SMPS on STM32N657

[matt-crc]

Hello,

Regarding the STM32N657, I'm not using the internal SMPS.  My question is:  Do we need to put capacitors on the 6 VDDSMPS pins or can they be left as no connect?   

Same question on the VLXSMPS pins... do we need the capacitors on these pins, or can they be left as no connect?

Also, on the STM32N657-DK board, there is a snubber circuit on the VLXSMPS pins, why?

Finally, when I say "No connect", I assume these pins are connected internally, so it does not require an external link on the PCB when not used.  Is this correct?  (this will probably make the routing of the PCB easier)

We are dealing with a very Very VERY small PCB, and saving the space used by 12 x 0402 makes a difference.

thanks

[TDK]

VDDSMPS must be grounded when SMPS is not used. Capacitors are not needed.

*Getting started with hardware development for STM32N6 MCUs - Application note

[matt-crc]

hi @TDK 

I'm not sure if you are responding to this post or: STM32N657-DK board bug or NC?

on the STM dev board, this pin is floating.

[TDK]

> I'm not sure if you are responding to this post or: STM32N657-DK board bug or NC?

This post.

My reply was made before you edited your post to add the VLXSMPS questions.

> on the STM dev board, this pin is floating.

It looks connected to me.

(So does VLXSMPS.)

[matt-crc]

hi @TDK 

1- thanks for the appnote - didn't see that one before.

2- No on the STM32N6570-DK board, the VCCA18PMU pin seems to be floating:

VDDA1V8PMU is connected to VDD1V8 thru and inductor L8

and also connected to VDDA1V8 thru R173 which is not installed

VDD1V8 is generated by external switcher, but R21 is a no install

and since the VDDA18PMU is floating, the VDD1V8 is all floating from the STM32

SMPS stuff table for the STM32N6570-DK board....

Very confusing design here....

[matt-crc]

AN5967,  RM6486, UM3300, and the STM32N657x0-dk development board schematics seem to contradict each other.... which one is correct regarding the SMPS configuration?

[RomainR.]

Hi @matt-crc 

The STM32N657-DK board was designated very early compared to AN5967 Rev3, so there are some differences regarding the internal SMPS and bypass (figure 4 of AN5967).

However, you must consider the recommendations of AN5967 as indicated in DS14791 Rev7 in section 3.4.4 SMPS usage.That is, connect VDDA18PMU and VDDSMS + VFBSMPS to GND when the design is SMPS Bypass.
The RM0486, UM3300 are not intended to explain how to implement the internal SMPS or Bypass; that is the role of the AN5967. 

Apart from this point, let me know if you see other contradictions between the different documents?

Best regards,

Romain,

[matt-crc]

Hi @RomainR. 

The other confusing issue was the VBAT connection.  It said that the Vdda1v8_AON is powering the crystal pins... So, does vdda1v8_aon need to be on when running from VBAT=3V? Or will it use the internal oscillator when on VBAT?

if it uses the internal oscillator while on VBAT, will that affect RTC accuracy?

thanks again, Matt

[RomainR.]

Hi @matt-crc 

Only the LSE (external crystal) is in VBAT domain, refer to RM0486 in Clock distribution for RTC/AWU clock.
This is the case for all other STM32.
If RTC use LSE as clock source, the accuracy depends on external 32768Hz crystal. We have the AN2867 which explain that. If the VBAT voltage recommendations are followed, there will be no impact on RTC accuracy.

VDDA18AON is the supply domain for HSE, HSI, MSI (and OTP). Figure 15. Power supply overview also in RM0486. No relation with LSE.

Best regards,

Romain,