ESP32-C3 power consumption doesn't match to the datasheet

[honestech]

Hi all,

I am checking the power consumption of the ESP32-C3-Devkit-M1. The section 4.4 of the datasheet (https://www.espressif.com/sites/default ... eet_en.pdf) says that it's 130uA when the light sleep mode and 5uA when the deep sleep mode. It's very interesting for me so that I think it could be used for low-power IoT applications.

So I have measured the current consumption using the Nordic Power Profiler Kit II and used two examples(examples/system/light_sleep and examples/system/deep_sleep) in esp-idf.

The result is like:
- light sleep: 1.1mA
- deep sleep: 750uA

And the README of the light_sleep example says me I am getting right result.
I am wondering if the datasheet is right or there is something that consumes the current out of the SoC like CP2102.
If it consumes about 1mA in the light sleep mode, it might not be used for the low power application.
Does anyone explain to me whether there is something to draw the current out of the module on the board or how can I get about a hundred of uA in the light sleep mode?

Regards.

[ESP_Sprite]

Are you measuring the 5V power usage? Asking because the Devkit-M1 uses a LM1117 LDO which has a pretty high quiescent current.

[ESP_igrr]

Even in case of powering the devkit through the 3.3V pin, as you correctly say CP2102 will draw additional current. To measure the current consumption of ESP32-C3 chip, you will probably need to connect a bare module to the current meter. I think the development board doesn't provide a way of doing this.

[stublen]

Hi！
I am using a bare esp2-wroom-32ue module to measure the current consumption, i used the deep_sleep example.
I only use the rtc and ext1 to wake up the esp32, but the measured current is:80-90uA which does not match the datasheet.
I want to know what i have missed.
regards.

[zwer97]

I also have this issue with ESP32C3 on custom board (not dev board). Current consumption is about 300uA instead of 5uA as in datasheet.
Also in esp-idf doccumentation https://docs.espressif.com/projects/esp ... modes.html they recommend to use

Code: Select all

rtc_gpio_isolate(GPIO_NUM_12);

to reduce power consumption, but this function is not available for esp32c3.
So how to achieve 5uA current consumption with this chip?

[zwer97]

Try this before deep sleep, it seems to help me.

Code: Select all

for (gpio_num_t gpio_num = GPIO_NUM_0; gpio_num < GPIO_NUM_MAX; gpio_num++) {
	if (GPIO_IS_VALID_GPIO(gpio_num)) {
		if (gpio_num <= GPIO_NUM_5) {
			REG_CLR_BIT(RTC_CNTL_PAD_HOLD_REG, BIT(gpio_num));
		} else {
			CLEAR_PERI_REG_MASK(RTC_CNTL_DIG_PAD_HOLD_REG, GPIO_HOLD_MASK[gpio_num]);
		}
	}
}

[stublen]

Have you tried esp_deep_sleep() and esp_deep_sleep_start().
When i call the first function, the current consumption is 10~20uA while the latter is 80~90uA.
So weird!

[axellin]

Have you tried esp_deep_sleep() and esp_deep_sleep_start().
When i call the first function, the current consumption is 10~20uA while the latter is 80~90uA.

esp_deep_sleep() simply calls esp_sleep_enable_timer_wakeup() + esp_deep_sleep_start().
So it looks lke what you observed is: https://github.com/espressif/esp-idf/issues/8240

[laurent.coldfusion]

Is there any update on this topic ? I am facing the same issue, using ESP32-C3-DevKit02, power supply directly on the 3.3v (to avoid LDO Regulator consumption and any other components connected to 5v). Result is a consumption of ~500 µA in deep sleep, significantly above the 5µA of ESP32C3WROOM02 datasheet.

[malachib]

I too am experiencing this. Measured right around 1ma. It seems though that the little red LED (during 5V pin usage) and the CP2102 are powered all the time, the combination of which plausibly could be 1ma. Because we're talking specifically about the DevKitM-1 right? That's what the schematic for the board indicates. Current measured at ~4V feeding LDO.

As far as light sleep goes for you OP, 1ma as compared to 750ua seems about right according to https://www.espressif.com/sites/default ... eet_en.pdf Section 4.6.2 Table 19