About spike voltage of 3V3 and 5V terminal of ESP32-DevKitC-32E

[rizkubo]

Hello everyone.

I'm having trouble with voltage spikes appearing on the 3V3 and 5V terminals of my ESP32-DevKitC-32E.
It is observed when using Bluetooth communication (a2dp_sink). Does anyone know how to dampen the voltage spike on the DevKitC's 3V3 terminal?

Figure 1

Figure1.jpg (587.21 KiB) Viewed 1858 times

This is an observation of the 3V3 terminal. 200mV/DIV, 1msec/DIV. It was amplified by an inverting amplifier circuit using an operational amplifier. Gain is 100. AC coupled with a 100uF capacitor.

I observed a voltage spike of 15mV. This spike seems to be adding noise to my own loudspeaker amplifier that shares the power supply with DevKitC.
Noise at the amplifier will not be generated unless the power supply is shared with DevKitC. However, I would like to create a loudspeaker amplifier that shares part of the power supply with the ESP32.

Figure 2

Figure2.jpg (738.18 KiB) Viewed 1858 times

Figure 2 was obtained by inserting a 0.47Ω current measurement resistor between the external 3.3V AC adapter and the DevKitC's GND. The voltage across the resistor is amplified 10 times using an inverting amplifier. 20mV/DIV, 2.5msec/DIV. Current consumption is about 40mA to 150mA. I believe this increase in current consumption is causing the voltage spike.

http://www.advanced-monolithic.com/pdf/ds1117.pdf
This is the specification of the 3.3V LDO installed in DevKitC. The load regulation is about 0.06%, which is about 2mV for 3.3V.
15mV is an order of magnitude worse than this.
It appears that DevKitC does not perform as per LDO specifications. Does anyone know?

[ESP_LJH]

You could measure the current at the same time, if there is current spike, too. If yes, it is normal because when you use wifi/bt function, there will be sudden increas of current and power IC could not deal with the sudden increase timely. If you want to reduce the ripple, maybe you should use very large capacitors. If it could not help a lot, maybe you could not share the power with wifi chip.

[rizkubo]

Thanks for your comment.
The current spike seems to be synchronized with the voltage spike. I connected a 1000uF capacitor to the 3V3 terminal. The voltage spike waveform has become more gradual but is still present. It seems that the power supply has to be separated.