Found a copy of Dx3 that was being sent to a cable labeled "BCM to Hall C".. in my handwriting. I guess Hall C disconnected this cable, and it was leading to some strange reflections and non-linearity in this BCM signal.
Also found that channel 2 of the Level Translator distributing Dx3 was also causing occasional reflections... re-organized some of the copies to avoid using this channel.
The connector from Dx10 at the down-converter seems a bit flaky. I found that I could get rid of some of the "noise" on the signal if I wiggled the cable a bit.
After injecting some voltage into the BCM v2fs, we assured ourselves that
the re-organization kept the signals in their expected locations.