About 10K worth of damage.
The unit controller (that's an AD MAX3430 IC, an RS485 interface), the Lenze VFD for the supply fan (through which the fault occured due to the pitiful isolation between the RJ45 jack and the analog output terminal block), and the Danfoss VFD for the DX compressor.
The Danfoss compressor is modular: I was pushing to have them just let me swap the I/O control board...cause the only thing that was wrong with it was the fact that it couldn't send or receive anything on its RS485 bus. Woulda been 1/10th of the cost. But...even though it's designed to be easy to service, the distributor didn't have any means of legitimately sourcing that component.
Insanity. Brand new unit, our controls guys decided to tie into the same 40va 24vac transformer that powered the unit's controls.
Well...that transformer had a common leg that was bonded to ground. Which meant that their controller's DC negative was 12vac to ground.
I found one crispy resistor on the Lenze VFD that tells part of the tale...between the grounded pins of the RJ45 modbus plug and the analog output "common".
What I don't quite understand yet is why it took a little time after they wired it for the fault to become apparent...and how that AC potential ended up making its way over to the RS485 modbus network.
Replaced the lenze VFD yesterday. Gonna poke around that circuit board and see what I can find.