Custom board keeps killing chips

[newsogn]

<t>Hello,<br/>
I am working on my own ESP32 board for home automation. I use an HLK 5M05 for a main power supply then I use an AMS1117 to bump it down to 3.3V. I am using basically the same circuit as the documentation shows for peripheral schematics. to program the chips I am using the arduino IDE with a CP2102 usb to serial module. <br/>
<br/>
I will build a board and it will work just fine for awhile, but then it will start to kill the program the serial message is error in main.c<br/>
<br/>
Then I will reprogram it and it will be just fine, but then it will start doing it again, but this time something happened to the internal pull-up on IO0 because the serial message then becomes "ready for download" as though I hit reset with IO0 pulled low. at this point if I hit the reset button then it acts as though IO0 is held low, but the boot button is not being pressed. I measure about 0.6vdc at IO0. To solve this I put a jumper from IO0 to 3v3. this allow me to upload the code a few more times, but after a bit I start to get errors like "could not write to target ram". <br/>
<br/>
as for the power supply I have tried using the 2200uf capacitor in my schematic, and I have replaced with a 10uf cap like the documents say to use. I have used a DSO nano from work to scope the 3.3v rail and it seems to be nice and smooth with little ripple, but it is a cheap scope its hard to tell. <br/>
<br/>
I have noticed that I have been using 0806 ceramic capacitors for the 0.1uf capacitor in the schematic and they seem to be getting fried as well because generally if I replace that then I can program the chip again. but I switched to an electrolytic cap and it seems to work fine now, I have 3.19vdc at the en pin when not pressed.<br/>
<br/>
My question is why do I keep killing these chips? is there something wrong with my schematic or maybe implementation? do I need a pull-up resistor on IO0?</t>

[ESP_Sprite]

Hmm, I took a look at your schematics and PCB and I don't really see anything that looks directly off. Only thing I would add is some 100n or 1uF or so decoupling capacitors from both the input and output pins of the AMS1117 to ground, but seemingly the AMS1117 datasheet does not mention those, so it should also work without it. You could still try adding it: to be honest, the 5V line with no decoupling at all on one hand, and some relays powered from them on the other hand, sounds somewhat scary; adding some capacitance there (eg an 100n plus an 22u) may fix your issues.

[Markus Becker]

... a bit late, but...
I would not recommend 1N4001 for D1 and D2, rather something faster, like 1N4148. And may be add small caps from the relay coils to ground.
Best
Markus