Module fails to power up; EN issue

[Gerry48]

Some modules fail to power up. Power source is a12V battery. A Mini MP1584EN DC-DC Buck Converter drops voltage to 6V. The 6V powers the ESP32 module. It takes about 70 ms for the output of the DC/DC converter to ramp to 6V (yellow trace). The ESP 3.3V ramps up in 40 ms (purple trace).

P1004336.JPG (61.72 KiB) Viewed 1239 times

Is it possible to add capacitance to the EN pin in order to delay turn-on?

[ESP_Sprite]

Yes. If you're using a module, the datasheet has a section about the power-up behaviour that allows you to size your reset RC network properly. Alternatively, just increase the capacitance on the EN line until it works, then add a bit more to have some margin.

[Gerry48]

The difference when adding 47 uf between pin EN to ground is significant.

Scope plots:
yellow = 6V power into 5V input
purple = IO pin 14

Code: [Select all] [Expand/Collapse]

1.   void setup() {
2.   pinMode(14, OUTPUT);
3.   delay(10);
4.   digitalWrite(14, HIGH);
5. }

GeSHi © Codebox Plus Extension

No added capacitance to EN:
During power-up IO 14 begins to ramp up to 3.3V after 23 ms delay. It takes IO14 550 ms to go low for 10 ms (before the digitalWrite High command.

No added capacitor

P1004362.JPG (22.78 KiB) Viewed 947 times

47 uF aluminum cap added to EN:
IO 14 does not ramp up. It goes to 3.3V after 90 ms delay. It takes 258 ms for IO 14 to go low for 10 ms. Adding the longer initial delay reduces the time IO 14 to become active by almost 50% (258 ms vs. 550 ms).

47 uF to EN

P1004361.JPG (22.1 KiB) Viewed 947 times

Are there any negative effects by add that much capacitance?

[ESP_Sprite]

Are there any negative effects by add that much capacitance?

No, not that I know of (aside from the implicit 'a slightly longer time before the module comes out of reset')

[Gerry48]

With the 47 uF cap the module comes faster out of reset: 258 ms vs. 550 ms.

If you look at the top purple trace, you can see multiple shallow dips in the signal. I think the processor is trying to get started but can't as 3.3V isn't stable yet. It takes 550 ms before it decides to get going; even though 3.3V is stable after 70 ms. As stated in the OP, sometimes it doesn't start at all. It seems the design could use some improvement in this regard.