ISM330IS ST | STMicroelectronics | iNEMO series 6-axis IMU with ISPU programmable core, 3-axis accelerometer + 3-axis gyroscope. The standout feature is the built-in 32-bit ISPU (Intelligent Sensor Processing Unit) that allows running C code and AI machine learning algorithms within the sensor itself, enabling edge computing at the device level. Ideal for industrial IoT, condition monitoring, and smart sensor nodes.

When MEMS sensors no longer merely output raw data but can execute algorithms and AI models within the chip itself, the ISM330IS is such a 6-axis IMU. The built-in ISPU programmable processing unit enables vibration analysis, anomaly detection, and motion recognition at the sensor end, significantly reducing the pressure on the main MCU and empowering industrial Internet of Things and intelligent sensing terminals.

Model: ISM330IS

Manufacturer: STMicroelectronics (STMicroelectronics)

Package: LGA-14, size 2.5 × 3.0 × 0.83 mm, no side pins on the bottom pad, ECOPACK eco-friendly package, RoHS compliant

Screen printing: abbreviated model on the top surface of the chip, with a small dot mark on the 1st pin; The complete model number will not be printed on the chip surface.

Power supply: VDD: 1.71-3.6V; Independent IO power supply 1.62 V ~ 3.6 V, suitable for multiple types of master level

Temperature range: ‑40 ℃ ~ +85 ℃

Communication interface: I2C, SPI; Built-in sensor hub, can connect up to 4 external sensors, expand sensing ability

WHO_AM_I: Chip identification ID, software driver identification key

Key features: Built-in ISPU intelligent sensor processing unit (32-bit RISC core), supports C language programming, paired with the NanoEdge-AI Studio toolchain, enabling the deployment of machine learning models. 32KB program RAM + 8KB data RAM, with FPU floating-point unit

Parameter item Typical indicator

Acceleration range ±2g / 4g / 8g / 16g programmable

Accelerometer noise 70 µg/√ Hz (@±2g)

Gyroscope range ±125/±250/±500/±1000/±2000dps programmable

Gyro noise 3.4 mdps/∘Hz

Gyro zero temperature drift ± 0.015dps/℃

Accelerometer zero temperature drift ± 0.1 mg/°C

6-axis high-performance mode current 0.59 mA (without ISPU operation)

Low power mode current 0.46 mA (without ISPU)

ISPU kernel 32-bit RISC Harvard architecture, 10MHz, with FPU floating-point unit

Storage resources 32KB program RAM, 8KB data RAM

FIFO cache Supports data batch processing, programmable interrupt output

Complementary toolchain: X‑CUBE‑ISPU, NanoEdge AI Studio, supports automated generation of AI models and C code development.

Application field materials (written in sections)

Industrial monitoring (core scenario)

Predictive maintenance and state monitoring of equipment: vibration detection of motors, pump bodies, and transmission mechanisms, ISPU local extraction of vibration features, identification of abnormal vibrations and signs of failure, no need for MCU to continuously read raw data.

Industrial robot joints: posture, vibration, impact detection, judging the operation status of the equipment.

Asset tracking and monitoring: detection of logistics assets, tilt, fall, and impact events of valuable equipment.

IoT and smart terminals

Intelligent alarm device, smart home event detection: tilt, vibration, impact event recognition, local judgment after the event triggers the interruption reporting to the host.

Battery-powered wireless sensor node: relying on ISPU local computation, MCU long-term hibernation, event wake-up, extending battery life.

Health and exercise recognition

Complex motion, gesture recognition, human body behavior recognition, ISPU locally performs motion feature classification.

⚠️Important boundaries (to improve the professionalism of the article)

Non-automotive grade components cannot be used for vehicle safety or autonomous driving.

Only 6 axes, no magnetometer, cannot directly output heading angle, navigation requires external magnetic sensor.

ISPU runs on RAM, the code is lost when the power is turned off, and the algorithm code needs to be loaded into ISPU when the MCU is powered on. This must be considered in product design.

The ISPU’s computing power is limited, making it suitable for lightweight signal processing and small AI models, but it cannot run large neural networks.

VII. Horizontal selection and comparison of materials

ISM330IS vs LSM6DSOX

LSM6DSOX: MLC machine learning kernel, hardware fixed operator, does not support C language programming; For consumer wearables and mobile phones.

ISM330IS: ISPU programmable 32-bit core, supports C language, floating-point operation, higher degree of freedom, industrial IoT, state monitoring, more flexible development.

ISM330IS vs ICM‑42670‑L（TDK）

ICM-42670-L: APEX hardware engine, hardware fixed motion detection function, cannot customize algorithm, mainly consumer headphones, wearables.

ISM330IS: ISPU programmable core; users can customize algorithms and deploy AI models, with a focus on industrial IoT and condition monitoring scenarios.

Common specifications of STMicroelectronics:

LIS2DWLTR

LSM6DSV16XTR  

LSM6DSV32XTR 

L9663-TR-1

LSM6DSOETR3

LSM6DSK320XTR

ISM330DHCXTR

ASM330LHHTR

SPC584B70E3EHC0X 

STM32H563MIY3QTR

STM32F0316DIE1

STM32G473VCT6

STM32H7B0ZBT6

STM32L053R8T6 

ST (STMicroelectronics) sensors; 3D accelerometer; Temperature sensing; ST sensor; Sensor Gyroscope

For technical and business consulting, please contact: 18277896808