That DOES present a problem - LOL.
I found that one does NOT want to power the Nano using the Vin pin if you are using an external supply greater than 12VDC! The on board LM1117 regulator can't take that with the load of the processor and the I/O on it. Smoke, fire and flames in the regulator - overvoltaged the processor on the Nano, put more than 5 volts onto the I2C bus so it wiped out the Si5251 Clock chip as well.
I got a real education in replacing the 5351 - LOL. I have a cheap hot air rework station and it did a nice job removing the dead chip. I have been playing with the idea of possibly breadboarding a Raduino and have some carrier PC boards for the 5351's. Still have to solder the chips on the boards though. I bough a handful of 5351A's from Digi-Key for 91 cents US apiece.
The way to go was Chip Quik's "Tack" flux - not cheap but 10ml for $20 bucks will last a LOOOONG time. Dab of it on each pad, position the chip carefully over it's pads and using a needle point tip on my soldering iron, tacked each pin to the board. I made up a probe for my DVM using a small sewing needle and was able to confirm continuity and no shorts on all the pins. Programmed a new Nano and replaced the one that burned up (put machined pins and sockets on things this time). Everything fired up and ran just fine. Just after I built the test fixture, I read the frequencies of the clocks on my nicely calibrated uBITX and wrote them down. With the test set, I can go into Factory Alignment mode and with a freq counter attached to the Clock 2 output, I can actually set the frequency accurately and save it. The BFO can be read on clock 1 and I had those figures too so setting it to the right frequency was a breeze. After the repair, I re-installed it in the uBITX, tuned to WWV at 5.000.00 MHz (the only one I can hear in Kansas during the daytime) and I was only off by 150 Hz. Sure saved a lot of time being able to pre-calibrate the clocks outside of the radio!
The only drawback is - the clocks have to be calibrated to the specific radio for accuracy, due to the variations in crystal filters. You can, however get very close so if I had to replace a Nano for someone, I could pre-calibrate the clocks to close proximity of what they should be and the user could fine tune them in his/her rig.
Jim - W0EB