[Rule 1] Design the most compact system that meets the requirements.
It is very important for a microcontroller system designer to correctly estimate the capability of the single-chip microcomputer, know what the single-chip microcomputer can do, and maximize the potential of the single-chip microcomputer. We should have such an understanding that the processing power of the microcontroller is very powerful. Early PCs, their CPU (8086) processing power and 8051 equivalent, but can handle fairly complex tasks. The key to the ability of the microcontroller lies in the software written by the software designer. Only by fully understanding the capabilities of the single-chip microcomputer, will it not make a "redundant" system design. However, many peripheral chips are used to realize the functions that the single-chip microcomputer can realize. Doing so increases the system cost and may also reduce the reliability of the system.
[Rule 2] Use a watchdog.
A watchdog circuit is usually a piece of hardware that is updated at regular intervals. The update is generally completed by the microcontroller. If the watchdog cannot be updated within a certain interval, the watchdog will generate a reset signal to reset the microcontroller. The specific form of updating the watchdog is to provide a rising edge of the level to the relevant pins of the watchdog chip or to read and write one of its registers. Using the watchdog circuit will reset the microcontroller when the microcontroller crashes and crashes. There are many kinds of watchdog chips at present, such as MAX802 and MAX813 of MAXIM Company. Moreover, there are many kinds of microcontrollers that have integrated watchdogs themselves. An external watchdog is best because it does not depend on the microcontroller. If possible, the watchdog update routine should not be placed in interrupts or subroutines, but in principle should be placed in the main program. I once saw an engineer who was debugging a program that occasionally caused the watchdog to reset when it was running, so he simply cleared the watchdog in a clock interrupt program that interrupted every 10ms. I'm sure he also knew about disabling the watchdog, but he didn't try to find the real cause of this phenomenon. Therefore, I would like to remind everyone: For whatever reason, never ignore the real cause of system failure. High-quality products come from high-quality engineers, and high-quality products create high-quality engineers.
[Rule 3] Make sure that the reset signal of the system is reliable.
This is an easy problem to ignore. Did you have this concept in mind when you were designing a microcontroller system? What kind of reset signal is reliable? Have you used an oscilloscope to view the reset signal of the product you designed? What are the possible consequences of an unstable reset signal? Have you ever found that the single-chip microcomputer system you designed, after each power-on, the data becomes messy, and the phenomenon is not the same every time. A dead state, and sometimes nothing at all works properly? In this case, you should check your system's reset signal. Generally, the requirements of the reset signal required by the microcontroller are mentioned in the data sheet of the microcontroller. Generally the width of the reset signal should be . The width and amplitude of the reset level should meet the requirements of the chip and should be kept stable. There is also a particularly important point that the reset level should occur at the same time as the power supply, that is, as soon as the chip is powered on, the reset signal has been generated. Otherwise, since there is no reset, the value of the register in the single-chip microcomputer is a random value, and the program will start to run according to the random content in the PC register when it is powered on, so it is easy to misuse or enter a crash state.
[Rule 4] Make sure that the initialization of the system is valid.
The start of the system program should be delayed for a while. This is a common method in many microcontroller programming, why? Because the chips and devices in the system often have a period of time from the power-on to the normal working state, the delay at the beginning of the program is to let all the devices in the system reach the normal working state. How much delay is appropriate? It depends on the time it takes for each chip in the system to reach normal operation, usually whichever is the slowest. Generally speaking, a delay of 20-100 milliseconds is sufficient. For "slow heat" type devices such as embedded MODEM in the system, it should be longer. Of course, this all needs to be adjusted in the actual operation of the system.
[Rule 5] Check the system when it is powered on.
It is a good design in the MCU program to detect the system when it is powered on. When designing the hardware, careful consideration should also be given to designing each chip and interface used in a mode that is easy to use software for testing. Many experienced microcontroller designers will conduct a comprehensive test when the system is powered on (especially the first time it is powered on), or go a step further, divide the operating state of the system into a test mode and a normal operating mode. The mode carries out detailed inspection of the system, which makes the batch inspection of the system more convenient and easy. In addition, it should be noted that a simple and clear fault display interface is also quite laborious. For example: the external RAM (data memory) of the system is a commonly used device in the microcontroller system. external RAM
If there is a problem, the program often becomes a wild horse. Therefore, the program must check the external RAM when it is started (at least when it is powered on for the first time). The detection contents include: 1) Detecting units in RAM. This is mainly by keeping the data written and read consistent. 2) Detect the address data bus between the microcontroller and the RAM. The buses are neither shorted to each other nor connected to "ground". In addition, many chips provide testing methods. For example, the serial communication chip UART has the function of loop test.
ã€Rule 6】Design hardware according to EMC test requirements.
EMC testing requirements have become mandatory for products. There are many articles about this.
[Rule 7] Compile in Small mode whenever possible
Compared with Large mode and Compact mode, Small mode can generate more compact code. In Small mode, the C51 compiler will not use keywords, such as idata, pdata, xdata special declaration variables are all placed in the data unit. In programming, for the data area, it can be specified to be placed in the external memory.
A docking station is simply an adapter with multiple interfaces, but the word "dock" is a bit confusing. It is a convenient one-stop connection between the notebook computer and multiple accessories or external devices by duplicating or even expanding the notebook computer's ports. Laptops of the same brand, usually a generic port replicator.
type c hub
Shenzhen Kingwire Electronics Co., Ltd. , https://www.kingwires.com