[Feature request] driver for AW86224 haptic device

[lyusupov]

HI Lewis!

At the moment you have only one haptic device supported by your library

Just in case if you consider to add one more device into this list as an alternative to Texas Instruments DRV2605 - there is an AW86224 IC which is a part of LilyGO T-Display P4 product hardware design.

1. Datasheet: https://github.com/Xinyuan-LilyGO/T-Display-P4/blob/main/information/AW86224AFCR.pdf
2. https://github.com/Llgok/cpp_bus_driver/blob/master/src/chip/iic/aw862xx.cpp

[lewisxhe]

I've added it to the to-do list; I'm currently adding a new power chip. I'll add the chip you mentioned once this issue is resolved.

[lewisxhe]

Full support has been added.

[lyusupov]

@lewisxhe

Lewis, thank you for the update!

I've made few attempts to execute your sketch ( HapticDrv_AW86224_Full.ino ) on LilyGO T-Display P4 AMOLED built with ESP32 Arduino Core 3.3.7

It fails to me upon AW86224 init procedure

ESP-ROM:esp32p4-eco2-20240710
Build:Jul 10 2024
rst:0x1 (POWERON),boot:0x20f (SPI_FAST_FLASH_BOOT)
SPI mode:DIO, clock div:1
load:0x4ff33ce0,len:0x164c
load:0x4ff29ed0,len:0xd64
load:0x4ff2cbd0,len:0x3364
entry 0x4ff29eda
I (25) boot: ESP-IDF v5.5-dirty 2nd stage bootloader
I (26) boot: compile time Jan 20 2026 17:36:58
I (26) boot: Multicore bootloader
I (27) boot: chip revision: v1.0
I (29) boot: efuse block revision: v0.3
I (33) boot.esp32p4: SPI Speed      : 80MHz
I (36) boot.esp32p4: SPI Mode       : DIO
I (40) boot.esp32p4: SPI Flash Size : 16MB
I (44) boot: Enabling RNG early entropy source...
I (49) boot: Partition Table:
I (51) boot: ## Label            Usage          Type ST Offset   Length
I (57) boot:  0 nvs              WiFi data        01 02 00009000 00006000
I (64) boot:  1 phy_init         RF data          01 01 0000f000 00001000
I (70) boot:  2 factory          factory app      00 00 00010000 00f00000
I (78) boot: End of partition table
I (80) esp_image: segment 0: paddr=00010020 vaddr=40040020 size=14898h ( 84120) map
I (103) esp_image: segment 1: paddr=000248c0 vaddr=30100000 size=00068h (   104) load
I (105) esp_image: segment 2: paddr=00024930 vaddr=4ff00000 size=0b6e8h ( 46824) load
I (117) esp_image: segment 3: paddr=00030020 vaddr=40000020 size=33048h (208968) map
I (154) esp_image: segment 4: paddr=00063070 vaddr=4ff0b6e8 size=05410h ( 21520) load
I (159) esp_image: segment 5: paddr=00068488 vaddr=4ff10b00 size=02e20h ( 11808) load
I (167) boot: Loaded app from partition at offset 0x10000
I (167) boot: Disabling RNG early entropy source...
E (932) esp_core_dump_flash: No core dump partition found!
E (932) esp_core_dump_flas
╔������������������������������������������╗
â•‘     AW86224 Haptic Driver Test           â•‘
╚�������������������������������������������

     0  1  2  3  4  5  6  7  8  9  a  b  c  d  e  f
00: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- 
10: -- -- -- -- -- -- -- -- 18 -- -- -- -- -- -- -- 
20: -- -- -- -- -- -- -- -- 28 -- -- -- -- -- -- -- 
30: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- 
40: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- 
50: -- -- -- -- -- -- -- -- 58 -- -- -- -- -- -- -- 
60: -- -- -- -- -- -- -- -- 68 -- -- -- -- -- -- -- 
70: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- 

=== Driver Initialization ===
Wire: SDA=20, SCL=21
[FATAL] AW86224 init failed!

To make sure that the AW86224 is present and that it works in general - I've installed the factory firmware first
https://github.com/Xinyuan-LilyGO/T-Display-P4/tree/main/firmware/%5BT-Display-P4%5D%5Blvgl_9_ui%5D
The AW86224 produces a vibration with factory firmware just fine.

I will continue to bring your sketch up...
Just curious if you have any suggestions from you to me to follow ?

[lyusupov]

I am not sure that HapticDriver_AW86224::parseChipID() is reliable enough.

When I've enforced _chipID = static_cast<uint8_t>(ChipID::AW86224); as follows:

#if 0
    if (reg == 0x00) {
        uint8_t ef_id = readReg(REG_CHIPID);
        if ((ef_id & 0x41) == 0x01) _chipID = static_cast<uint8_t>(ChipID::AW86223);
        else if ((ef_id & 0x41) == 0x00) _chipID = static_cast<uint8_t>(ChipID::AW86224);
        else return false;
    } else if (reg == 0x01) {
        uint8_t ef_id = readReg(REG_CHIPID);
        if ((ef_id & 0x41) == 0x01) _chipID = static_cast<uint8_t>(ChipID::AW86214);
        else return false;
    } else {
        return false;
    }
#else
     _chipID = static_cast<uint8_t>(ChipID::AW86224);
#endif

the sketch has started to work properly:

     0  1  2  3  4  5  6  7  8  9  a  b  c  d  e  f
00: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- 
10: -- -- -- -- -- -- -- -- 18 -- -- -- -- -- -- -- 
20: -- -- -- -- -- -- -- -- 28 -- -- -- -- -- -- -- 
30: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- 
40: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- 
50: -- -- -- -- -- -- -- -- 58 -- -- -- -- -- -- -- 
60: -- -- -- -- -- -- -- -- 68 -- -- -- -- -- -- -- 
70: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- 
=== Driver Initialization ===
Wire: SDA=20, SCL=21
[OK] Driver initialized
  Chip Name:    AW86224
  Chip ID:      0x2
  Actuator:     LRA
  F0:           1811 Hz
  VBAT:         3306 mV
  RAM Waveforms: 4
  F0 Cali Data: 0x1B
[TEST] Device Capabilities
  F0 Calibration:    Yes
  VBAT Compensation: Yes
  RTP Mode:          Yes
  Sequence:          Yes
  Break Effect:      Yes
  Continuous Mode:   Yes
  Max Effects:       4
  Max Sequence:      8
  Max Duration:      Unlimited
[OK] Done
========== Quick Test ==========
Running auto demo sequence...
=== Auto Demo Sequence ===
[1/12] Basic playEffect()...
[2/12] Basic short vibration...
[3/12] All waveforms (short)...
  Wave 1...
  Wave 2...
  Wave 3...
  Wave 4...
[4/12] Continuous vibration (HapticBase)...
[5/12] Long vibration (2s)...
[6/12] Intensity sweep...
[TEST] Intensity Sweep
  Gradually increasing intensity
  Gain: 0x20
  Gain: 0x40
  Gain: 0x60
  Gain: 0x80
  Gain: 0xA0
  Gain: 0xC0
  Gain: 0xE0
[OK] Done
[7/12] Non-blocking long vibration (3s)...
[8/12] Sequence test...
[TEST] setSequence() - HapticBase Interface
  Sequence set: [1, 2, 0]
  Playing sequence...
[OK] Done
[9/12] Status check...
[TEST] getStatus() - HapticBase Interface
  Status: IDLE
[OK] Done
[10/12] Rapid waveform switching...
[TEST] Waveform Switching (Non-Blocking)
  Rapid switching between waveforms during vibration
[OK] Done
[11/12] Gain test...
[TEST] shortVibration() - Different Gains
  Gain 0x20...
  Gain 0x40...
  Gain 0x60...
  Gain 0x80...
  Gain 0xA0...
  Gain 0xC0...
  Gain 0xFF...
[OK] Done
[12/12] Final short vibration...
=== Demo Complete ===
========== Commands ==========
Initialization:
  i   - Re-initialize driver

HapticBase Interface:
  e   - playEffect(1)
  E   - playEffect() all waveforms
  c   - continuousVibration()
  S   - setSequence() test
  p   - stop() test
  t   - getStatus()

AW86224 Extended:
  1   - Short vibration
  a   - All waveforms (short)
  g   - Different gains test
  l   - Loop count test (1-5)
  2   - Long vibration (blocking, 2s)
  L   - Long vibration all waveforms
  n   - Non-blocking long vibration (5s)
  w   - Waveform switching

Special Tests:
  x   - Sweep: gradual intensity increase
  d   - Auto demo sequence

Info:
  k   - Show capabilities
  v   - Read voltage
  f   - Read F0
  h   - Help (this message)
==============================
[?] Unknown: 'ÿ (0xFF)

[lewisxhe]

Replace the parseChipID function; I want to see what ID your device returns.

bool HapticDriver_AW86224::parseChipID()
{
    uint8_t reg = 0;
    for (int i = 0; i < 5; i++) {
        reg = readReg(0x00);
        if (reg != 0) break;
        hal->delay(2);
    }

    if (reg == 0x00) {
        Serial.printf("Read 0x00: 0x%02X\n", reg);
        uint8_t ef_id = readReg(REG_CHIPID);
        Serial.printf("Read CHIPID: 0x%02X\n", ef_id);
        if ((ef_id & 0x41) == 0x01) _chipID = static_cast<uint8_t>(ChipID::AW86223);
        else if ((ef_id & 0x41) == 0x00) _chipID = static_cast<uint8_t>(ChipID::AW86224);
        else return false;
    } else if (reg == 0x01) {
        Serial.printf("Read 0x01: 0x%02X\n", reg);
        uint8_t ef_id = readReg(REG_CHIPID);
        Serial.printf("Read CHIPID: 0x%02X\n", ef_id);
        if ((ef_id & 0x41) == 0x01) _chipID = static_cast<uint8_t>(ChipID::AW86214);
        else return false;
    } else {
        Serial.printf("Unexpected value at 0x00: 0x%02X\n", reg);
        return false;
    }
    return true;
}

In my tests, the ID was successfully read and the operation was executed.

[lyusupov]

Replace the parseChipID function; I want to see what ID your device returns.

     0  1  2  3  4  5  6  7  8  9  a  b  c  d  e  f
00: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- 
10: -- -- -- -- -- -- -- -- 18 -- -- -- -- -- -- -- 
20: -- -- -- -- -- -- -- -- 28 -- -- -- -- -- -- -- 
30: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- 
40: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- 
50: -- -- -- -- -- -- -- -- 58 -- -- -- -- -- -- -- 
60: -- -- -- -- -- -- -- -- 68 -- -- -- -- -- -- -- 
70: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- 
=== Driver Initialization ===
Wire: SDA=20, SCL=21
Read 0x01: 0x01
Read CHIPID: 0xFF
[FATAL] AW86224 init failed!

Lewis, I have one more side note -
I can feel that haptic.begin() test sketch procedure activates one vibration session.
It is expected that begin() should initialize the IC only but not to vibrate with the motor.

Is it a bug or feature?

[lyusupov]

@lewisxhe
Lewis, I have one more log for you.
It was taken with another T-Display P4 that has TFT display.

     0  1  2  3  4  5  6  7  8  9  a  b  c  d  e  f
00: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- 
10: -- -- -- -- -- -- -- -- 18 -- -- -- -- -- -- -- 
20: -- -- -- -- -- -- -- -- 28 -- -- -- -- -- -- -- 
30: -- -- -- -- -- -- 36 -- -- -- -- -- -- -- -- -- 
40: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- 
50: -- -- -- -- -- -- -- -- 58 -- -- -- -- -- -- -- 
60: -- -- -- -- -- -- -- -- 68 -- -- -- -- -- -- -- 
70: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- 
=== Driver Initialization ===
Wire: SDA=20, SCL=21
Read 0x00: 0x00
Read CHIPID: 0xB6
[OK] Driver initialized
  Chip Name:    AW86224
  Chip ID:      0x2
  Actuator:     LRA
  F0:           1761 Hz
  VBAT:         3318 mV
  RAM Waveforms: 4
  F0 Cali Data: 0xF

We can see that the sketch works fine on the T-Display P4 with TFT display.
I suppose that these devices may have different AW8622x IC hardware versions.

[lyusupov]

I have a question to @Llgok

Are there any GPIOs connected to Moto_INT and Moto_RSTN signals ?

[lewisxhe]

Test this

#include <HapticDrivers.hpp>
#include <SensorWireHelper.h>

#ifndef SENSOR_SDA
#define SENSOR_SDA  20
#endif

#ifndef SENSOR_SCL
#define SENSOR_SCL  21
#endif

HapticDriver_AW86224 haptic;

#ifdef ARDUINO_ESP32P4_DEV
#include <IoExpanderXL9555.hpp>
#define IO_EXPANDER_3V3_POWER_EN 0
#define IO_EXPANDER_5V0_POWER_EN 6
#define IO_EXPANDER_P4_VCCA_POWER_EN 8
#define P4_TOUCH_SDA  7
#define P4_TOUCH_SCL  8
IoExpanderXL9555 expander;
#endif

void testInit()
{
    Serial.println("\n=== Driver Initialization ===");
    Serial.print("Wire: SDA="); Serial.print(SENSOR_SDA); Serial.print(", SCL="); Serial.println(SENSOR_SCL);

    Wire1.begin(SENSOR_SDA, SENSOR_SCL);

#ifdef ARDUINO_ESP32P4_DEV
    Wire.begin(P4_TOUCH_SDA, P4_TOUCH_SCL);
    SensorWireHelper::dumpDevices(Wire);

    if (!expander.begin(Wire, XL9555_SLAVE_ADDRESS0)) {
        while (1) {
            Serial.println("Failed to find XL9555 - check your wiring!");
            delay(1000);
        }
    }
    Serial.println("Sensor Expander Initialized");

    expander.pinMode(IO_EXPANDER_P4_VCCA_POWER_EN, OUTPUT);
    expander.digitalWrite(IO_EXPANDER_P4_VCCA_POWER_EN, LOW);

    expander.pinMode(IO_EXPANDER_5V0_POWER_EN, OUTPUT);
    expander.digitalWrite(IO_EXPANDER_5V0_POWER_EN, HIGH);

    expander.pinMode(IO_EXPANDER_3V3_POWER_EN, OUTPUT);
    expander.digitalWrite(IO_EXPANDER_3V3_POWER_EN, LOW);

#endif

    Wire1.begin(SENSOR_SDA, SENSOR_SCL);
    SensorWireHelper::dumpDevices(Wire1);
    if (!haptic.begin(Wire1, AW8624_SLAVE_ADDRESS, SENSOR_SDA, SENSOR_SCL)) {
        Serial.println("[FATAL] AW86224 init failed!");
        while (1) delay(1000);
    }

    Serial.println("[OK] Driver initialized");
    Serial.print("  Chip Name:    "); Serial.println(haptic.getChipName());
    Serial.print("  Chip ID:      0x"); Serial.println(haptic.getChipID(), HEX);
    Serial.print("  Actuator:     "); Serial.println(haptic.getActuatorType() == HapticActuatorType::LRA ? "LRA" : "ERM");
    Serial.print("  F0:           "); Serial.print(haptic.getF0()); Serial.println(" Hz");
    Serial.print("  VBAT:         "); Serial.print(haptic.getVbat()); Serial.println(" mV");
    Serial.print("  RAM Waveforms: "); Serial.println(haptic.getRamNum());
    Serial.print("  F0 Cali Data: 0x"); Serial.println(haptic.getF0CaliData(), HEX);
}

void setup()
{
    Serial.begin(115200);
    delay(1000);
    testInit();
}

void loop()
{
}

[lewisxhe]

The motor vibrates once during initialization; that's to calibrate motor f0.

I have three Display-P4 displays:

Display-P4 version 1 (AMOLED) works without requiring any power-on operation;

Display-P4 version 2 (AMOLED) requires power-on to detect the motor;

Display-P4 version 3 (TFT) detects the motor and executes commands without requiring power-on operation.

I suspect the issue is caused by 3V3 not being enabled.

[lewisxhe]

The reset pull-up resistor for the AW86224 is in the Low3v3 power domain. The fact that Low3v3 is not enabled is causing the AW86224 to malfunction. I think this is the reason.