If you use VREFBUF generated from the MCU, use buffer capacitors on the VREF+ pin. Always: it's a power supply. (That hint is in the hardware design guide, but it uses two sentences.)

[waclawek.jan]

From your other posts I see you use the 'U5xx, so I took a look at the 'U575 DS.

In Functional overview/VREFBUF chapter, there's a schematics with capacitor with no marking, so that may be the first clue.

In Electrical characteristics/VREFBUF chapter/table the capacitor is specified, and in fact in a rather narrow range, min.0.5uF typ.1uF max.1.5uF, with a footnote requiring an additional 100nF HF bypass. Then later in the table, the startup time is given as function of the actual cap used.

But I agree that this all is fineprint too easy to miss when designing a board, being overwhelmed by all the many detailed requirements.

Thanks for this post.

JW

[EDIT] I've had a look also at the 'L0/'L1 DS. These two families are relative outliers in several respects from other STM32, and in this particular respect, they have a buffered VREFINT (the buffered version called VREF_OUT), connectable to two GPIO pins, for whatever purpose. Granted, it has only 1uA current capability, as compared to VREFBUF's 4mA, and no direct connection to VREF+. Nonetheless, the DS for this VREF_OUT calls for *max* 50pF output load. If you'd had experience with this feature in the past, it's just natural then to tend to expect the similar with VREFBUF.

@Peter BENSCH​ , see here potential for one of those dozens of appnotes per peripheral I'm talking about all the time?

[Peter BENSCH]

Thank you @Community member​ for bringing this to our attention. There is certainly room for improvement in the appnotes.

@Imen DAHMEN​ 

@Amel NASRI​ 

[Mohamed Aymen HZAMI]

Hello @Community member​,

Thank you for your feedback.

As I understand the request is to improve the following topics:

• Update the figure VREFBUF block diagram in the U5 datasheet, add a reference to the capacitor value (referring to the VREFBUF characteristics table).
• Change the VREF_OUT naming to VREFBUF_OUT in several STM32MCUs datasheets.

Could you please clarify this?

With Regards,

Mohamed Aymen

[waclawek.jan]

Hi Mohamed Aymen,

Yes, you're absolutely right, in both points.

Note, that there are STM32 families ('L0, 'L1) which do have a genuine VREF_OUT signal, so don't change those.

What exactly would you need to clarify further?

Thanks,

Jan

 @Mohamed Aymen HZAMI​ 

[Mohamed Aymen HZAMI]

Hi @Community member​,

What do you mean by:

>Note, that there are STM32 families ('L0, 'L1) which do have a genuine VREF_OUT signal, so don't change those.

Mohamed Aymen

[waclawek.jan]

Hi Mohamed Aymen,

In the 'L0 and 'L1 families, which are older, and don't have the VREFBUF module, the internal VREFINT module - which is present in all STM32 but in the vast majority its entirely internal and has no outside connection - but in 'L0 and 'L1 VREFINT can optionally be output to certain pins. This is marked as VREF_OUT and this should not be changed.

The VREFBUF module is relatively new, first appeared in the 'L4 family.

It's output is connected directly to VREF+ pin, but if it needs to be indicated in the datasheet, it ought to be called VREFBUF_OUT and not VREF_OUT. So VREF_OUT is incorrect only in the STM32 models which are newer than 'L4 and do have VREFBUF (as in 'G071:

https://community.st.com/s/question/0D53W00001pJhC3SAK/st-in-models-with-vrefbuf-dont-call-its-output-vrefout

 )

JW

@Mohamed Aymen HZAMI​ 

[Mohamed Aymen HZAMI]

Hello @Community member​,

Thanks for the clarification.

Mohamed Aymen