Internal diagrams show no backfeed diode (but I remember having seen one many moons ago).But the text talks about "the internal diode between the output and input pin".Output capacitance is not an issue here. Input capacitance is.We are putting 5volt on the output, while the input cap is still discharged."the internal diode between the output and input pins can withstand microsecond surge currents of 10A to 20A"The root of the problem stays the same.The input cap on Uno and Mega boards is 47uF. Less on smaller boards.I think we're safe to assume backfeeding (without extra capacitance connected to Vin) is not problem.Extra capacitance in a supply, connected to the DC socket, is not a problem because of the reverse protection diode.Leo..
Overall, the biggest issue appears to be making sure that 5V isn't connected to the output of the regulator while a voltage is applied to Vin.
Don't get that part.
It is not a problem to parallel two regulators.The problem could be that only one does all the work.The mosfets on the USB line prevent backfeeding into the computer.But they only switch off when Vin is above 6.6volt (see opamp circuit).When USB is just a diode drop higher than the onboard 5volt regulator, USB supply takes over through the reverse protection diodes inside the mosfets. Leo..
Even without the above, the UNO does it all of the time when USB is connected and there is no voltage applied to Vin, and I don't see them dying all over the place from back-feeding the regulator.
Maybe I'm missing something here but I just looked at the UNO schematic and when the USB is connected , with nothing plugged into the external dc barreljack, the USB 5V is connected to the output of the onboard 5V regulator (which has no input). I don't see anything shorting Vin to GND I don't know what you mean by the UNO does it all the time. Can you explain (what you meant ?)
I meant that the UNO connects a voltage to the output of the regulator all the time when USB is powered and that no damage is done.
Yes, I agree. That's what I was describing in the schematic. I guess it was someone else who was saying the regulator would self destruct if you shorted the input pin to GND and connected 5V to the output pin.
I'll add that to my bucket list...