ESP8266 Community Forum

I think this is a power issue.
I have a setup running now where I use the two GPIOs of an ESP-01 to control one of those 5V double relay modules. The box is powered from an external 230VAC-12VDC supply. Then in my board with the ESP I have a small DC/DC buck step down converter that provides 5V for the relay module. Then (I'm lazy) a miniature module with a 1117 provides 3.3V for the ESP.
I have not seen the need even for decoupling C's, as the unit runs nicely without.
It has not been running for long, but I haven't seen a reset yet.
My theory is that those little buck converter boards do a great job in providing the current needed with a fluctuating load. Unfortunately I do not have a scope, so I cannot investigate further, I just see that it works.
And they are cheap; you get twenty for 10$:
http://www.aliexpress.com/item/20-pcs-Ultra-Small-Size-DC-DC-Step-Down-Power-Supply-Module-3A-Adjustable-Step-Down/32262311443.html

I'm almost sure it's not a power supply issue. I still use the same cheap AC/DC 5V Chinese PSU http://www.ebay.com/itm/371231570130 and only modified the RST line after measuring spikes on it with a oscilloscope. Since this modification more then 3 day's ago, the ESP-01 runs flawless switching a 5V relay using MQTT.

Sweetpants wrote:I'm almost sure it's not a power supply issue.

After having done some experimentation, I agree with you. It is known that the ESP draws current in burst, so needs a solid power supply. But the phenomena discussed here surely an "over sensitive"

Touching RST with my finger makes the ESP reset when I switch the relay (with my soldering iron as load).
Touching RST with the tip of a screwdriver resets it immediately, no switching required.

Interestingly, I have switched my circuit (with load) manually maybe 100 times now; not a single reset detected. I will program it to sit there and switch continuously for some time and see what happens - and report back.
EDIT: I ran 500 on/offs now without a glitch. So considere my module sufficiently insensitive to be installed.

Apparently, some of us have substantial (noise-induced?) reset problems while others (like me) are more lucky. Maybe this is a chip batch issue, so that some modules become more sensitive than others?

Here is my not yet fully assembled relay and ESP board in test mode. Short distance to "live" relay.

(Deleted)