W |
| Finder 36.11.9.005 | 5V DC / 40mA | 125 Ω | 10A @ 250VAC | 2500VA |
| Panasonic TQ2-5V | 5V DC / 28mA | 178 Ω | 2A @ 30VDC / 1A @ 125VAC | 60W / 125VA |
Which Rating Column Governs Your Load?
The most common mistake in relay selection is looking only at the Resistive Contact Rating (e.g., '10A @ 250VAC'). If you are switching a motor, a solenoid, or a transformer, the resistive rating is meaningless. For inductive and motor loads, the Maximum Breaking Capacity (measured in VA or Watts) and the specific Inductive/Motor Derating Curve govern the limit. A 10A resistive relay will typically only handle 3A to 5A for an AC motor load due to the severe arcing caused by the motor's inrush current and inductive kickback upon opening.
Coil Wiring and the Mandatory Flyback Diode
The coil side is an inductor. When your MCU drops the GPIO pin LOW to turn off the relay, the collapsing magnetic field generates a massive reverse voltage spike (inductive kickback). Without protection, this spike will avalanche your driving transistor and eventually destroy the MCU's GPIO bank.
Load Selection Decision Path & Firmware Timing
Relay coding must account for the physical time it takes for the armature to move. A typical 5V relay has an operate time of 5ms and a release time of 5ms. If your code attempts to PWM or toggle the relay faster than this physical limit, the armature will hover in the middle, generating massive heat and failing to make contact.
| Load Type | Physical Hazard | Hardware Mitigation | Firmware / Coding Action |
|---|---|---|---|
| Resistive (Heaters, Incandescent) | Minimal inrush, low arcing. | Standard relay sizing. | Standard toggle. Add 10ms software delay after state change to allow armature settle before reading feedback. |
| Inductive (Solenoids, Valves) | Severe arcing on contact break. | RC Snubber across contacts (e.g., 100Ω + 0.1µF). | Implement a mandatory 'cool-down' timer in code. Prevent rapid cycling (limit to max 1 switch per second). |
| Motor (Compressors, Pumps) | High inrush (LRA) welds contacts. | Use a contactor or Solid State Relay (SSR). Derate EMR by 70%. | Code a hard interlock: require a 3-second dead-time between reversing polarity or restarting to prevent back-EMF spikes. |
| Capacitive (LED Drivers, SMPS) | Inrush current acts like a short. | NTC Thermistor in series with load. | Implement 'pre-charge' logic if using multiple staged relays, closing a pre-charge relay 50ms before the main relay. |
When writing the firmware, always use non-blocking timers (like millis() in Arduino or FreeRTOS software timers on the ESP32) to enforce these physical delays. Never rely on delay() in a production environment, as it halts the MCU and prevents safety interlocks from being monitored.
Firmware Implementation: Debounce and State Machines
If your code reads the state of a relay's auxiliary contacts for feedback, you must treat the relay contacts exactly like a mechanical pushbutton: they bounce. When the armature slams into the normally-open (NO) contact, it physically bounces for 1ms to 3ms, creating a rapid series of open/close spikes.
According to Texas Instruments application notes on relay drive circuits, ignoring contact bounce in feedback loops can cause microcontrollers to register multiple false triggers. Your relay coding must include a software debounce routine—ignoring state changes on the feedback pin for at least 5ms after the initial transition.
Furthermore, structure your relay control as a Finite State Machine (FSM). Instead of simple IF/THEN logic, define states: IDLE, ENERGIZING, SETTLED, DE-ENERGIZING, FAULT. This prevents race conditions where a user commands a motor to reverse before the forward relay has physically released and broken the circuit.
Diagnostics: Testing Dead, Live, and When to Replace
When a relay circuit fails, you need to determine if the fault lies in the MCU code, the driving transistor, the coil, or the contacts. Here is the definitive diagnostic sequence.
How to Test Dead (Power Removed)
- Coil Integrity: Set your multimeter to resistance (Ω). Probe the coil pins (A1 and A2). You should read the exact coil resistance listed in the spec sheet (e.g., ~70Ω for a Songle 5V relay). An infinite reading means an open coil; a reading near 0Ω means a shorted coil. Both require replacement.
- Contact Continuity: Set the meter to continuity. Probe the Common (COM) and Normally Closed (NC) pins. It should read < 1Ω. Manually press the armature with a non-conductive tool; the COM to Normally Open (NO) pins should now read < 1Ω. If resistance is high or erratic, the contacts are pitted or carbon-fouled.
How to Test Live (Energized)
Safety Warning: Live testing involves mains or high-current DC. Use an oscilloscope with isolated probes or a CAT-III multimeter. De-energize and lock-out/tag-out before making physical connections.
- Flyback Clamp Verification: Connect an oscilloscope across the relay coil. Trigger on the falling edge when the MCU turns the relay off. You should see the voltage spike clamped to roughly VCC + 0.7V. If you see a spike exceeding 20V on a 5V system, your flyback diode is missing, backwards, or blown.
- Contact Voltage Drop: With the relay energized and the load running, measure the DC voltage drop across the COM and NO pins. A healthy relay will drop less than 50mV. If you read >200mV, the contacts are degraded, generating excessive heat (I²R losses), and are nearing failure.
When to Repair vs. Replace
The golden rule of modern electromechanical relays: Never attempt to repair a sealed PCB relay. If the contacts are welded shut, pitted, or the coil is open, the component is e-waste. Filing down contacts destroys the factory-applied plating (usually silver-tin-oxide or gold-flashed), which will cause immediate re-welding upon the next use.
However, if a relay fails prematurely (e.g., contacts welding after only a few hundred cycles), the fault is usually in the relay coding or hardware design, not the relay itself. If you find welded contacts on a motor load, do not just replace the relay. You must rewrite the code to enforce longer dead-times, add an RC snubber across the contacts, or upgrade to a Solid State Relay (SSR) or a heavy-duty contactor designed for high L/R time constants. For deeper standards on contact arcing and electrical endurance, refer to the IEC 61810-1 standard for electromechanical elementary relays.
By treating relay coding as a physical engineering problem rather than a simple software toggle, you ensure your embedded systems operate reliably for millions of cycles, rather than failing catastrophically on the workbench.
