I once spent four nights in a drafty diner in New Jersey trying to diagnose a flickering letter “E” in a neon sign that had been there since the . I was convinced it was the transformer. I had read a bulletin in a trade circular about a specific batch of mid-century transformers failing under high humidity, and since the diner was near the coast, I spent nearly two hundred dollars on a replacement that I didn’t really need.
I replaced the housing, I re-soldered the electrodes, I checked the gas pressure in the tubing, I wiped the dust off the glass supports, and I waited for the hum to stabilize. The light did not return. I had fixed the wrong problem because the problem I had read about was the only one I was prepared to see.
It was a mistake of recency. I had allowed a single piece of external data to override the physical reality of the circuit in front of me. Later that night, I walked into the supply room to grab a voltmeter, stood there among the crates of argon and mercury, and completely forgot why I had entered the room. My mind was so cluttered with the “transformer failure” narrative that I had lost the thread of my own diagnostic process. This
