Many years ago I was dispatched to a Lennox GCS series unit that was down. I found the main breaker tripped on a 480V 3 phase unit, and without an obvious good reason. So I was following some kind of intuitive sequence as I often do and I had the system running with the blower panel off as I was following something down. The conduit entrance to this machine came through the blower section as often they do. So what it comes down to is an intermittent Phase to Ground fault in that conduit run, near enough to the RTU to see a flash and hear it. Its an intermittent arc at potential that is not dead in contact with ground. Sometimes it would not trip the breaker.
I then had to " show " the attending electrician the fault because he did not " get " why he was being asked to replace the wiring back to the nearest junction which in this case may have been the entire run. After he saw the demonstration he understood I knew what I was talking about.
Around the same era I was in a tight crawlspace repairing a leak on top of a walk in cooler by flashlight. When I cut the suction line I got lit up with electricity. I broke out my meter and found about 90V between the now separated suction line. I was " hot " to say the least. I seem to recall asking the boss to come by the job and file through circuit breakers while I remained reading the potential that had got my attention. It turned out to be the box lighting circuit. Now start to imagine how everything was grounded with this system, or not. As far as wiring goes I personally never assume it is right. I have corrected major control issues on very large and expensive systems that were not wired right from the get go.