I have a batch of cortex-M3 boards, some of which are slow to power on (the startup program is in the FLASH inside the MCU), and it takes two or three minutes to start, while the normal boards start in a few seconds after powering on.
1. For the slow-starting board, the voltage is normal before starting, but the 2 crystal oscillators (32.768K/12M) have not started to vibrate.
2. After the slow-starting board is powered on once, it is normal to restart after power-off, but after a few hours or put it under the air conditioner for a minute or two to cool down, the start-up becomes slow again. Yes (heating but not enough to melt tin) can be normal again. At this time, the MCU welding problem is ruled out, and it is suspected that it is the part of the crystal oscillator.
3. Replace the MCU, 2 crystal oscillators and load capacitors on the normal board and the slow-start board. The original slow-start board still starts slowly, and the original normal board is also OK.
4. The load capacitance of the slow-starting board was changed to a smaller value according to the requirements of the MCU specification, and did not try to become larger.
Have you ever encountered such a situation, please give me some suggestions.
Savannah postato su October 15, 2020
Could it be a problem with the reset circuit?
Luis postato su October 15, 2020
Check the reset circuit.
Carlos postato su October 15, 2020
Pay attention to the parallel resistance of the crystal oscillator. Have you tried it?
Gabriel postato su October 15, 2020
There is no parallel resistor on the crystal oscillator, and the string resistor has been debugged, but it has no effect.
