I'm pretty sure I have burned 2 Nano Every boards when connecting them both to USB and an external power supply but I'm not sure why.
The conditions where a little different each time but the outcome was the same: it powers on for a couple of seconds then the ATMega4809 gets excruciatingly hot (without burning or smoke) and stops working. After unplugging everything and reconnecting to the board, with USB only, the Nano get recognized (I can pull board info) but again the chip gets extremely hot so I unplug to avoid further damage. Also the external switching converter (LMR50410) failed both times.
I'm saying the two cases are a little different because the first time VIN was provided by a 9V wall adapter while the second time the source was a 12V automotive battery. I think that both times USB was connected first and afterwards the external power source was started, however I'm not sure 100%.
The following schematic shows the auxiliary board layout, the only missing part would be the wiring to from the battery to the connector J_OBD (shouldn't matter as the LED lit up, confirming correct polarity).
Looking at the schematic the Arduino and buck converter are connected only through VIN, my understanding is that USB power gets automatically disabled once an external VIN is detected. Even if that not the case it shouldn't be an issue for the onboard regulator which is downstream of both connections.
Nano worked correctly through USB only before the event.
No relevant code was running at the time of the incident, even though it was different each time.
What am I missing? Is the failure caused by something not included/described above?
Since the ATMega is heating up but not the onboard voltage regulator I would think it an "current overdrawing" issue but it doesn't make much sense.
The USB 5V passes through the onboard regulator to Vin. The regulator will be damaged (short) if you have a big load on Vin.
The 9-12V applied later on Vin blows the 4809.
Only solution is to add a diode to Vin which blocks the current coming out of it and still allows it to be used as power input.
I guessed as much, thank you for the explanation.
So the regulator fails and allows more current to pass through than what it's meant to? I would have guessed that regulators would be designed so that in case of failure they cut off all outputs, this doesn't seem to be the case.
Very interesting.
More likely designed for best figures. Any element in the main current path will cause a higher voltage drop.
This means the characteristics for efficiency will drop. As the maximum current is a heat-problem, that will drop too, while the minimum drop-out voltage will increase.
The chipmaker knows that a board-designer will compare those figures to other chips, so…
When powered by USB there is no power on Vin and the regulator is out of control. But the return path can be seen in the above diagram. From OUT, coil, parasitic diode of mosfet, RSEN to IN. When that current is to high the parasitic diode shorts allowing the full voltage from Vin to the output.
Arduino knows about this problem and a simple diode on Vin would have prevented this.
I don't see much load on the 12V (Vin) in your schematic and this isn't explained yet:
It looks like your regulator might have shorted out when it received only something like 4V and tried (to) hard to make 5V from this?
I'm not sure about this. My experience tells me that powering via USB also outputs power to VIN (the LED in the buck converter lights up).
That's true, the J_OBD plug connects to the diagnostic port of a motorcycle which provides power and CAN data.
I suspect the Nano was already plugged in to the motorcycle when I tried to connect through USB, the regulator must have tried to provide power for the whole electrical system and burned out (I distinctly remember the license plate light turning on when the USB was plugged in, leaving me quite perplexed). Subsequently when the motorcycle was switched on it went straight through the failing regulator to the ATMega burning that as well.
This remains unexplained. I have previously powered the Nano (and the LMR regulator) by USB to a PC (which from my readings provides around 3.8V) without any issue.
I think I have a similar situation, but just wanted to clarify. I have a Nano Every controlling some 3W NeoPixels. Power in to the Nano Every is 12V, via the Vin. That same 12V is supplying a 12V to 5V converter, to power the Neopixels. All was working fine, until I updated the software on the Nano Every, with the 12V unplugged. Now the ATMega4809 gets very hot, and a board powered by the Nano Every 5V (a Adafruit Audio FX Sound Board) no longer functions.
So, if I've understood correctly, when I updated the Nano Every s/w there was a voltage path out the Vin, which would try to power the 12v to 5V converter & Neopixels, and then bad things happened.
I'm going to put a diode on the Vin, which should prevent this.
The internal voltage regulator on the Every shorted from input to output when powered by USB only because the current from the Vin pin to the Neopixels was too high.
When you then reconnect the 12V to Vin: That 12 V passes straight to the 5V destroying the Every and anything else that is connected to 5V.
