How about the Qinheng RISC-V core CH32V103 microcontroller?

nmg

How about the Qinheng RISC-V core CH32V103 microcontroller? [Copy link]

 

The CH32V103 evaluation activity of Qinheng RISC-V core will end next Thursday (March 25). The 10 partners who participated in the evaluation shared that they have done a lot of evaluations ( >>Click here to view ).

What do the actual test engineers think of this microcontroller? Welcome to leave a comment. We will also feedback to Qinheng if you have any suggestions.

@yuzhang.zheng
@Single-piece Chinese cabbage
@eew_Violet
@jennyzhaojie
@wintonson
@ddllxxrr
@小天501
@MianQi
@hzz592788
@bigbat

单片大白菜

The effect is not as perfect as I imagined. Let me first talk about my situation:

1. I am not familiar with development tools and software development environment (IDE). If I have relevant Keil packages and fool-proof configuration, it will be much easier.

2. I am not familiar with the entire architecture. It would be much better to organize an offline seminar so that I can become more familiar with the architecture.

3. This chip is generally good, but the DAC is a bit useless and the ADC accuracy is not as high as expected.

ddllxxrr

I have tested the Qinheng RISC-V core microcontroller. I feel that it is not much different from the CORTEX-M3 core. I think this is a good opportunity to catch up with foreigners. Because, once foreigners refuse to supply due to war, weather and other factors, then our country also has a single chip with the same function. For me, I just need to modify the corresponding hardware to suit the pin allocation. And the price, if the price is lower and we promote it vigorously, maybe our products will be used.

MianQi

I didn't expect time to pass so quickly, and two months have passed without me noticing. First of all, I would like to thank eeworld for providing me with this review opportunity, which gave me the opportunity to test the RISC-V chip that I have been paying attention to. I would also like to thank the website staff "Liu Muer" for communicating with me on the progress of the review work.

Looking back on this review, my overall impression is:

Qinheng's kit (board + uploader) 1. Exquisite hardware. 2. Robust software tools. 3. Neat examples. 4. A good foundation. It would be a pity not to continue.

I discussed this with "Liu Muer" via WeChat a few days ago. The plan I proposed was too grand and could not be implemented. After some revisions, I now have a reduced plan:

I noticed that the examples provided by Qinheng are limited to brushed DC motors. The size and weight of the board being evaluated are suitable for use as a motor control board. I tested it in this post (https://en.eeworld.com/bbs/thread-1159389-1-1.html) and found that it can only accelerate and decelerate, but not start, stop or turn. This may be related to the lack of understanding of the motor driver board mentioned in the example.

In addition to the motors in the post above, 4 motors and a rotary encoder are also prepared:

The goal is to manually adjust the speed and commutation through a rotary encoder, in order to enrich Qinheng's existing examples and make a small contribution to domestic chips.

I hope to experiment with people who are interested. If you are interested, please leave a comment in the post above.

jennyzhaojie

I think it is easy to use, especially its USB read and write capabilities, which can make the development work no longer worry about the reading and writing of data, and it is highly efficient. Unless the performance requirements are too high, I will choose this chip first for general projects in the future. In addition, we should also support and choose domestic products!!!

bigbat

The overall impression of CH32V103 is OK. When testing, I was a little hasty in writing the program and some problems occurred. The calculation of the clock was a little different. I did not have a deep understanding of the entire system, which led to some ambiguity. I spent a whole day carefully reading most of the code in the routine again. I have some new understandings. I also made some summaries in the previous posts. This time I will talk about it again.
The instruction set of the chip is open source RISC-V, which everyone knows, but there are many types of RSIC-V instruction sets, just like ARM has M0, M3, M4, R7, A9... and so on. The instruction set of RISC-V is organized in a modular way, and each module is represented by an English letter. The most basic and only mandatory instruction set part of RISC-V is the basic integer instruction subset represented by the letter I. Using this integer instruction subset, a complete software compiler can be implemented. Other instruction subsets are optional modules, and representative modules include M/A/F/D/C, etc. I guess the instruction set of ch32v103 should be RV32E. The RISC-V instruction set was designed with "modularity" in mind, and various extensions and instruction set density were also considered. The system clock counter in CH32V103 is 64, which can support GHz frequency, while ARM mx is 24 bits. For a 72Mhz main frequency, 64 bits may not be necessary. The debugging interface of ch32v103 is also different. It is a 2-wire RVSWD interface, which is different from the JTAG used by most RSIC-V. I don't know why the system freezes when using debug. It doesn't feel very good here. But this does not mean that it must be a chip problem. It may also be a software problem. The interrupt management is different, it is PFIC (Programmable Fast Interrupt Controller), but it doesn't feel like a vector interrupt. As for any more advanced functions, I don't know much about it. The bus and peripherals of CH32V103 use the same bus as ARM cortex-M3, so everyone thinks that the programming is similar to the ARM system. I don't know why WCH puts the information of RISC-V and Cortex-M3 in the "CH32x Series Application Manual". You need to pay extra attention when reading it.
Some functions are not very meaningful, such as CRC32 calculation. This cannot change the check formula, nor can it calculate the "16-bit experience and". Its application is very narrow, and I hope it can be strengthened.
As for how to choose a chip, everyone has their own opinions. You need to consider it according to your own situation. The first thing to consider for a chip is performance, and then price. Choose a product with good cost performance according to your needs. In actual production, it is usually not easy to replace a verified chip. Once a problem occurs, there may be heavy losses. If there is no good cost performance, it is generally not easy to replace the solution.

bigbat

In addition, the routines of wch are somewhat incomplete. For example, the USB program only does a bootstrap, and the subsequent read/write data and error interrupt are not done. However, the USB function is still good. The functions of SPI and I2C are not fully tested, and they feel so-so. The main reason is that there is no description of the clock in the routine. I hope the information can be improved. Because it is a new product, there is not much information on the Internet, so the manufacturer tries to gather us grassroots personnel as much as possible. Let's work together to popularize RSIC-V microcontrollers as soon as possible.

eew_Violet

Qinheng's IDE is based on eclipse, so it's easy to get started. Some other manufacturers are also based on this. In the process of using CH32C103, it is basically the same as ST's, which reduces the difficulty of getting started. What's rare is that it can be directly powered by 5V, saving LDO and PCB area, which is very good. Compared with the registers, it is the same as 48 pins and below, which can basically be a perfect replacement. It's just that the PFIC developed by itself does not have too much introduction and routine reference. I think the manufacturer should issue a more detailed PFIC introduction and NVIC transplantation reference. In addition, I still feel that the burning time is long during use. In general, it is still a good MCU. Especially with the sister CH32f103, Qinheng has made it compatible and can be universal. You can choose arm or risc-v without changing the design.

hzz592788

The IDE is OK, I don't think there are any big problems. But it would be nice if I could use KEIL. Overall it's good. There are project requirements in the middle, and I hope to have the contact information of the agent. Maybe it's not over yet during the Spring Festival, so there is no response. Later, there was no news from the customers who had needs.

Everything is based on price: cheap means good, expensive means bad.