ESP8266 power supply oscillating with audible noise and noticeable droop

Thread Starter


Joined May 3, 2010

I have recently built an electroencephalogram (EEG) headset and acquisition hardware using the OpenBCI open-source platform. Instead of using their Bluetooth dongles, I opted to build their WiFi Shield board. I have been able to successfully acquire data and all is well, but I have noticed that the power supply for my WiFi shield has an annoying audible noise and has occasional dropped packets when transmitting data. The WiFi board has an ESP8266 module which on average draws about 80mA but can draw up to 150mA when in operation and has an operating voltage of 3.0 to 3.6V

The WiFi Shield uses an AP2112 LDO regulator with 3.3V output. The design used by OpenBCI has no capacitors connected to VIN but four ceramic SMD capacitors connected to the 3.3V VOUT (0.1uF, 1uF, 10uF, 100uF).

For my application, I have been using a 3.7V LiPo 400mAh battery, but to investigate further I hooked up the WiFi Shield (on its own) to my adjustable DC power supply (Agilent E3620A).

First thing that I notice is that the ESP8266 puts out some kind of beacon (10 Hz, 1 ms pulse-width) that causes a droop in both the power supply and output of the 3.3V LDO. Considering that the ESP8266 requires a minimum of 3.0V and the 3.3V line looks to be dragging down to 2.7V (0.6V droop), this is not ideal. Also, interestingly, if I increase the supply voltage from 3.5V to 5.5, there is noticeable oscillation that shows up, which corresponds to the audible sound that I am hearing.

Below are screen-captures from my scope while varying power supply voltage Vbatt+.

Probe on power supply Vbatt+ (AC-coupled):

Probe on 3.3V LDO output (AC-coupled):

The AP2112 LDO datasheet does recommend 1uF capacitors on VIN and VOUT, but this OpenBCI design uses many values. I wonder if having the four capacitors in parallel in this design is throwing off the expected ESR for the LDO and causing instability at certain supply voltages.

Ideally, I would like to suppress the oscillations as much as possible (so I don't hear them) and make the LDO 3.3V output more solid so that the ESP8266 doesn't run into minimum supply issues.

Wondering if anyone has any suggestions. Thank you


Joined Mar 10, 2018
I would say no since they state Ceramic MLC is OK, which has super low ESR.
And they allow a broad variety of C technologies for the 1 uF. Plus there is no
mention anywhere concerning ESR issues.

The spike sure sounds like stray L somewhere. Scope probe ground short and
well connected ?

Regards, Dana.