The domestic substitution of components was initially a positive development, but recently, counterfeits have started to appear in the domestic market.
A user named imdx posted that they purchased a “brand new original” GD32F103C8T6 from a certain online marketplace. The pins looked quite neat, and the silkscreen was very clear, but there were some differences in the marking compared to previous versions. Actual testing with debugging tools confirmed that it was indeed a GD32F103C8T6, and it could be programmed and operated normally.However, some chips (approximately 10%-40%) exhibited low resistance to ground when powered with a 3.3V supply, resulting in high standby currents ranging from 10mA to 300mA.Imdx noted that for a normal chip, the input current at 3.3V before programming is very low, below 2mA, so such a high current must indicate that the chip has been refurbished. Upon checking the chip UID, the UIDs from the same batch were quite chaotic, whereas chips from other suppliers had very similar UIDs within the same batch.Ultimately, this batch of counterfeits was returned, and a new supplier was chosen. They received 10 pieces of new stock from week 44 of 2020, and the UIDs of the 10 pieces only varied in the first 16 digits. It can be confirmed that these are new products.Although the price is slightly higher than the previous counterfeits, counterfeits cannot change their UIDs no matter how they are refurbished. Judging whether chips are refurbished or not based on UIDs from the same batch is relatively reliable.Currently, the new GD chips have the GigaDevice logo in the marking. On this note, he reminds everyone to be cautious of counterfeits.To determine whether a product is counterfeit, the definition is whether the physical item matches what the seller claims. If it is sold as “loose new,” then it cannot be considered counterfeit; “loose new” refers to refurbished items, which can still be genuine. If it is sold as “brand new original,” then it is counterfeit.Another user also reported falling victim, stating that they bought samples from two different suppliers, with one being fine while the other had CPUIDs that were all FF. The default internal oscillator frequencies varied significantly, which explained why they could not download using ISP; they could only use SWD to download and found that the internal RC oscillation frequency was actually 8.69M, which was corrected for normal use.After checking on the online marketplace, the reporter indeed found many similar sellers labeling their products as “brand new original.”
Searching for information on the online marketplace does not imply that the seller is engaging in fraudulent behavior, nor does it represent the stance of this publication; please choose wisely.LZSXlijuqing suggested that the defect rate is about 10%-40%, which is clearly indicative of refurbished goods. The pins look quite neat, and the silkscreen is very clear, but the pins have been cleaned and the silkscreen has been reworked, creating a very attractive appearance. The IC chip market is quite complex, and purchasing from the online marketplace is not recommended; it is better to find a long-term supplier of electronic components and start with small batches for long-term cooperation.Another user expressed concern, stating, “Even domestic products are starting to have refurbished goods; this is terrifying. When I first started with this, I bought STM32 chips from the online marketplace, and the yield was quite poor. I didn’t know whether the issue was with my circuit design or something else. Later, I switched to buying brand new products from authorized distributors, and all problems were resolved.”Clockwin commented that when individuals are experimenting with small quantities, they often do not know which suppliers are legitimate. If it is a distributor recommended by the manufacturer, they might not pay attention to you if your company or order volume is too small.Xiongguoling noted that a defect rate of 0.5% is still a bit high; previously, when using ST products, they had about 5K units per month with very few customer complaints. This time, after switching to a domestic brand with 10K units over two months, they received many customer complaints and are quite worried, especially since ST’s prices have recently become quite appealing.Some believe that the emergence of counterfeits indicates that GD is performing well. They generally deal directly with reliable distributors, ensuring no defective products, and are glad they did not go through the online marketplace.Others believe that the cost of producing counterfeit MCUs is too high, so they are mostly refurbished products.Those who refurbish GD32 chips are often insiders; some sellers are more reputable and will not sell refurbished items. If you buy them for debugging a prototype, that is fine, but do not use refurbished items for formal shipments.Imdx responded, stating that refurbished goods have their value. It is fine to use them during the development phase. However, as a seller, refurbished goods are refurbished goods, and brand new goods are brand new goods. Selling refurbished goods as brand new is indeed selling counterfeits. For products like MCUs, what you think are counterfeits cannot be produced by unscrupulous merchants.
In addition to the rampant counterfeiting, the increasingly rapid replacement cycle has also exposed some issues, and this problem happens to be very similar to the aforementioned counterfeit products, leading to a mix-up.The user dujunqiu discovered that the GD32F103RET6 had excessively high standby power consumption during use. They mentioned that to save costs, the company decided to switch from STM32 to GD32. During testing, they found the following issues:1: When testing the sleep power consumption, they found that the standby current was too high, reaching 15mA; compared to STM32, the standby power consumption was only 1mA;2: Using J-link to erase the GD32 program, the power consumption was 20mA after powering on; compared to STM32, the power consumption after erasing the chip program was only 5mA; the standby program should be fine, as they compared it with STM32 and other users’ shared cases and found no issues..
/**************System enters standby mode**********/void Sys_Enter_Standby(void){ RCC_APB2PeriphResetCmd(0X01FC,DISABLE); //Reset all IO ports RCC_APB1PeriphClockCmd(RCC_APB1Periph_PWR, ENABLE); //Enable PWR peripheral clock PWR_ClearFlag(PWR_FLAG_WU); //Clear wake-up flag 20180523 PWR_WakeUpPinCmd(ENABLE); //Enable wake-up pin function PWR_EnterSTANDBYMode(); //Enter standby (STANDBY) mode}void PWR_EnterSTANDBYMode(void){ /* Set SLEEPDEEP bit of Cortex System Control Register */ SCB->SCR |= SCB_SCR_SLEEPDEEP; /* Select STANDBY mode */ PWR->CR |= PWR_CR_PDDS; /* Clear Wake-up flag */ PWR->CR |= PWR_CR_CWUF;/* This option is used to ensure that store operations are completed */#if defined ( __CC_ARM ) __force_stores();#endif /* Request Wait For Interrupt */ __WFI();}After reading the UIDs of three chips, they found them to be quite normal [GD32] UID: 30451B67-5343932-35323647 –3[GD32] UID: 3045837A-4343932-35323647 –2[GD32] UID: 30458378-4343932-35323647 –1Some speculated that perhaps the original poster also encountered counterfeit issues. Others pointed out that some IO ports of the GD103 series MCUs have leakage issues and suggested checking them one by one.Ultimately, the original poster discovered that the actual GD32 datasheet was incomplete, and from the STM32 datasheet, it was noted that for the IO ports that need to be controlled, pull-up or pull-down resistors need to be added in hardware when entering standby mode.
Although it was just a mix-up, there are still some issues with domestic products that need improvement, such as the datasheets.What are your thoughts on the counterfeit issues arising during the domestic substitution phase? Do you have any tips to avoid scams? What other problems do you think domestic MCUs have? Feel free to leave comments for discussion.
Electric Competition Column
Familiarize yourself with these for the 2021 Electric Competition! 
Some thoughts on the 2021 Electric Competition, seeing is earning!
Detailed technical solutions for the 2020 Electric Competition power supply question, bookmark it now!
2020 Electric Competition questions, authoritative analysis by topic experts!
How to prepare for the Electric Competition? Summary of experiences from the 2019 Electric Competition!
Electric Competition | 2019 National First Prize, a memoir from a student of Beihang University (Part 1)
Electric Competition | 2019 National First Prize, a memoir from a student of Beihang University (Part 2)
“Electric Competition Sharing” Power Supply Question, Provincial First Prize!
How to conduct a comprehensive evaluation for the 2019 Electric Competition? A senior from National University shares the ultimate evaluation!
Electric Competition | How to prepare software for the power supply question?
“2020 Electric Competition” Detailed Technical Solutions for Power Supply Questions, bookmark it now!
-END-
The annual Electric Competition is here, so I will be writing some articles about it in the near future. Stay tuned!