Configuring the Brownout for when VDD3P3_RTC is fed 1.8v

[hassan789]

ESP team,
I am developing a 1.8v ESP32-D2WD board. The entire digital domain: VDD3P3_RTC, VDD3P3_CPU, and VDD_SDIO are powered from 1.8v
Looking into the menuconfig and the Technical Ref Manual 2.8, the brownout has a 3-bit (eight settings). However, it is centered around 2.5v. It is not clear on how to center this around 1.8v...

How can I trigger the brownout at ~1.7v?

Thanks

[ESP_Angus]

ESP team,
I am developing a 1.8v ESP32-D2WD board. The entire digital domain: VDD3P3_RTC, VDD3P3_CPU, and VDD_SDIO are powered from 1.8v
Looking into the menuconfig and the Technical Ref Manual 2.8, the brownout has a 3-bit (eight settings). However, it is centered around 2.5v. It is not clean on how to center this around 1.8v...

Hi Hassan,

This is not possible because 2.3V is the minimum voltage for the digital domain VDD supplies (except for VDD_SDIO). Consult ESP32 datasheet section 5 Electrical Characteristics, first row in "Absolute Maximum Ratings" table.

For this reason the minimum configurable brownout detector voltage is ~2.43V. Anything lower is too risky (the voltage may dip a small amount very temporarily without triggering the brownout detector, hence the 0.13V of headroom).

If a chip seems to run (with brownout detector disabled) at 1.8V anyhow, be advised that some features will probably not work and you may get data corruption, glitched instructions, and other weird problems.

Sorry to be the bearer of bad news.


Angus

[hassan789]

Hi Hassan,

This is not possible because 2.3V is the minimum voltage for the digital domain VDD supplies (except for VDD_SDIO). Consult ESP32 datasheet section 5 Electrical Characteristics, first row in "Absolute Maximum Ratings" table.

For this reason the minimum configurable brownout detector voltage is ~2.43V. Anything lower is too risky (the voltage may dip a small amount very temporarily without triggering the brownout detector, hence the 0.13V of headroom).

If a chip seems to run (with brownout detector disabled) at 1.8V anyhow, be advised that some features will probably not work and you may get data corruption, glitched instructions, and other weird problems.

Sorry to be the bearer of bad news.


Angus

Angus,
According to the datsheet, it says 1.8v is possible. Please see attached pic.
Also, I am using the chip already in 1.8v mode...

[ESP_Angus]

I'm sorry, I think this is a mistake in the datasheet. As per the footnote 1 to table 5.1, this range is only valid for the VDD_SDIO pin & associated power domain.

I've asked my colleagues to confirm, will get back to you.

[hassan789]

Here is another excerpt for the latest Hardware Design Guidelines v2.1

[ESP_Angus]

The hardware team has confirmed with me that they don't recommend running the other digital domain supplies below 2.3V. The lower voltages mentioned in the datasheet will be corrected to reflect this.

If you run these power domains at 1.8V then the ESP32 will probably appear to work but you may encounter very occasional memory corruption and also problems writing to efuse.

I'll let you know once the documents have been updated.

Sorry to be the bearer of bad news.

[hassan789]

wow.. thanks for clearing that up.
anyway to minimize the effects of this? perhaps running at 80MHz would help?

[ESP_Angus]

Hi hassan,

Thanks for your patience with this update.

A new ESP32 datasheet (V2.2) has been released which corrects the VDD3P3_RTC voltage range (2.3V to 3.6V, with a footnote - see section 5.2 in the datasheet). VDD3P3_CPU voltage range remains 1.8V to 3.6V. You can download the new datasheet from here.

Angus

[hassan789]

thanks! will review this in detail