The chips I have used before include: AT89C51, 2051, PIC16C711, 16F84, LPC76X, ADuC812/834
However, after seeing AVR, I think AVR has many advantages:
1. It has a 10-bit A/D converter, which generally meets the basic requirements for industrial control;
2. It comes with FLASH and EEPROM, especially the EEPROM, which can store logic bombs to prevent certain people from using it without payment;
3. It has fewer pins, no external bus, and decent anti-interference performance;
4. Development is easy, and the hardware cost is low (I am currently using assembly language, not C language, as C is too slow for high-speed real-time control). Using assembly language, writing a relatively complex control program (like a high-power welding device SCR regulator controller), including key display, PID, triggering, and protection, 4K * 16 bits is sufficient. Even for very complex industrial control systems, 16K * 8 bits of ATmega16 is more than enough; the key is whether your programming skills are sufficient;
5. The encryption performance is relatively good;
6. It is inexpensive; for high-power industrial control devices (at least over 10,000 yuan), that cost is nothing;
7. I originally wanted to use uPD7810 or 68HC11A8, but the development system’s price was just too high, and you can’t use it in just a few days.
So, I am still using 51, mainly with software simulation and dedicated simulators. I use software simulation for AVR because the JTAG interface consumes too many resources, and modifying programs is not as quick and convenient as ISP.
AVR is the chip I want right now; it has all the basic functions, no external bus, and the possibility of being cracked by others is far less than that of 51. I really can’t stand the paging of PIC; it is just too annoying. The operating temperature range of AVR can also well meet my industrial control requirements, so while I continue with 51, I am also learning AVR.
For true industrial control, DSP is essentially useless except for space vector transformation and complex signal resolution; it wastes money and national resources 🙂 and is not easy to process. If the volume is small, it is not worth it, so I do not use 320LF3407/2812, ARM, etc.
To determine whether a chip is useful, I believe the following points are key:
1. Can it meet market demands for your product;
2. The cost is relatively low;
3. Development costs are low; including hardware and software costs;
4. PCB design is easy;
5. Excellent encryption performance;
6. There is some room for upgrades;
7. Can store logic bombs (because, in China, the one who owes money is the boss, and the one who collects debts is the subordinate);
8. High pin driving capability, allowing for fewer external components;
9. Development languages can easily incorporate software anti-interference and occupy less code resources;
10. Wide operating temperature range and strong power supply adaptability.
I am not developing Mars rovers or the “Shenzhou” spacecraft; I only deal with some simple devices in military electronics (not phased array radar fire control systems or sonar processing equipment), so there is no need for fancy things like VXworks or uC-OSII. For a good assembly language system, program portability is not difficult as long as you modularize the program and carefully allocate resources. However, I have not yet seen anyone capable of directly porting C51 to AVR; portability often refers to compatibility within the same series of chips, not across different MCUs. Perhaps it will be unified in the future, but at least it is not unified now, so just bear with it, haha. Moreover, for a company, for example, if you originally made mobile phone chargers and now suddenly have to make a 250KW high-power welding device, it seems impossible to get started without three to five years of experience, let alone make a profit. If a company or an employee tries to do everything, they certainly can’t do anything best, just like Haier appliances in China.
Therefore, I still use assembly language, still do not use C language, and continue to control high-power electronic devices, although the development speed is somewhat slower (compared to C language), but at least it won’t lead to users returning the products I send, which is the law of survival in the market, not the survival law of academic fraudsters.