Update on 17th November 2021: having now played with an ESP32-S2 based board, I should say that it draws a lot less power than the combo described here.

Power Measurements

I was thinking about using a Raspberry Pi with an ESP32 based Wireless Pack for a battery-powered monitor, so I was interested to see how much current it draws.

All the tests below use Adafruit’s Circuit Python. Normal operation is what the name suggests, and includes time sleeping with the time.sleep() call.

Light sleep uses the alarm.light_sleep_until_alarms() call; deep sleep uses the alarm.exit_and_deep_sleep_until_alarms() call.

Pico only

The first tests use only the Pico board, powered at 5V through the USB port.

StateCurrent draw / mAPower consumption / mW
Normal operation1890
Light sleep1260
Deep sleep525

Accuracy for the current measurement is about ±0.5mA. There’s a bit of high-frequency noise, presumably from the DC-DC converter, but the very little variation otherwise.

Pico and Wireless Pack

The second set of tests use the Pico board with a Pimoroni Wireless Pack1, all powered at 5V through the USB port.

StateCurrent draw / mAPower consumption / mW
WiFi not in use36180
WiFi in use91455
Light sleep29145
Deep sleep22110

As can be seen just plugging in the Wireless pack increases the current consumption by about 17mA. Thus if the Pico is running normal code the current consumption is roughly doubled if it’s sleeping it more than quadruples. It is rather a shame that the sleep current is so high.

If you actually use the WiFi consumption rises sharply: nearly three times higher. I have no idea if the current will change in different settings: in my test the ESP32 was about a metre away from the WiFi Access Point, so presumably very little RF power was needed. The single number above hides significant variation when the WiFi is active.

Pico, Wireless Pack, 12V supply

I used a Pololu D24V10F52 buck converter to drop 12V to 5V, and then supplied that to the Pico. Internally it uses the Intersil ISL85410 DC-DC converter.

StateCurrent draw / mAPower consumption / mWEfficiency
WiFi not in use16.519891%
WiFi in use40.548694%
Light sleep13.516290%
Deep sleep1012092%

Conclusions

I learned a couple of things:

So one disappointment and one nice surprise!

Update: If you care about low-power, it is probably more sensible to use a pure ESP32 design.