Hello,
I am powering my M66 module with the TPS27081A, but from time to time (sometimes once every 15 days or once a month), the module locks up and stops sending data. However, when I disconnect and reconnect the power socket of the circuit, it starts sending data again. What could be the reason for this? Could it be caused by soldering issues?
Hi,
This problem is like a “half-reset” problem caused by a 90% power supply/sabotage. The M66 requires a momentary 1.8–2.2 A peak, especially during GSM TX bursts; a millisecond sag in VBAT or a slow/half-cut power module will lock up. Removing and reconnecting the power connector recovers the module because it’s “full” cut.
The problem appears to be hardware-related, so you can check if the problem reoccurs by following these steps:
Add a 470–1000 µF + 2×100 µF MLCC to the bottom of the module, thickening the VBAT trace.
Add a bleeder after the TPS27081A → full cutoff.
Verify minimum VBAT during TX with the scope. If you see <3.5 V, redesign the power path.
Additionally, it would be a good idea to send a reconnection command when disconnected from the base station.
Periodically query AT+CREG?
If not connected for a certain period of time:
AT+CFUN=0 → shut down the terminal
AT+CFUN=1 → restart
Alternatively, you can reset the module with the AT+CFUN=1,1 command.
Hello Emre,
I have a circuit that is running in the field. However, as I mentioned, sometimes I lose the signal once a week, sometimes once a month, and occasionally it works fine for up to 3 months. But after that, the issue reappears and I can’t receive any signal, sometimes even twice a week. When I unplug and reinsert the connector of the circuit, the problem is resolved.
As you suggested, I focused on the input capacitors of the M66 module, but the VBAT pins and the VRTC are powered from the same source, and frankly, I don’t understand the reasoning behind this. Isn’t the VRTC pin supposed to be for a battery? Could powering both from the same source cause the signal loss issue I’m facing?
Apart from that, on the VBAT and VRTC inputs there are capacitors connected to GND: 100µF tantalum, 10µF, 100nF, 10pF, and 33pF. Could the issue be caused by this configuration?
Hello,
You can monitor VBAT with a data logger to see if there’s a voltage drop.
VBAT and VRTC are two separate pins. One is used for power supply, the other for clock. Connecting them to the same location is not recommended. This can be one of the freezing issues. You can use 470uF tantalum (Low ESR).
Connecting VRTC to VBAT can bring around 4.2V to the VRTC pin (charged battery/boost). If this pin isn’t designed for this, internal domains may behave abnormally: not responding to AT commands, half-open state, or locking.
2A bursts on the VBAT line can also trip the VRTC connected to the same line. The RTC/low-power block will shake → in some modules, URCs will be interrupted, socket flows will be distorted, and there will be a delay or no response to AT commands.
The best thing to do: Disconnect VRTC from VBAT.
If you don’t want an RTC: leave it blank (connect a 0.1–1 µF C if necessary).
If you want an RTC: Power it from a 3.0–3.3 V low-noise source (small LDO + 0.1–1 µF). If you’re using a coin-cell/Supercap, isolate it from the main 3.3 V with a Schottky diode.
For VBAT: low-ESR 470 µF (close), +100 nF bypass, thick line/ground plane, load-step test if possible (aim for <300 mV droop on the oscilloscope).
Thank you very much, I will try it. Most likely the issue is caused by the reasons you mentioned. I also noticed something else that I would like to ask. The TPS02781A IC has an R1/C1 pin for slew rate control. In the circuit, C1 is connected from the line going to the M66 to the R1/C1 pin, while R1 is connected from the same pin to the input pin of the IC. R1 seems to be positioned correctly, but shouldn’t C1 go from the R1/C1 pin to GND instead of the line going to the M66?
Thanks in advance, sorry for taking your time.
Honestly, I don’t understand the purpose there. Normally, an internal current source charges a capacitor connected to GND, slow-ramping the MOSFET gate. Therefore, that pin should be connected directly to GND. The voltage fluctuation experienced during TX will affect slew control if connected to the line going to M66.
R1 is connected correctly.
I hope applying these will solve the problem.
You’re welcome 