Summary:
MCU is a chip with a very high usage rate on the hardware circuit board. Many of its purposes are used to control IO, simple input and output high and low levels, or use IIC and SPI to read data from another device, etc. Therefore, the circuit design of MCU is especially important for hardware engineers. It is one thing to make it usable, but another thing to make it perfect.
The simple things like power supply filter capacitor and reset circuit will not be repeated here. Let’s talk about the overall design thinking of single-chip microcomputer.
For the circuit board, the essential is the power supply, DCDC control. Some high-end DCDC. Its chip will have an enable pin (EN) and a status detection pin (PG).
Enable control:
Enabling EN is to turn on the power supply output of this DCDC. The state detection PG is whether the output voltage is within the set range. For example, the setting is 5V output. If it is actually 5V output, the PG level is high (usually high Level, abnormally low, specific chip analysis), the following chips have these two signals, generally connected to MCU or CPLD, etc.
Generally speaking, there may be many different power supplies on a circuit board, such as the figure below, 0.8V, 0.9V, 1.1V... When it is necessary to control the switches of different power supplies, we assume that the circuit timing is based on P0.8V_EN has the highest priority, then detects P0.8V_PG... and finally SFP_3V3_PG. (For the MCU, EN is used as an output signal, and PG is used as an input detection signal).
The software needs to turn on the power supply of a certain power supply, and then check whether the PG of this power supply is in a normal state, and then turn on the power supply of the next power supply. Here, it needs to operate in a logical order, and the next EN can only be operated after receiving the previous PG signal. signal, which is an important timing requirement and cannot be skipped.
Design Thinking:
When designing in the R&D stage, how can it be convenient for hardware engineers to debug? Here comes the answer. In the final stage of the timing sequence, add an LED, which is usually green. When the LED lights up, it means that the previous timing sequence on the software is complete. After running, the power supply of each DCDC is normal (at least in theory).
This action is very friendly for the first debugging of the board. Considering the power consumption or actual demand when updating iteratively, it is not necessary to install the LED in the later stage, or just not light it up in the software.