What does Doug have to say about all this? If it's new, then that means there are a bunch of new components in there. Who knows if a resister that cost pennies is changing with it gets hot....
Here are some suggested "living room tests"?
- run your stereo line out > 671 line in, and see how that sounds over time. Put the same song on "repeat" for 2 hours.
When you pull that waveform up on the computer it should be even and cyclical, not creeping up.
- I have some attenuator cables I could loan you... designed for running line levels into the XLRs on an AD-20...
Then you could run line out of your stereo > attenuators > mic XLRs. Repeat above test. This is using 671 XLRs components, but not phantom power yet.
- Set your rig up, mics and all. Blast stereo speakers > mics > rig, again, same boring song and see how the waveform looks.
- Let it all cool off. Find a voltmeter, and run the rig (sound isn't important). Check over time to if the 48VDC phantom power drops off when it warms up.
If the voltage drops off, that would affect your mic's headroom. You can probably test this with a particular resistor instead of real mics... not sure what value.
I suspect at that point you will have narrowed it down a bit. Good luck.