STM32 drives INA226 to measure current, voltage, and power

 INA226 + STM32 is a rock-solid combo for precision DC current, bus voltage, and power. Here’s a clean, copy-ready guide with wiring, calibration math, and working HAL code.

---

STM32 + INA226 (I²C) — Measure Current, Bus Voltage & Power

1) What the INA226 gives you

• Bus voltage (VBUS): up to 36 V, LSB = 1.25 mV

• Shunt voltage (VSHUNT) across your sense resistor, LSB = 2.5 µV

• Current (after you program a calibration)

• Power (after calibration), Power_LSB = 25 × Current_LSB

• Programmable averaging, conversion times, and alert (over-current, over-voltage, power, etc.)

---

2) Typical wiring (single-supply)

• Shunt resistor (RSHUNT) in series with the load on the low side or high side (INA226 supports both; high side is common).

• Connect:

  • VIN+ to load side (after shunt toward load)

  • VIN− to source side (before shunt toward supply)

  • VBUS internally sensed from VIN+ (no extra wire for bus voltage)

  • SDA / SCL → STM32 I²C pins (e.g., PB9/PB8 on many Nucleo boards)

  • A0, A1 to GND/VCC for I²C address select (0x40–0x4F)

  • ALERT (optional) to an STM32 EXTI-capable GPIO for interrupts

  • Decoupling 0.1 µF close to INA226 VCC

I²C address: base 0x40 when A1=A0=GND (others up to 0x4F).

---

3) Key registers (16-bit)

Register	Addr	Notes
Configuration	0x00	Averaging, conv. times, mode
Shunt Voltage	0x01	2.5 µV/bit (signed)
Bus Voltage	0x02	1.25 mV/bit (unsigned)
Power	0x03	25×Current_LSB (W/bit), needs CAL
Current	0x04	Current_LSB (A/bit), needs CAL
Calibration	0x05	Sets Current_LSB & scaling
Mask/Enable	0x06	ALERT config
Alert Limit	0x07	Threshold (unit depends on mask)

All registers are big-endian (MSB first).

---

4) Calibration math (do this once!)

1. Choose RSHUNT and your IMAX (expected max current).

2. Pick Current_LSB ≈ IMAX / 32768 (A/bit).

3. Compute CAL (write to Calibration register):

   $$\text{CAL} = \frac{0.00512}{\text{Current\_LSB} \times R_{\text{SHUNT}}}$$

   (CAL is a 16-bit integer; clamp to ≤ 0x7FFF if needed.)

4. Then:

   • Current (A) = Current_Reg × Current_LSB

   • Power (W) = Power_Reg × (25 × Current_LSB)

Example

• RSHUNT = 0.010 Ω, IMAX = 3.0 A

• Current_LSB = 3.0 / 32768 = 9.1552734e-5 A/bit (≈ 91.55 µA/bit)

• CAL = 0.00512 / (9.1552734e-5 × 0.01) ≈ 5592 = 0x15D8

• Power_LSB = 25 × Current_LSB = 0.0022888 W/bit (≈ 2.289 mW/bit)

---

5) Recommended configuration

• Averaging (AVG): e.g., 16–64 (trade noise vs speed)

• VBUS/VSHUNT conversion time: 1100 µs (or 140 µs…8.244 ms)

• Mode: Shunt+Bus, continuous

Configuration register (0x00) fields (typical):

• AVG = 0b010 (16x) or 0b011 (64x)

• VBUSCT = 0b100 (1.1 ms)

• VSHCT = 0b100 (1.1 ms)

• MODE = 0b111 (Shunt+Bus continuous)

---

6) STM32 HAL code (I²C)

6.1 Helpers

#include "main.h"
extern I2C_HandleTypeDef hi2c1; // or your I2C handle

#define INA226_ADDR      (0x40 << 1) // 7-bit addr <<1 for HAL
#define REG_CONFIG       0x00
#define REG_SHUNT        0x01
#define REG_BUS          0x02
#define REG_POWER        0x03
#define REG_CURRENT      0x04
#define REG_CAL          0x05
#define REG_MASK_ENABLE  0x06
#define REG_ALERT_LIMIT  0x07

static HAL_StatusTypeDef ina226_write16(uint8_t reg, uint16_t val) {
    uint8_t buf[3];
    buf[0] = reg;
    buf[1] = (uint8_t)(val >> 8);
    buf[2] = (uint8_t)(val & 0xFF);
    return HAL_I2C_Master_Transmit(&hi2c1, INA226_ADDR, buf, 3, HAL_MAX_DELAY);
}

static HAL_StatusTypeDef ina226_read16(uint8_t reg, uint16_t *val) {
    uint8_t data[2];
    HAL_StatusTypeDef st = HAL_I2C_Master_Transmit(&hi2c1, INA226_ADDR, &reg, 1, HAL_MAX_DELAY);
    if (st != HAL_OK) return st;
    st = HAL_I2C_Master_Receive(&hi2c1, INA226_ADDR, data, 2, HAL_MAX_DELAY);
    if (st != HAL_OK) return st;
    *val = ((uint16_t)data[0] << 8) | data[1];
    return HAL_OK;
}

6.2 Init (config + calibration)

// Example: AVG=16, VBUSCT=1.1ms, VSHCT=1.1ms, MODE=Shunt+Bus Continuous
HAL_StatusTypeDef ina226_init(void) {
    uint16_t config = 0;
    // AVG[2:0]=0b010 (16x) -> bits 11:9
    config |= (0x2 << 9);
    // VBUSCT[2:0]=0b100 (1.1ms) -> bits 8:6
    config |= (0x4 << 6);
    // VSHCT[2:0]=0b100 (1.1ms) -> bits 5:3
    config |= (0x4 << 3);
    // MODE[2:0]=0b111 (Shunt+Bus continuous) -> bits 2:0
    config |= 0x7;

    if (ina226_write16(REG_CONFIG, config) != HAL_OK) return HAL_ERROR;

    // Calibration for Rshunt=0.01Ω, Imax=3A (CAL=0x15D8 from example)
    if (ina226_write16(REG_CAL, 0x15D8) != HAL_OK) return HAL_ERROR;

    return HAL_OK;
}

6.3 Read & convert to units

// Match these to your calibration choices:
static const float CURRENT_LSB = 9.1552734375e-5f; // A/bit (example)
static const float POWER_LSB   = 25.0f * 9.1552734375e-5f; // W/bit

HAL_StatusTypeDef ina226_read_measurements(float *bus_V, float *shunt_V,
                                           float *current_A, float *power_W) {
    uint16_t raw;

    // Bus voltage (1.25 mV/bit)
    if (ina226_read16(REG_BUS, &raw) != HAL_OK) return HAL_ERROR;
    *bus_V = (float)raw * 0.00125f;

    // Shunt voltage is signed (2.5 µV/bit)
    if (ina226_read16(REG_SHUNT, &raw) != HAL_OK) return HAL_ERROR;
    int16_t shunt_raw = (int16_t)raw;
    *shunt_V = (float)shunt_raw * 2.5e-6f;

    // Current (A) from CURRENT register
    if (ina226_read16(REG_CURRENT, &raw) != HAL_OK) return HAL_ERROR;
    int16_t curr_raw = (int16_t)raw;
    *current_A = (float)curr_raw * CURRENT_LSB;

    // Power (W) from POWER register
    if (ina226_read16(REG_POWER, &raw) != HAL_OK) return HAL_ERROR;
    *power_W = (float)raw * POWER_LSB;

    return HAL_OK;
}

---

7) Using ALERT (optional)

• Configure Mask/Enable to select alert source (e.g., over-current).

• Write Alert Limit with a threshold in the corresponding units (depends on alert function).

• Route ALERT to an EXTI pin; in ISR, read Mask/Enable to clear/identify cause.

Example idea: over-current alert → set Alert Limit = threshold / Current_LSB (in Current register’s units) and enable the Conversion Ready bit to pace reads.

---

8) Accuracy tips

• Choose RSHUNT so that your typical/peak current uses a healthy fraction of the shunt’s full-scale voltage (keep power dissipation acceptable: $P=I^2R$).

• Prefer low TCR (e.g., 50 ppm/°C) 1% or 0.5% shunts; 2512 or 1206 packages for thermal stability.

• Use averaging for noisy loads; raise conversion time for better resolution.

• Kelvin-connect the shunt; keep high-current paths short and away from sense lines.

• If measuring at high bus voltages, respect creepage/clearance on the PCB.

---

9) Quick checklist

• I²C at 100–400 kHz, pull-ups 2.2–4.7 kΩ

• Correct I²C address from A0/A1 strap

• Write CONFIG, then CAL

• Use correct LSB scales (1.25 mV, 2.5 µV, Current_LSB, 25×Current_LSB)

• Convert signed registers properly (shunt, current)

• Optional: set ALERT mask & limit, handle EXTI

---

10) Troubleshooting

• All zeros? Check CAL register was written and mode is continuous.

• Crazy values? Endianness, wrong LSBs, or mis-wired VIN+/VIN-.

• I²C NACK? Address pins (A0/A1), pull-ups, or wrong 8-bit vs 7-bit address in HAL.

• Noisy readings? Increase averaging and conversion time; improve shunt layout.