ESP32 Pico D4 Custom PCB not working

[dan_jim_23]

Hello,

I have realized a custom PCB design for the ESP32 Pico D4 and I’m not able to program it.
To reduce the components needed I use a separated USB-FTDI adapter to program the SIP.
I tried to reproduce the schematic of a similar board that successfully use the USB-FTDI adapter.
Because I don't need the Wifi part I didn't connected anything on the antenna input.
Debugging with the scope on the USB-FTDI lines I cannot see a “TX” signals, RST and IO0 are good.

I have seen in the forum that many other had similar problem but yet not many have solved..
Please can someone advise me if I did something wrong in the design.

Thanks
Dan

[mspider65]

Did You also received recently the pcb assembled from JLPCB?

[dan_jim_23]

sorry no, I just let it assembly in another factory.
Anyway I'm aware this BGA components aren't easy to solder, but after reading how many have similar problems (no communications on the TX) now I'm wondering if there is a lot of defective chip around..
because I made two boards and no one works.
I still have two others chips and honestly I would like to know if they are working.

but honestly I know I could have made a mistake in the design, for this I hope someone could point out the error.

[ESP_Sprite]

Do you get any communication with the ESP32 at all, that is, does the chip send you anything over serial when you reset it?

[dan_jim_23]

if you mean on the U0TXD (IO1) unfortunately not, it alwyas stay high (when the other working board which usie the USB-FTDI adapter, sent bytes).
checking on other design I also tried a modification with a pull-up for EN/RST with a 10 k resistor to +3.3V but didn't changed anything
how said I check also the EN/RST line on the scope and this during the upload process goes regularly to 0.

Anyway something is probably wrong with the design, chip or soldering: when I supply the board externally with the 5V I see only 3-4 mA, when I read from datasheet must be ~80 mA.

[deandob]

I had a similar problem with the Pico, and solved it by ensuring the decoupling capacitors were placed close to the chip and was very careful with the earth under the chip so that only current destined for the ESP32 was routed via the ground pad, so no other transient current could interfere. My board is working reliably now.

[dan_jim_23]

in the last days I have been able to make some improvements: the chip seems to works (most of the time) and probably the biggest problem was that it has not been soldered correctly.

[NiclasH]

I have had a similar problem before, and it was due to too much solder paste on the GND under the chip. Using a "checkered" pad resolved the problem. I.e. add solder mask on top of the pad so that you get 9 "islands" of solder paste, or even just make the solder paste layer into such pattern. A good PCB assembly place should probably have spotted this, but mine didn't (nor did I think about it).

[jpfaguirre]

Did You also received recently the pcb assembled from JLPCB?

I just received the pcb assembled from JLPCB and I have a similar problem with my ESP32 Pico D4 Custom PCB. Is there a problem with JLPCB?