custom ESP32 board produces "garbage" on serial

[pietnl]

I'm having some trouble with a ESP32 based custom board I had created recently. I had 5 prototypes created, 3 of them work fine, but 2 of them only spit out garbage on the serial output, like this:

serial garbage.PNG (111.33 KiB) Viewed 7170 times

I can't get these boards into programming mode, not even when I manually pull gpio0 lo and reset them.

The board is using a ft232rl ic as usb to serial bridge.

The odd thing is that with a previous batch of prototypes, I had the same kind of problem (2 out of 4 boards working) while using a different usb to uart bridge (CP2102N) and while having them assembled at a different party.

I tried reflowing the boards, but that did not make any difference (although that could easily be due to my soldering skills).

Is there anything I can do in my design to prevent these kind of issues?

[ESP_Minatel]

Hi,

Maybe sharing your schematic/design could let others help you!

[ESP_Sprite]

Generally, given that there's some repeating boot messages in your terminal, I suggest you check your power supply. If the ESP32 gets 3.3V somehow, but the power supply can't even deliver enough current to get through the initial boot process, this is something you can see.

[pietnl]

Sorry, forgot to attach the schematic

dyli screenshot.PNG (20.7 KiB) Viewed 7078 times

I basically started out with the schematic of the Adafruit Huzzah32 and adopted it to my needs

[ESP_Minatel]

Hi,

Not sure if I can read the part numbers properly. The image is in low resolution for me.

As my colleague ESP_Sprite mentioned, you need to be sure that you are providing enough current or using a good quality USB cable (if powering using USB).

I think you are using the AP2112 LDO. Is that right?

[chegewara]

One more question, what XTAL you are using?

[pietnl]

Hi,

Sorry for my late reply. I'm indeed using the AP2112 3.3.

What do you mean by xtal? I do not have any crystal on the board I believe.

[ESP_Sprite]

+1 for a better resolution schematic. How is your LiIon fuel gauge connected, btw? Because as it is, I see a 0.01 resistor effectively shorting Vbat and +3.3V... the ESP32 is not gonna like running from a fully charged LiIon in that case.

[pietnl]

Attached a better resolution image

schematic.png (33.18 KiB) Viewed 6889 times

The fuel gauge (BQ27411) circuit does indeed have a 0.01 ohm resistor between 3.3v and VBATT, I copied that part of the circuit from the sparkfun battery babysitter: https://cdn.sparkfun.com/assets/learn_t ... ematic.pdf and it does seem to work well for the boards that do not produce "garbage".

One more note about the AP2112, I do have another issue that when the ESP32 is in deep sleep, the board still uses to much power (3mA), so I am considering swapping the AP2112 for a different LDO. I found this guide in which they conclude that the LDO on the FireBeetle (RT9080-33GJ5 ) consumes less power, so I might go for that one.

[WiFive]

I think you are misunderstanding the sparkfun schematic. It should go battery->resistor-->vbat. Vbat is not directly connected to the battery.