ESP32 Forum

I've now connected my board (from the same supplier) to a 3.3V supply. It is drawing 10.8mA in deep sleep. The red LED is drawing 1.5mA so I expect most of the remainder is being drawn by the UART bridge and perhaps by back-powering of the regulator.

Anyway with 10.8mA being drawn, the batteries should last over a week in theory. The booster will be wasting some power but conversely should allow operation down to a lower battery voltage than the voltage on which the figure of 2000mAhr is based.

As your batteries are lasting only 15 hours, that implies that your board is drawing 133mA on average in theory. Of course, with the approximations involved, the actual current may not be close to this calculated current.

Kazuya91 wrote:

meneldor wrote: Unfortunately the Lolin32 is not in production anymore.

Really? That would be a pity. Currently plenty of them are still available.

On the Lolin32 I wonder how to put the USB UART in sleep mode.
Is it enough to disconnect it from USB?

I'm using a 3.3V Arduino Mini (8MHz) with ripped-off voltage regulator and power led, which I supply by two Alkaline batteries connected to the 3.3V VCC pin directly. In deep-sleep, this should last for 57 years. When waking up, it powers a step-up-module via mosfet to send a high-power 433MHz ping, then cuts it off again.

I wonder if you could power your board with 3.0V on the 3.3V pin directly, just like I do. You could eliminate the converter which eats power all the time.
Or use a 3.7V LiPo and put about two diodes in series to drop the voltage from 4.2 (when fully charged) to 3.3V or below. You would have to test some diodes as their voltage drop rises with the current running through (at least that's what I measured).
Just my 2 cents, I did not read the whole thread.

Hexman64 wrote: Just my 2 cents, I did not read the whole thread.

You should I just received today some ultra ldo's to drop from 3.7V directly to the 3.3v pin.

Hexman64 wrote: When waking up, it powers a step-up-module via mosfet to send a high-power 433MHz ping, then cuts it off again.

ESP32 could be VERY hungry when initializing the Wifi.

Hexman64 wrote: I'm using a 3.3V Arduino Mini (8MHz) with ripped-off voltage regulator and power led, which I supply by two Alkaline batteries connected to the 3.3V VCC pin directly. In deep-sleep, this should last for 57 years.

This is 20 times slower than esp32. My is halfed down to 80MHz. I have some projects using attiny's but they can't be compared.

Ok, so I missed that you're eliminating your power converter(s) already.
Alkalines should be able to provide 1A, by far enough to initialize Wifi (I read up to 800mA somewhere. Never measured >400mA, but I might have missed 800mA peaks which could drop the batteries' voltage below a working value).
I just mentioned the Mini as an example for a working setup on batteries. Of course the esp32 requires more power, but still could be powered the same way, I guess.
You know about it, you have a good solution, all good.

So, i tested with 700mAh LiFePo4 battery - 23 hours. Sleeping only a minute and reinitializing the Wifi doesn't seem to fit a battery project Ive never used BLE but if its less power hungry i might try it.

Those boards eventually need some efficient buck-boost converters instead of voltage stabilizers that waste and dissipate battery energy into heat.

Oh, i forgot to mention. Lifepo4 is connected directly to the 3v pin. Its 3.2v and doesn't need stabilizer.

Not directly ESP32, but for LoLin32 boards:
Get the onboard CP2102 to low power:
viewtopic.php?f=12&t=5305