English is currently available.
Cellular, GSM & LTE DISCUSSION
My SIM800L breakout answers AT commands from an Arduino Uno, but when it searches for the network the LED flashes quickly for a few seconds and then the module restarts. I am powering it from the Uno's 5 V pin through two diodes to get roughly 4 V, with jumper wires on a breadboard.
What supply does this module really need, and what is the proper way to provide it in a project that otherwise runs on 5 V or on a single lithium cell?
The module wants 3.4 V to 4.4 V on VBAT, with about 4.0 V recommended, and it draws current in short bursts of up to 2 A while transmitting. GSM is a time-division system: the module transmits in one slot of 577 µs out of every 4.615 ms frame, so the supply sees a pulse of up to 2 A about 217 times a second. The average current is modest, but the peak is what matters.
When the supply or the wiring cannot deliver the peak, the voltage at the module dips below its shutdown threshold during the first high-power transmission, which happens exactly when it tries to register. The module then reboots, and that is the loop you are seeing.
The diode trick fails on two counts. The Uno's 5 V rail comes from USB or the on-board regulator, and neither can supply 2 A pulses. And the drop across a diode is not constant: a 1N4007 drops roughly 0.6 V at a few milliamps and close to 1 V at 2 A, so the module sees a rail that swings by several hundred millivolts on every burst.
Wiring adds to it. Breadboard contacts and thin jumper wires can easily total 0.2 Ω in the supply and ground path, and 2 A × 0.2 Ω is a 0.4 V dip by itself. Use short, thick wires soldered to the module, with the ground return taken straight to the supply.
A workable supply is a buck converter rated for 3 A, set to 4.0 V and fed from 5 V or more, with a low-ESR bulk capacitor right at the module: 470 µF to 1000 µF plus a 100 nF ceramic. The capacitor helps with the edges but cannot carry the burst alone. From I × t = C × ΔV, a 2 A pulse lasting 577 µs would pull 1000 µF down by 2 × 0.000577 / 0.001 = 1.15 V if nothing else supplied current, so the converter and wiring must handle most of the peak.
Set the converter's output voltage before connecting the module, since more than 4.4 V can damage it, and avoid linear regulators unless they are rated for the peak current.
A single lithium-ion or lithium-polymer cell is the other standard answer. The module was designed for phone batteries, so a cell's 4.2 V when full is within its range and a healthy cell delivers 2 A easily; the module shuts itself down as the cell approaches empty. Connect the cell to VBAT with short leads and still fit the bulk capacitor.
Treat the cell with care: use one with a protection circuit, charge it only with a proper lithium charger IC or module, protect the leads against short circuits, and do not leave it charging unattended on the bench. If the rest of the project needs 5 V, boost from the cell for the microcontroller rather than bucking down for the modem.
Two related checks. The SIM800L's UART uses about 2.8 V logic. Its TX output is normally read correctly by a 5 V Arduino, but the Arduino's 5 V TX must not drive the module's RX directly. Use a divider, for example 10 kΩ in series and 12 kΩ to ground, which gives 5 × 12 / 22 = 2.7 V.
To verify the supply fix, ask the module: AT+CBC reports the supply voltage it measures in millivolts. The status LED shows the state as well: a flash about once a second means it is still searching, and one flash every three seconds means it is registered. Better still, put a scope on VBAT at the module pins and look at the dips during registration.
Back to Cellular, GSM & LTE community
