I keep burning out ESP32 boards

Either those regulators you bought are bad (not uncommon for Aliexpress) or you have a short somewhere, maybe even on one of the flap unit modules.

Soldering that WROOM module right to the motherboard makes for a lot of extra rework.

1 Like

The unit module is not connected yet, I don't connect the unit module if the control unit isn't working ( I do this so that I know that I will not damage other components).
I don't think the problem is the PCB itself because I manage to get one to work but on C2 instead of having the 10uf 50v capacitor I used by mistake a 100uf 16v one.
It might be a coincidence, I don't know but it's the only one that works.
All other don't work, maybe it's a bad batch of capacitors? I have no clue but all the other boards have the same problem, (out of 6 only 1 worked.)

I know I had to buy a desoldering tool for it, but I already ordered a 19pin connector so for the future boards I will be able just to swap.

That could mean bad/cheap capacitors. The LD1117 needs a minimum of 10uF to work properly. If it's a +/-20% capacitor then it could be as low as 8uF and that could cause problems. From now on, I recommend you use at least a 33uF for C2. If you already have a bunch of 100uF, then use them.

Ok I will try again with the 100uf 16v about the 33uf what voltage should I get?

Anything greater than 10V, 16V is fine.
Also it does not have to be 33uF. 47u, 68u, 100u are all OK

If it's the same I will use the 100uf for testing since I have a lot of them, I will try and let you kno if something changes

No hurry.
Rather that burn out another ESP connect a 100 ohm resistor between 3V3 and ground and check that the voltage is 3.3V

Regulator looks the right way around. I always use female pin-headers on the PCB so the Devboard can be removed easily and change if needed. I am also out of ideas other than to check the voltage coming out of the regulator. If the capacitor beyond the regulator is to big that may cause damage to it on rare occasions. ESP32 boards do warm quite a bit when they are using WiFi and even more so when running at 240MHz. But if the whole thing isn't working properly then probably there really is something wrong. I mean if they burn out something is wrong. Check all of the ESP pins for shorts between the pins and between GND & pins and Vcc(3.3v) and the pins. Unless there is also a diode somewhere, the resistance should always be at least 100 Ohms (more usually)

With a bare card, nothing at all installed, do you read any resistance at all, or infinity, between all power paths? Test GND/Vin/3.3V etc, be sure all are infinite, OR that you can explain why it isn't.
This, of course, may be too late if you have no untouched boards left.

↑ ↑ With all external power 100% Disconnected ↑ ↑

1 Like

I decided to make the next board step by step, first I wanted to see which voltage regulator the PCB that is working is using and it's this one:

I connected the PSU jack, C1 the voltage regulator and C2.
The voltage regulator works fine and it doesn't make heat, I mesured C2 and it give 3.3V
What should I do next?

let's assume that the voltage regular was not working then could it be possible that the 9V fried the ESP32 and the rest of the components?

I think too late for that board, but what do you mean by infinity?

Most meters show OL when there's no path between meter pins - that's by definition infinite, as far as the meter is concerned. If you read resistance, try to figure out why, because two traces even a few thou apart should have that level of resistance at the voltages most meters use for reading resistance.

Please note that this is applicable only to a board that hasn't been soldered to. If you've soldered, then desoldered, there's a good chance you've left residue on the board that may turn up as a very high resistance, confusing the issue.
But you're not hunting for high resistances. The problem you're describing can only be

  • wire trace from one power to another power, or from a power trace to a gnd trace
  • solder bridge from power to ground
  • solder bridge from power A to power B
  • failed component
  • miswired component - voltage routed to a ground pin, ground pin routed to voltage, etc. etc.
1 Like

Connect a 100 ohm resistor between 3V3 and ground and see if it's 3.3V at the 3V3 pin.
Are you using the 100uF cap?

1 Like

Then check each of the ESP pins on the PCB and make sure that only the 3V3 pin has 3.3V.

Then, remove the 100 resistor and the connection to the 9V supply. Connect the negative lead of your ohmmeter to the 9V minus (-) terminal and check each of the ESP pins on the PCB to make sure none are shorted to the minus. If your meter has a continuity beeper use that, it should NOT beep on any pin.

Ok so this is something I noticed on the one that is not working, the ESP pins are shorted with the ground, why do you think this happened?

Before to solder the esp32 to the new board I believe it's better to wait for the 19 pin header this time, I have a 20 pin, maybe I can just cut one so I can start testing? because it will take another week before I will get the 19 pins from china, I can't find them on amazon

I will try thanks for the info