Page 1 of 1

Higher current consumption (100uA) in Deep Sleep on ESP32-WROOM-32D

Posted: Tue May 14, 2019 11:21 am
by marekg
Dear all,
I am struggling with excessive current consumption on ESP32-WROOM-32D in deep sleep mode.
According datasheet it should be 10uA but it consumes 100uA instead.

I have used functions rtc_gpio_isolate to isolate GPIO pins but I could not lower current.
I measure current directly at module power supply input.

Do you have same experience, please?
Did you reach current 10uA in deep sleep on ESP32-WROOM-32D.

Thank you very much for your help.

Re: Higher current consumption (100uA) in Deep Sleep on ESP32-WROOM-32D

Posted: Tue Dec 24, 2019 10:03 am
by borelg
Hi marekg,

I am stuck with the same question here.
I am actually consuming around 140uA when in Deep Sleep.
However, I set up the wake-up to be triggered by RTC_GPIO.

Any suggestion about how to lower the power consumption?

Thanks :)

Re: Higher current consumption (100uA) in Deep Sleep on ESP32-WROOM-32D

Posted: Tue Dec 24, 2019 10:40 pm
by SO_yeah
I also see this problem.
The only board I have which doesn't present this problem is the "ESP32 SparkFun Thing".
The only meaningful difference is that the sparkfun has a 26MHz quartz the others 40MHz.
I know that many quartz draw a lot of current but could anyone post the dataheets for the SparkFun and the WROOM quartz to verify?

Re: Higher current consumption (100uA) in Deep Sleep on ESP32-WROOM-32D

Posted: Wed Dec 25, 2019 1:08 am
by ESP_Sprite
Not sure if that's the difference at play here, as the main (40MHz/26MHz) clock oscillator is shut down in deep sleep mode.

Re: Higher current consumption (100uA) in Deep Sleep on ESP32-WROOM-32D

Posted: Wed Dec 25, 2019 10:56 am
by SO_yeah
In light of your observation I headed to the second possible culprit and I think I solved it.

the sparkfun uses the ftdi serial adapter which runs on the 5v of the usb, while all of my other boards use the cp210x which run on the 3.3v from the LDO (these serial adapters when active draw around 100uA).

so when you use a battery on a board with the ftdi type programmer this will not be powered, unlike on the boards with the cp210x type.
Then I cut the 3.3v trace to power the cp210x on a board with ESP32-WROOM-32 mounted. Now that the programmer is inactive power consumption is comparable to that of the sparkfun board.
I think the same reasoning should apply to esp8266 boards.

hope this is a solution for the OP because it wasn't specified if a serial adapter was present in the setup.

Re: Higher current consumption (100uA) in Deep Sleep on ESP32-WROOM-32D

Posted: Fri Jan 03, 2020 10:35 am
by borelg
Thank you very much for your replies!

I believe that So_yeah highlighted the problem here.
Moreover, I was doing a mistake while measuring the current consumption in the first place.
It is not 100uA, but it is in the order of 2mA.
I am supplying directly 3.3V, but I noticed from the schematic of the board (a Lolin32) that the USB-Serial adapter (CP2104) is directly connected to the 3.3 line and not to the VUSB.

Thus, I will now isolate the ESP32 module and perform again the measurements.
I will then update the post.

Thanks again!!