this week I spent way too many hours debugging a mesh repeater that stopped working after I relocated it across the city.
the story felt familiar:
• moved hardware to a new location
• deleted neighbors
• changed a network prefix
• added a cavity filter
• upgraded firmware
• analyzer no longer showed the node
so naturally I assumed I broke something in software.
I reflashed firmware.
I reset configs.
I removed the filter.
I verified identity keys.
I compared settings with a known good node.
still nothing.
a friend of mine, Jade, said something simple:
“when I move a node and it stops acting normal, I suspect hardware first.”
that annoyed me a little.
because I wanted it to be clever.
but she was right.
on the drive to the new location I had slammed on the brakes and the repeater flew off the seat and hit the floor antenna-first.
I swapped the antenna.
everything came back instantly.
neighbors.
analyzer visibility.
healthy traffic.
thousands of packets flowing.
the firmware had never been the problem.
the lesson was not about LoRa.
it was about variable stacking.
I changed location.
I changed prefix.
I changed firmware.
I added and removed a filter.
I deleted neighbor tables.
when multiple variables change at once, the human brain reaches for the most interesting explanation.
but physics does not care about interesting.
the simplest explanation was physical damage.
and the simplest test solved it.
swap the antenna.
done.
engineers love elegant software explanations.
but sometimes the right answer is:
check the damn hardware.
before you blame the stack.
