• Spot the Difference

    04/07/2024 at 09:05 • 0 comments

    Can you tell the difference between these two pictures?

    The astute among you will realize that the board works when the USB-C is plugged in one-way and not the other.

    I bought this ESP32-S3 board as part of a batch and it just wasn't working like the others.  Firstly, I discovered that the RGB link needs to be joined for the on-board WS2811 pixel to work.  But more important was finding the cause of the fault.

    My first, and incorrect assumption was to replace the USB-C socket.  Which I did without anything coming to any harm, but it didn't fix the issue.  What I'd failed to catch was the obvious: cable orientation behaviour is based on the CC1/CC2 lines.  ALWAYS check those first.

    Sure enough, my old friend the 5k1 was shorted, but only on CC1.  So I removed it, but because I didn't have any 5k1 "in-stock", I then moved the 5k1 resistor from CC2 to CC1; which is why if you're really observant you'll see a missing resistor on the board.  The CC1 orientation then powered the board without issue, and confirmed that the shorted 5k1 was the only problem.

    So - after a few days I'll have the 0403 5k1 SMD resistor I need to make this board fully working again.  My advice to board debugging?  Never discount the most obvious things before leaping to the wrong conclusions - if I'd done that I'd have saved myself the dicey swapping of the USB-C socket with a hot air gun.

  • Never What You Think

    04/07/2024 at 08:31 • 0 comments

    So I bought an ESP-1, ESP-12, ESP32 programming jig from AliExpress.  They look like this:

    It arrived what I thought was DoA.  I was unable to get any life out of it.  So, as they're cheap, I just ordered another.  A second one arrived also DoA, which then got me thinking, what had happened?

    After extensive checking of the USB-C connector, I eventually realized there were no 5k1 ground pulldown resistors on the USB-C CC1/CC2 lines.  More so, CC1/CC2 weren't even connected beyond their pad.

    Once I knew this, I knew what to do to make the board "work".  By taking a USB-A to USB-C adapter, and building a frakencable to go USB-C to USB-C VIA USB-A I was able to power the board and communicate with it.

    Win.

    So what happened here?  Well as far as I can work out, old designs in Chinese factories are just getting USB-C sockets slapped on rather than microUSB ones.  The result is no regard of the USB-C protocol and its electrical requirements.

    Once you know what's happened, everything falls into place.  I could add CC1/CC2 5k1 ground wires or I could continue to use a frakencable.  But all I can say for sure is, when debugging a problem, it's never what you think.