ESP32 Forum

Hi all,

I am working on a ESP32 project that is powered by 3.7V Lipo batteries. I need the circuit to have very low current consumption when the ESP32 is in sleep mode.

I was looking at the ESP32-MeshKit-Sense Hardware Design Guidelines (https://github.com/espressif/esp-iot-so ... uide_en.md) and it recommends a ETA3425 for the power supply management portion. The ETA3425 is a 1uA Ultra-Low Quiescent Current 0.6A Output Sync BUCK.

This part seems hard to purchase and is not easily available on Digikey/Mouser etc. Is there a similar part that I can use? Would anyone be able to recommend something that you have used in your project?

Also, what is the reason for using a buck converter in the reference design? Why wouldn't a regular LDO work? I used the MIC5219-3.3 extensively in many of my designs for battery powered projects in the past. Am I doing something wrong all this while with just using a standard LDO?

Thanks in advance!

Also, what is the reason for using a buck converter in the reference design? Why wouldn't a regular LDO work? I used the MIC5219-3.3 extensively in many of my designs for battery powered projects in the past. Am I doing something wrong all this while with just using a standard LDO?

I'm not familiar with the ETA3425, but to answer this question, a switching regulator will generally be more efficient when there is a large enough voltage difference between Vin and the regulated output AND the switching regulator itself draws very little current, so, in your case, where Vin is 4.2V in when the battery is fully charged, the voltage difference is 0.9V.

An LDO will draw an equivalent current (plus a very slight amount for the regulator circuit) on Vin as the load so the efficiency is virtually equivalent to Vin/Vout, or in your case 3.3/4.2 = 78.5% (or I * 0.9V wasted power).

The switching regulator, on the other hand, can operate above 90% efficiency at 'normal' operating currents, significantly reducing power consumption at the normal operating currents and is recommended when the circuit is alway ON (not sleeping). According to the ETA3425 datasheet, the Efficiency vs Iout with 3.3V out and 4.2V in, is above 78% down to about 8uA load current and is above 90% when the load current is above 50mA (normal operating range for the ESP32 when awake). That's very good.

An LDO isn't a bad choice in this case, it's easy to implement and will be more efficient when sleeping, but a switching regulator will save a lot of power if you spend a significant amount of time operating in a normal state.

Which is optimum depends on the ratio of time spent sleeping vs time spent awake.

NOTE: Updated from an earlier version where I was using the wrong efficiency chart (1.2V Vout rather than the 3.3V).

Thanks TomWS1!

I am looking at a design that sleeps most of the time, waking up only once every 5 mins to read a sensor and post to a MQTT broker.

Looks like a LDO would be a better choice in this case.

I managed to shortlist a few from another thread in the forum: https://www.esp32.com/viewtopic.php?f=2&t=5018

In the end I am going with the XC6210.

Why would you go for a 6210 over the 6220? I think it's perfect for ESP32: up to 1 amp in high power mode, and seamlessly switches to Green Operation mode (nominal 8uA but even less than that in my real world experience) when load current drops to less than 1mA (ie. deep sleep).
Also slightly better dropout voltage, though the 6210 is decent too.