BLE initate current consumption

[R0815dd]

Hey guys,

I know it´s not the first thread about BLE current consumption. But I can´t find a proper solution for my case.

I use a custom built PCB with an ESP32 -WROOM-32E. I want to transmit data via BLE, but I get a brownout every time BLE initate. If I power the PCB via ESP-PROG it works fine. So I measure the current consumption at the boot sequence. I measure the current, drawn from 5 V input to voltage converter. The peak current is about 700 mA peak and draws more than 400 mA for more than 30 ms. After the boot sequence it just draw 50 mA.

I also tested this with the Espressif Dev Kit and the sample code BLE_server for Arduino and get the same values for the current consumption.

As I read a lower CPU frequency might help, I can´t confirm this. It takes the same current at 80 MHz, as well as 240 MHz.

The image shows the current consumption over time.

Thank you for your help.

[becorey]

The recommended min power supply is 500mA 3.3V. Is your power supply smaller than that, causing the brown outs?
Little larger power supply and/or more output capacitance can help avoid the voltage dip brown out.

[R0815dd]

I know, that a 500 mA power supply is recommended. But I used a 400 mA (continuous) switching converter instead of an 500 mA LDO. The switching converter is able to deliver up to 700 mA for short time. Maybe the rise time is too long for the ESP32. I am now able to to supply the esp with an inrush current limiter (p-mos) and a 1 mF capacitor.

What I´m really looking for is a way to reduce the current during the BLE init. The 1 mF is quit big and makes the circuit board not pretty . Maybe you somebody know a solution for this.

[nanoant]

I can only confirm these measurements. I was having problems with my ESP32-WROOM-32E based BlueRetro joystick to Bluetooth dongle compatible with 9-pin Atari style joysticks. While it was running fine on my Atari 800XL, it was is most of the cases unable to go through init sequence when powered from Amiga A500 game port that uses 4.7Ohm current limiting resistor.

When I measured the dongle with oscilloscope connected to 1Ohm shunt resistor on +5V side, I could clearly see 400mA for 30ms, which is IMHO extremely high and explains why when connected to Amiga 4.7Ohm current limiting resistor will cause such a high voltage drop so AMS1117 LDO will be unable to deliver reasonable voltage to ESP leading to brownout and reset.

000.png (2.89 KiB) Viewed 663 times

Therefore I have same question, if there is any way to reduce current consumption during BLE init. And if not, if any of newer ESP chips are better in this respect.

[boarchuz]

That'll be PHY calibration. Unfortunately it has only been getting longer and more power hungry as ESP-IDF grows, and espressif devs have confirmed that the process will not be open sourced.

So your options are to accept it and spec your power supply appropriately, or some non-standard approach (eg. reverse engineering).