Company products - rare problems with STM32 MCU ADC

xinbako

Company products - rare problems with STM32 MCU ADC [Copy link]

The company has been mass-producing models for more than a year, with an average of 1,000 units per month. One of the motherboards uses STM32F051C8T6. Recently, a rare problem with the STM32 microcontroller was discovered (not sure if it has occurred before but was not discovered). The power supply voltage is normally 3.3V, but the voltage of all AD input ports of the microcontroller (each board uses 4 AD ports) becomes about 2.8V. About one out of every 100 units will have this problem, as shown in the figure below: WENDU2 is connected to the AD input port of the microcontroller, the CN10 terminal is disconnected and not connected to anything, the upper end of R65 is normally 3.3V, and the lower end of R65 becomes 2.8V, causing inaccurate ADC testing, especially the full scale can only reach 2.8V. It seems that there is a resistor inside the microcontroller, or the AD module is damaged. It has been confirmed that the chip is bad. The problem has been solved by replacing the chip. I just want to know what may be the cause of the chip damage?

Now I have the following doubts:

1. The PCB supplier purchased a non-genuine MCU. Because STM32 is out of stock recently, the supplier once told us that they also purchased from everywhere and were not sure about the reliability of the source.

2. The PCB supplier's production equipment or production process damages the chip

3 The circuit hardware has defects. The only suspicious point is that several MCU pins are used directly without protection, as shown in the figure below (TDS_PIN2, TDS_PIN1 are used directly as output ports to control the output high and low levels. These two pins are not protected by series resistors in order to test accuracy. The TDS_VALUE pin should be connected in series with a large resistor. This is indeed an oversight. I don’t know if the impact is significant. The CN12 terminal is connected to a probe located in a water tank). However, the MCU has its own protection clamping diode, and the chip with the problem damages the ADC. The input and output of these two pins are normal, so it is unlikely.

Jacktang

It may be that the AD module is broken. What is the reason?

damiaa

You can also remove the soldered resistor (such as 10K, preferably not more than 27K) at C21, and then measure the difference between the upper end voltage of this resistor and the calculated voltage, and you will know the reason. Is it a chip problem or not?

xinbako

damiaa posted on 2021-9-28 10:06 Connect a resistor to C21, or remove the soldered resistor (such as 10K, preferably not more than 27K), and then measure the voltage at the upper end of the resistor...

This problem has been going on for a long time. We have determined that the chip is broken. We have tried to replace the chip and the problem has been solved. I just want to know what caused the chip to be damaged.

w494143467

There must be a reason why the module is broken. It can't be because of poor quality.

freebsder

Consider whether static electricity leaked into it and damaged it.

xmj0615

Replacing the temperature module

szkalwa

It could also be a refurbished chip, the quality of which varies.