Dead M4 Express after a few days Moderators: adafruit_support_bill, adafruit

[zuritero]

Hello,

I have 6 of these units that failed after a few days or running (1-4 weeks); normally these units will operate continuously for many months. The symptom: 5 / 6 are no longer detected by the computer, the 6th one is detected, but can no longer upload FW. These units "turn on". meaning that the yellow LED blinks but I am unable to communicate with them using a computer. When pressing the reset button twice, the big LED turns red (not green), and it stays that way.

We have multiple other units that were bought at the same time which haven't failed. These working units have bootloader v3.10 (latest one is 3.14); I will assume that the failed units also have the same bootloader FW. These units bought at the same time have a chip date code of 1903, and a board code (8 part length, left of the word "express") of 3320. I mentioned this because one of my suspicions is a bad batch, however, it's hard to conclude at this point.

My primary focus is to better identify units at risk, and weather I can do anything to prevent the failure upfront (e.g. upgrade bootloader FW to v3.14 which I am already doing); this is causing painful / expensive downtime. My secondary goal would be to recover the units which already failed. Any ideas on the primary and secondary focus would be much appreciated.

Thank you

Omar

[mikeysklar]

Omar,

Thank you for the summary.

What kind of things do you a have connected to these boards (sensors, power supply)? Do you have a photo of how they are connected?

What is the environment like (temperature, mositure)?

What code are you running? Is it CircuitPython or Arduino? Can you post it?

The bootloader can be normally be recovered on these boards using JTAG/SWD debugger.

https://learn.adafruit.com/how-to-progr ... ootloaders

The solid red LED after double-reset indicates a USB failure.

The NeoPixel is powered by the 3.3V power
supply but that hasn't shown to make a big difference in brightness or color. The
NeoPixel is also used by the bootloader to let you know if the device has enumerated
correctly (green) or USB failure (red).

https://cdn-learn.adafruit.com/download ... samd51.pdf

[zuritero]

Hi Mike,

Thx for your comments, please read below for my replies:

[quote}
What kind of things do you a have connected to these boards (sensors, power supply)? Do you have a photo of how they are connected?[/quote]
The M4 Express is powered via the microUSB connection, with a Raspberry PI on the other side, we have sensors connected to it as well, and other electrical components like resistors, capacitors, etc ...

What is the environment like (temperature, mositure)?

They fit inside a water tight aluminum housing, which is exposed to the elements (sun, wind, rain, snow in some cases). But to point out, the units that failed pre-deployment were always kept insider in our office space.

What code are you running? Is it CircuitPython or Arduino? Can you post it?

It's running Arduino, I can't post it due to IP (hope you understand); however, we've been running the same code for more than a year, and we are only now seeing this issue.

The solid red LED after double-reset indicates a USB failure.

Good to know, can you think of a reason why it would fail in my case?

Thank you

[mikeysklar]

Your units that respond with a solid RED LED when you double press reset to enter bootloader mode are showing a bootloader that is stuck so USB enumeration fails. Powering direct off the USB of a Raspberry Pi might be an issue. Are you using the 3x or 2x ports? Can you use a powered USB2 hub instead?

Do you have a JTAG programmer you can use restore the bootloader?