I have an XDRduino Mega clone with ATmega2560 and ATmega16U2 USB support chip. I also have a Sensor Shield (same vendor), which provides screw terminals to Vin and Ground.

I connected the board with USB, loaded "Blink", and then changed the time interval. Next I connected a 20x4 LCD using the Sensor Shield and a breadboard and loaded LCD "Hello world". All seemed fine with the new board.

Each time I powered on or reset the Mega, the TX and RX LEDs flashed a couple of times and then my sketch started running.

Then I disconnected it from the PC (no USB for power anymore) and connected a (working from another project) S-60-12 power supply, DC Output 12V 5A to the shield screw terminals. I verified polarity -V to GND, +V to VCC.

OK, I swear I hooked it up correctly. But if you tell me that hooking it up backwards would explain all the symptoms below, then yes I'll believe you that I screwed up {sob}.

(And yes, you're right, there is no diode protection on Vin, just on the dc power jack. Dummy me. Double dummy me, as there was a screw type power plug laying unused on the bench {groan}.)

About 4 seconds later, I heard a snap and immediately pulled the plug on the power supply. Yes, something probably smoked, but let's refine at bit, please.

I moved the Mega to my PC, (disconnecting the external power supply so to use only USB for power). It is no longer recognized as a USB device, and the Arduino IDE (1.6.5) no longer finds the USB serial port.

Voltages at 3.3V pin and 5V pin are correct. I've also probed the regulator and comparator and all looks well for the power circuitry. F1 is not blown (or there would be no power as USB is the only source).

Item: The TX and RX leads are on solid. I expected them to flash then turn off. (Power LED is on solid, which is normal.)

Item: The reset pin for the Mega reads 1.41 V. It's connected through a 10K pullup to +5 V (reads 4.63 at USB pin, 4.53 at the 5V pin). This seems way too low.

Item: The reset pin for the 16U2 reads V+ (4.35) and is has a 10K pull-up.

Item: AREF is 50 mV. I believe it should be 1.1 V or 2.56 V.

Question: Is it reasonable to think the 2560 may be dead, and the LEDs are on because the 16U2 can't talk to it? Any way to test this theory?

(If the 16U2 were bad, it would still be possible to program the 2560 via the ICSP header, thus salvaging something...)

Question: The 16U2 won't do the USB connection dance with the PC because the 2560 is dead? Should the 16U2 to try to connect no matter what the state of the 2560?.

Question: If the 16U2 is dead, should TX and RX LEDs be lit?

Question: If I had reversed the hookup, should not the regulator (P1117 chip) have saved me?

Question: Has anyone else had a failing component (other than the 2560 & 16U2) with these symptoms?

  • Are you sure the shield takes 12V? You didn't supply a link to its circuit diagram. I suspect it is intended for 5V.
    – jippie
    Nov 8, 2015 at 15:57
  • The shield is just a giant break-out box. It contains no active and only two passive components (LEDs and their current resistors) replicating those on the Arduino itself (because installing the shield hides them). All the rest is header pins and edge connectors and PCB traces. For example, the power terminals I used to connect the supply just feed back to Vin and Gnd. Nov 9, 2015 at 15:40
  • @jippie I lied. There is also a push button switch replicating the reset button on the base Audrino because the shield covers the base. BTW, the shield was "clean". I'd removed all the leads to my breadboard, so only the pins were showing. And there wasn't a paper clip or wire shard visible (to me) that could have shorted the pins. The shield still works, to the extent that the power led lights (on both the shield and the Arduino itself). Link to shield board. Nov 9, 2015 at 15:55
  • How did you verify the power strip is connected to V(IN)?
    – jippie
    Nov 9, 2015 at 16:10
  • Excellent question. I trusted the internet instead of metering it out. DummyPoints++. Taking meter in hand, power goes straight to +5v. It doesn't go to Vin. Much is explained!!! Nov 11, 2015 at 3:47

2 Answers 2


I want to make this obvious to the casual reader: Mega Sensor Shields -- all versions I've data on -- run the power connector to +5V.

The power connector is not power in, as defined by the base board. In fact, there is no direct way to get to the base board regulator from the shield. (A pin exists for Vin, but it isn't connected to anything on the shield. It's mechanical support.)

It's intended to supply more current to devices connected via the shield than the base board does.

There is a jumper that disconnects the sensor shield +5 rail from base board power to allow independent +5 volts distribution to the shield attached devices.


If what i read is correct, then you've probably cooked your duino.

As for what exactly failed, i'm going to guess from your description, that one or both of the capacitors failed. Usually when they blow, you'll hear a loud snap like you describe.

I would suggest testing those, and replacing them if they are faulty.. This is no guarantee that other components weren't damaged, but it's a good place to start.

As for why the LED's may still light up, that could be any number of reasons really, and not a particularly good indicator of what exactly failed.

  • You're right, it's cooked. Both chips got 12v right through Vdd/Vcc, cause I didn't verify that "power" meant Vin. LEDs must be on because the chip pins/paths "melted" to ground. Nov 11, 2015 at 3:51

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.