Arduino Fails to Program Custom NRF52, But Has Worked In Past

Question

Background:

I have a custom board that consists of two PCBs connected via a 100mm Picoblade connector. The "Aux" board has a USB port, and CP2102N Serial=>USB converter. The "Main" board has an NRF52 moduel (MDBT42Q by Raytac).

I have 5 boards that I successfully flashed with the boot-loader via a JLink programmer. One board after flashing the boot-loader I was able to program ~5 times via USB with Arduino.

Come back a week later, I can not program the original board, nor can I program any of the other 4 boards. My friend is also unable to program the boards using both a Mac and PC.

Investigation: For anyone familiar with this boot-loader, it expects a "softdevice" to be programmed, if one is not found, such as after burning the boot-loader, it starts up in DFU mode, which can be identified by a flashing LED. The "MODE" led on my board is flashing thus indicating the boot-loader is working.

Examining the physical signals on the board, it looks like the RX line is not being pulled to 0V, more like ~1V on a 3.3V logic level. There does appear to be some serial data on it though. Arduino is giving an error:

"Failed to upgrade target. Error is: No data received on serial port. Not able to proceed."

So this seems like it is realistically being caused by the NRF52 not driving the RX line low fully. I tried placing a 22k pull-down resistor from the RX line to ground, and then the CP2102N/NRF52 faults out upon communication and the RX line sits forever at ~1V.

The most frustrating thing is that I have programed one of these board multiple times, and "magically" it no longer works. Since I have 5 boards with the same issue, I find it hard to believe it is local damage like ESD etc.

Any ideas would be much appreciated.

Schematic:

enter image description here enter image description here

• Tony Stewart EE since 1975: Can you isolate Rx with no load to measure output Vol then for input pull down Rx input to measure input current. V/R. Perhaps grounding or supply sequencing has changed. Verify
• MadHatter: That's a good idea, I can disconnect the lead on the picoblade connector. I'll check that tomorrow.

Answer 1

Turns out I forgot to hook the ground pin up to anything as is visible on the schematic... I guess programming once was luck, or something with the way wires were located they provided enough of a ground... Amazing it worked at all...