- Hacker Embedded STM32: Understanding ADC Oversampling, Averaging, and Noise Reduction

STM32 HAL Tutorial: High-Precision Analog Measurements

Abstract

Learn how to implement ADC oversampling, averaging, and noise reduction on STM32 using HAL. Step-by-step guide for high-precision analog measurements.

1. Introduction

Real-world analog signals often contain noise or require higher resolution than the ADC’s native capabilities.

STM32 ADCs support techniques to improve measurement quality:

Oversampling – Sampling multiple times per conversion to increase effective resolution
Averaging – Computing the mean of multiple samples to reduce noise
Noise reduction – Using hardware and software features to stabilize readings

By the end of this episode, you’ll be able to:

Implement ADC oversampling and averaging
Reduce measurement noise
Achieve higher precision readings without external hardware

2. Prerequisites

STM32 board with ADC
STM32CubeIDE installed
Knowledge of HAL, Interrupts and ADC

3. Oversampling Basics

- Concept: Take N samples and sum them → divide by N
- Effect: Each factor of 4 in oversampling increases resolution by 1 bit
- Formula: V_out =ΣNi=1ADC_i /N
- MCU ADC oversampling can be hardware-assisted or manually in software

4. HAL Example: Software Oversampling

Reduces random noise and increases effective resolution:

				
					#define OVERSAMPLE_COUNT 16
uint32_t adcSum = 0;
uint16_t adcValue;

for (int i = 0; i < OVERSAMPLE_COUNT; i++)
{
    HAL_ADC_Start(&hadc1);
    HAL_ADC_PollForConversion(&hadc1, HAL_MAX_DELAY);
    adcSum += HAL_ADC_GetValue(&hadc1);
}

adcValue = adcSum / OVERSAMPLE_COUNT; // Averaged and oversampled value

5. Using ADC Hardware Oversampling (if supported)

STM32H7, STM32G4, and STM32G0 MCUs support hardware oversampling
Configure in CubeMX:
- Oversampling ratio (2x, 4x, 8x, …)
- Right shift (divide accumulated value automatically)

				
					ADC_OversamplingTypeDef OversamplingConfig;
OversamplingConfig.Ratio = ADC_OVERSAMPLING_RATIO_16;
OversamplingConfig.RightBitShift = ADC_RIGHTBITSHIFT_4;
HAL_ADCEx_ConfigOversampling(&hadc1, &OversamplingConfig);

ADC returns averaged and higher resolution value automatically

6. Noise Reduction Techniques

Increase ADC sampling time → reduces input impedance mismatch noise
Enable internal voltage reference (VREFINT) → improves stability
Filter software data: moving average, exponential smoothing
Use differential mode if supported for sensor signals

7. Hands-On Lab Example

1. Configure the selected analog channel

2. Configure ADC oversampling (hardware)

3. Configure the DMA

4. Configure the desired clock

5. Read multiple samples and compute average by hardware

				
					/* USER CODE BEGIN PV */
uint16_t u16ADCBuff[16];
/* USER CODE END PV */

				
					 /* USER CODE BEGIN 2 */
 HAL_ADCEx_Calibration_Start(&hadc1);
 HAL_ADC_Start_DMA(&hadc1, (uint32_t *) u16ADCBuff, 16);
 /* USER CODE END 2 */

6. Compare noisy single reading vs oversampled averaged reading

a. Oversampling

b. Single Read

7. Optionally plot results on PC via UART

Tip: Combine hardware oversampling + DMA + low-pass software filtering for ultra-stable measurements

8. Advantages

Improves effective resolution without external ADCs
Reduces noise from sensors and analog circuits
Compatible with DMA and low-power modes
Ideal for precision measurement, sensor fusion, and control applications

STM32: Understanding ADC Oversampling, Averaging, and Noise Reduction