I emailed Jeremy (aka Clever Clogs ..) and here is what he said...
Ok… technical stuff follows.
The way these sub-node boards work is pretty simple (SPI protocol).
The main node sends a binary stream down the MOSI (Master Out, Slave In) line to sub-nodes clocking on SCK.
Data is passed to shift registers on each sub-node (74HC594D) and onto mosfets which drive the LEDs on that board.
Each shift register also passes on data to either the next shift register (more LEDs) or to the next sub-node in the chain via a dual inverter (NC7WZ14) acting as a buffer.
It’s the dual inverter I’ve circled in the node picture you have and is what’s busted.
So, how’d I diagnose this?
It really helps having an electrical engineering background or at least understand what components do and be able to read and understand datasheets.
It's also good (mandatory?) to have the right tools for the job (multimeter, oscilloscope, microscope, hot-air rework station etc).
I was told the LEDs on node 8b were working fine. As this is the first sub-node in the chain that meant node 8 could not be the issue.
Quite thankful for that as node 8 is a way bigger problem to diagnose.
Using a multimeter on continuity mode you can see most of the connector pins on 8b are pass through (RCK, SCK, MISO, 5V, GND) so none of those can be a problem.
That’s leaves MOSI.
Basically I wired up a micro controller to the sub-node, wrote some code and sent data down to the board.
Following the MOSI traces, I tested the appropriate pins on the shift register and dual inverter on my scope.
The signal was good on the inverter input (A1) but wasn’t inverted on the output (Y1).
I also tried to manually drive the inverter by pulling A1 high then low but still no output.
The dual inverter is a 7c (NZD) component!
Replaced it on 8b and 8c and all is well again.
Now the mystery is why this has happened?
Possibly a bad batch of inverters.
Maybe a voltage spike on the 5V rail blew them up.
So there you go! Clear as mud.