Just a quick topic/thread, because I haven't found any information about this on the web so far.
I got my hands on the new ESP32C3 chip (Lilygo Tigo Plus) and already had several Esp8266 boards (D1 mini pro clones of the Wemos D1 mini pro) laying around.
I want to build several moisture sensors which monitor my plants and submit the results via MQTT to a server every 6 hours and therfor was wondering which of the two boards could be run with the lowest power consumption, without having to modify the boards too much or to use barebone ESP32/ESP8266 instead (because it looks I suck at soldering).
I have used a simple script which simply first calls a delay() functiond then then puts the ESP into deep sleep using ESP.deepSleep() and measuerd the power consumption with a poer profiler kit II (something I really can recommend). The profiler kit was used as source meter, which gives more precise results compared when used as ampere meter, and I tried three different ways to power up the esp: via USB cable, 5V in and 3.3 V in (to check the effect of the internal step down converter).
The results:
D1 mini pro:
USB port: 15.13mA (delay) / 206 uA (deep Sleep)
5V in: 15.01 mA / 95 uA
3.3V in: 15.32 mA / 121 uA
Short note: the D1 requires a cable between D0 and RST, otherwise it doesn't restart at the end of the deep Sleep cycle. My clone has no power or other LED on board. Some models have one and they draw more power.
LilyGo (original)
USB port: 53.83mA / 34.36mA
5V in: 19.55mA / 855 uA
3.3V in: 20.83mA / 840 uA
3.7V in: 19.53mA / 875 uA
Short note: the Lilygo has several additional features on board, like the grove Pin or a battery holder for a 3.7V Li 16340 accu. This pin (plus a blue LED used as charging indicator) draw some power, and at least when running the ESP on USB it constantly checks for reloading and the blue LED always flashes. On top, there is a power led (red) which is always burning. This all consumes power.
When run on 5V in/3.3V in, there is no charging indicator. The blue LED is turned off, the whole system draws less, but still a lot.
There is one adavantage: no need for a shortcut between D0 and RST, the ESP32 awakes from deep sleep without the need of the cable.
LilyGo modified
I removed the power LED "by force" from one of the boards
USB port: 51.98 mA / 33.53mA
5V in: 17.28mA / 34 uA
3.3V in: 18.93mA / 34 uA
3.7V in: 17.89 mA / 29 uA
Conclusion:
At least from my perspective it is really worth using the LilyGo instead of any ESP8266 board if I don't want to spend a lot of time in redusing the power consumption of the 8266 by using a barebone chip and adding the required components to run it (resistors and a step down converter).
All you have to do is to remove the red power LED and then you have a system which can easily be programmed via USB and run on a battery for a long time. Yes, the 29uA are still higher than the optimum (5uA according to the data sheet of the C3), but it's good.
If the price (1.75 USD D1 mini pro vs. 4.00 USD Lilygo) is relevant, then use the 5V in on the D1 when running on batteries.