STM32G474RET6 device marking and function issue

[WLai.1]

Dear Sir,

I have questions about STM32G474RET6. I have 25 MCUs in my hand. After SMD and power on, I found some abnormal conditions. Then, I swap the MCU (NUCLEO-G474RE and my PCB) and use the same firmware, the MCU can work normally on my PCB.

1. I took two MCU pictures, one was used in NUCLEO-G474RE, and the other was purchased from MOUSER. You can find some different information in the product identification. At NUCLEO-G474RE, the product identification show "STM32G474RET6U" and revision code is "Y". At my MCU (from MOUSER), the product identification show "STM32G474RET6" and revision code is "Z". Can you explain what different at there two MCU ?
2. Continue to question 1. If these two are the same, why my MCU cannot work on my PCB but NUCLEO's MCU can work on my PCB ? If these two are different, how to confirm that the MCU I bought is correct and working?
3. ​At NUCLEO-G474RE, there have use 24 MHz crystal (External HSE CLK), after testing (removing SB25 and SB26), it is not necessary. What is its function here? If use it, have any register setting need add in firmware ?

Thank you.

[TDK]

> Can you explain what different at there two MCU ?

Revisions are typically made to correct hardware issues. You can find out the known issues by reading the errata sheet, which notes which ones have been fixed in which revision.

https://www.st.com/resource/en/errata_sheet/dm00500968-stm32g471xx473xx474xx483xx484xx-device-errata-stmicroelectronics.pdf

> If these two are different, how to confirm that the MCU I bought is correct and working?

No silver bullet here. Could be a hardware issue. Look like there could be a solder bridge on the top-left on the bottom picture. Inspect power rails, step through code to find issues.

> At NUCLEO-G474RE, there have use 24 MHz crystal (External HSE CLK), after testing (removing SB25 and SB26), it is not necessary. What is its function here?

To use the HSE clock, you need to enable it and select it as a clock source. CubeMX can do this for you within the SystemClock_Config function. The benefit of HSE is higher precision and (pretty sure) lower power compared to HSI.

[Uwe Bonnes]

The Y revision has an incredible amount of _2_ issues fixed w.r.t. Rev Z. Both are ADC issues.

[waclawek.jan]

The U suffix is used to mark, that licence fees have been paid for software which accompanies the related Nucleo board. That has no relationship to functionality at all.

JW

[Piranha]

> The benefit of HSE is higher precision and (pretty sure) lower power compared to HSI.

Datasheet shows several times higher current consumption for HSE compared to both of HSI.

[WLai.1]

Hi TDK,

Thank you for your reply.

But there are some errors in your reply.

"It seems that there may be a solder bridge in the upper left corner of the bottom picture." -> The bottom picture works, but the top picture does not work.

I double-checked the "solder bridge" in the bottom picture, there is no "solder bridge" on the actual PCB.

[TDK]

> But there are some errors in your reply.

I didn't say there definitely is, just that there may be, based on the low-quality pictures you provided. My actual quote was:

> > Look like there could be a solder bridge on the top-left on the bottom picture.

 

Also, HSI is lower power, as pointed out by Piranha. That one was an error.

[WLai.1]

Hi TDK,

Thank you for paying attention to this point. We have confirmed this part and there is no problem.

Based on the information currently obtained, the MCU and EVM purchased by MOUSER are actually only used to correct internal problems, not a major revision. I still can't understand why after transplanting the MCU on the EVM to the PCB I designed, the PCB and the designed functions can work normally, but the MCU I bought from MOUSER cannot work.

P.S.: EVM = NUCLEO-G474RE.

[Igor Cesko]

Hi WLai.1 ,

Both devices differs only in limitations described in erratasheet.

What is exactly the problem which is present in the device with Revision Z ?

Regards

Igor