In the Field of EMC / EMI and RF, we call this Black Magic!.
The problem you are describing, sounds like Capacitive coupling caused by variation in the Cable Dielectric due to temperature / moisture variation.
I have not experienced this, but this is what I would be trying:-
1. Does the pixel have an In-Series Termination Resistor inside. Usually 75R. The Sender should also be fitted with 75R (or ish). in Series with the data line.
2. What happens if you put a Null Pixel in the Controller.
3. Someone mentioned hand capacitance stopped it, in this case, you might want to try adding a small Value Cap between Gnd and Data. Some thing like 47pF.
4, OR possibly add 1M Resistor across Gnd and Data, at the Pixel End.
5. If the data signal is reflecting, you could also try adding or removing a couple feet of cable if possible. This will change the Matching / characteristics of the wire.
6. One final test, tear the +v Strand away, Twist the the Gnd and Data and then lightly twist +V in the reverse direction.
7. I am questing that no one who has the issues had a scope. Although, just attaching the scope probe makes these problems go away.
Would be interested in the out come?
I did some more experimenting tonight. I found a few turns through ferrite on both ends of the CAT5 going into and out of my E682s helped the most. For the few strings that still had flicker, adjustments in the orientation and length of the cables to them resolved the flickering (some may have been moisture in connections, so redoing the connections fixed this). It was -10F last night and no noticeable flickering.
