ESP32-S3-WROOM-1,IO0 AND IO47 PUII-UP?Why voltage unusual

[emyjob]

I made a product using ESP32-S3-WROOM-1 and put it into mass use. After working continuously for about a month when powered on, I found the following problem: the product uses IO0 (pin 27 of the module) and IO47 (pin 24 of the module) each connected in series with a 330R resistor, and the external wiring is led out to the button on the small board for user input. The program is configured for internal pull-up, and each IO at the wiring is placed with an electrostatic protection tube SM05. Of course, there are other pins that also have series resistors and lead out button inputs. However, only these two IOs have been found to have problems so far. The problem is that after the first period of operation, although there was no button pressed, the program detected that the button was pressed. The voltage of the pin was measured by a multimeter to be below 2.5V, and some boards even had pin voltages as low as around 1.8V. Some boards also have a pull-up voltage of 3.9V (I'm sure I read it correctly, my board's power supply voltage is 3.3V). After replacing the new module, it returned to normal. Due to the large number of faults (currently 6% of which are still being discovered...), some boards have faults with IO47, some boards have faults with IO0, and some boards have faults with both. Overall, the fault rate of IO47 is much higher than that of IO0. So I started looking for the cause of the problem, and later found that when measuring the voltage of these two pins with a multimeter, different pull-up voltages would randomly appear. Then, when I touched the multimeter rod with my hand, the highest voltage could reach 4.7V, confirming that I had read correctly. Touching the series resistor with your hand produces the same result as the solder pads on the buttons on the external board, even if a new module is replaced without burning the program. So I suspected that there was a problem with these two IOs. Could it be a problem with my own hardware design? Because I bought an officially designed module, I never thought I would suspect that there might be a problem with the official hardware at first. After reviewing the hardware design guide again, I didn't find any issues with the layout. So I thought of using a new module, without soldering it to the board, to directly connect 3.3V and GND with the power supply, connect EN to pull up, measure IO0 directly with an external meter, and try to touch the IO port with my hand. The good guy jumped out at 4.6V. In order to eliminate the issue with the multimeter, the multimeter was replaced and the meter stick was touched with both hands without any problems. At this point, the problem is with these two pins of the module or S3 chip. After flipping through the boxes and cabinets, I found a C3 module sample that I had purchased before. I measured it in the same way and found it to be very good and normal. It seems that not all ESP models have problems. So the ultimate problem is whether there is a problem with the module design or an internal issue with the chip. I just want to know what to do now? I'm so worried, thousands of products are out there..... Now we plan to replace series resistors with diodes on a small number of boards to try replacing products on site.

Please ask the expert how to solve it.

[MicroController]

Some boards also have a pull-up voltage of 3.9V (I'm sure I read it correctly, my board's power supply voltage is 3.3V).

This is very unlikely. When fed from a (stable) 3.3V supply, the source of the 3.9V is pretty certainly not the ESP.
Once you've verified your meter is ok, and you're measuring correctly, you could share the circuitry connected to the pin(s) in question.

[emyjob]

At first, I didn't believe it, and I also suspected that there was a problem with the measurement. However, before I suspected it, many modules of my ESP32-S3 were already broken. After pulling up, the voltage level was only about 2V, and the code was judged to be low. It was because the pins were already broken that I started measuring them with a voltmeter, which revealed the problem. Moreover, this phenomenon has been found on ESP32-S3-WROOM-1, but not on ESP32-C3 module.
It's very simple now. I don't have any other peripheral circuits, only the EN pin pulls up 3.3V, and the external power supply 3.3V. After powering on, touching the pin with my hand will cause problems with voltages higher than 3.9V.

I have replaced the power supply and the voltmeter, but the measurement results are the same.

20240722083201.jpg (162.61 KiB) Viewed 1244 times

20240722083147.jpg (101.77 KiB) Viewed 1244 times

pics link:
http://www.elitebaba.com/download/20240722083550.jpg
http://www.elitebaba.com/download/20240722083539.jpg
http://www.elitebaba.com/download/20240722083201.jpg
http://www.elitebaba.com/download/20240722083147.jpg
http://www.elitebaba.com/download/20240722084859.png
http://www.elitebaba.com/download/20240722084911.png

[emyjob]

picture1

20240722084911.png (789.26 KiB) Viewed 1249 times

picture2

20240722084859.png (889.19 KiB) Viewed 1249 times