Problem: board seriously fried. Desired outcome: measures to take in future.

Background: This project has been running for about a week on the 2560, and previously for a few weeks on an UNO clone. Board switch a result of program size/expansion. It sits on a shelf with others of the same ilk. Nothing is touched unless I need to plug in USB cable for update. The Arduino is a clone, connected to DS3231, LCD1602 via I2C Serial Interface board, LCD2004 also via 12C SIB, SD card adapter, HC-12 wireless, 5 leds [normally off, used to flag specific conditions, each in series with 470 ohm resistor. The unit receives data from another Arduino and logs to the SD, updating the two LCD displays.

Steps leading to failure: Plugged in USB cable [no other physical changes, nothing else touched, power jack remained connected]. went to Arduino s/w to make minor changes, would not upload to Arduino. While scratching my head, smelled something burning, then saw smoke coming from the shelf.

Post failure: Board seriously scorched around 5v regulator, can provide photo, but seems unnecessary. Visual wiring inspection revealed no obvious problems. Input to power jack measured at 10.06v. Replaced with new Arduino 2560 clone, matching connection for connection, connected USB but not power port [just in case of ?, seen ancedotal posts of possible conflict; a clone after all]. Uploaded my modified program, and all works fine. Disconnected USB, reconnected power jack and all still fine, 16 hours in. Also searched internet for possible problems related to such a failure, but found nothing to suggest probable cause.

Seeking: Words of wisdom from others far more knowledgeable and experienced than I. Could the dual connection of USB and power jack, be the root cause? Never experienced this with UNO clones.

I do not want to put more boards at risk in the future.

  • Maybe they tried to save a few cents and not include the the opamp and mosfet, that would turn of usb-power if the jack is present. I think what happened, is that the mega was trying to supply 5V back into the USB port, back into your PC. This would happen if the 5V of your PC is below the 5V the regulator is supplying. Just don't plug in both at the same time, seems like a good solution. Other that reverse engineering the clone, and see what's different from the original.
    – Gerben
    Commented Dec 1, 2016 at 19:41
  • You said it yourself: a clone after all...
    – Majenko
    Commented Dec 1, 2016 at 19:42
  • Sometimes the big decoupling caps you put around the regulator have a low voltage limit on clones; I have seen board with a grilled cap. If regulator is also a cheap brand, not up to specs, then you are at risk. The first wise thing to do is to buy genuine boards.
    – jfpoilpret
    Commented Dec 1, 2016 at 21:08
  • 1
    It could be the board, but don't assume it is - could also be a problem like a hot ground or excessive ripple with your power supply or even your computer. Commented Dec 1, 2016 at 21:44
  • I suggest pulling out your multimeter and measure that you don't have a short between 5V & GND and 3V & GND. I also suggest checking these 10 possible causes: ruggedcircuits.com/html/ancp01.html
    – sa_leinad
    Commented Dec 2, 2016 at 0:35

1 Answer 1


They are using the same SOT223 LDO as every other product. These have a very low limit of output current when run from an input voltage such as 12V. If you replace the LDO with a switching regulator, you can eliminate the waste heat regardless of the output current (Except for the 5V to 3.3V secondary LDO which is a constant). The PDF linked below shows a comparison between the LDO and drop in switcher. Not that the switcher performance at 1900mA is equivalent to the LDO at 250mA... at 250mA the switcher temperature rise would only be 8.5C.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.