Build one yourself!
Two fuses and a zener-diode and a normal diode should do the job. Those parts can limit the current depending on the value of the fuse, the positive voltage to about the zener-voltage and the reverse voltage to ~0.7V.
@Couka: definitly not a bad idea! (especially regarding costs).
But I had to prevent USB's Data+ and Data- Lines as well - especially when overvoltage comes into play.
And I have to cut the USB-Line to insert the fuses & zeners in between.
And all that stuff has to meet the USB specs in the end - I do guess, it's not that easy
(otherwise I would have expected to find this solution in different places in the internet, but I didn't)
That's why I prefer an out-of-the-shelf solution.
Any more suggestions / field experiences?
If you're using an actual Uno, see attached schematic, there is a resettable PTC fuse protecting the USB power line.
P_channel MOSFET T1 FDN340P Intelligent Power and Sensing Technologies | onsemi
should also shutdown if Vin is > 3.3V to isolate +5V from VUSB when Vin is higher then 6.6V. If Vin < 6.6V (resulting in a low on the Gate and thus being on) and +5V goes high, T1 might reverse conduct and allow VUSB on the 16U2 to go high and fry it. The USB section of the 16U2 is separately powered from VUSB, I don't know if a high VUSB could make it thru 16U2's onchip 3.3V regulator and out to the D+/D- lines. You could put a schottky diode in series with the PTC F1, MF-MSMF050-2, to keep that higher voltage out of the USB port, but +5V will then be more like 4.2V when powering from USB due to diode drop and voltage drop across the PTC's 0.15ohm to 1ohm resistance. http://www.bourns.com/docs/Product-Datasheets/mfmsmf.pdf
The D+/D- lines travel mere mm to 22 ohm resistors before traveling mere mm to enter the 16U2, so you'd have to really work at shorting something to the back of the USB connector to them. I've read of 16U2s being damaged from overvoltage on the 5V pin from elsewhere, but not about the USB port being damaged by that.
My Lenovo W530 workstation has shut down the port on me a couple times when externally supplied 5V conflicted with USB 5V (on custom design boards without autovoltage selection), resetting the W530 was all that was needed to recover tho. It reports something like a power surge and shuts down the port.