ESP32 LiPo/Where the power comes from/09. From the cell to 3.3 V
Where the power comes from · 09 of 13

From the cell to 3.3 V

The regulator makes a steady 3.3 V from the cell for as long as the cell stays a few tenths of a volt above that. Below about 3.5 V the rail starts to follow the cell down on every Wi-Fi burst, which is why a lithium cell's last few percent are not worth chasing.

A cell is not 3.3 V

A lithium cell is 4.2 V full, spends most of its charge near 3.7 V, and is empty at 3.0 V. The ESP32 and the flash chip beside it want 3.3 V, steady. The regulator on the board sits between them and turns whatever the cell is into 3.3 V — by dropping the difference.

A regulator like this needs a little room to work: its input has to stay a few tenths of a volt above its output. While it has that room, the output does not move. When it runs out, the output follows the input down.

Watch the cell run down

How low the cell can go before the 3.3 V rail follows it
3.70 V
Cell voltage3.70 V
Rail in a burst
3.30 V
Headroom
0.15 V spare
Held at 3.30 V. The cell is above about 3.55 V, so the regulator has room to drop what it needs even while the radio transmits. This is where a lithium cell spends most of its charge.

The drop drawn is a typical figure for a small regulator at the 240 mA of a Wi-Fi burst, not a measurement of this one — the part is not documented. The shape is what matters: flat until about 3.5 V, then following the cell.

What the ESP32 does at each stage

CellRail in a Wi-Fi burstWhat happens
4.2–3.6 V3.3 VNothing. Most of the cell's charge is here
3.5–3.4 VDips a little below 3.3 VStill works. The start of the end
3.4–3.0 VDips furtherWorks, with little charge left; resets if the cell is small or cold
Below 3.0 VThe cell is being damaged. Its protection cuts it at 2.5–3.0 V

Where to stop

Stop at about 3.4 V, in software: read the cell, and when it gets there, finish what you are doing and go to sleep for good or switch the radio off. Between 3.4 V and empty there is very little charge left, so stopping costs almost nothing and saves the cell being run down to its cut-off.

The protection circuit on the cell is the last resort, not the plan. It cuts the cell off when it gets there, which is abrupt — no chance to save anything — and a cell taken there often wears faster than one that is not. Reading the cell is the sketch that knows when to stop.

When it does not work

It resets on battery, but never on USB

The rail is sagging when Wi-Fi transmits. Charge the cell and try again; if it still resets, fit a 100 µF capacitor across 3V and GND on the shield, and turn the radio down with WiFi.setTxPower(WIFI_POWER_8_5dBm) after WiFi.begin().

It stops working with the cell at 3.6 V

The cell cannot deliver the burst current without its voltage collapsing. Small, old and cold cells all have more internal resistance than the voltage suggests. Try a larger or newer cell, or a capacitor across 3V and GND.

Can I power sensors from the 3V pin on battery?

Yes: 3V is the regulator's output whichever source is feeding it. Every milliamp a sensor draws comes out of the cell too, so it shortens the run time, and anything that draws in bursts — a servo, a motor — needs its own supply.

Where this goes next

Capacity over the average current, and what decides the average.

How long a cell lasts

Edit this page — content/books/esplipo/from-cell-to-3v3.mdx

Community

Questions about this product

See what other owners have asked, and read their solutions.

Ask a question ↗

ESP32 LiPo Dev Kit: 3 Boards with LiPo Charger + 3 Shields

Loading discussions…

Discuss this article

Ask about this page. The answer stays here, on the page it belongs to, for whoever hits the same wall next.

Browse Modules and blocks on the forum