Yes, the Weller 'splat test' is always useful when tracking down or trying to generate occasional errors. We keep one of the old irons for this very purpose.
Another interesting source is from the (plastic) wheels on office chairs running on a polished lino floor - if the lights are out you can see the sparks as you move.
Good catch. How well(er) I remember similar soldering irons with the transformer in the light blue stand. Every time we ran bit error tests on T1 6000 foot lines all soldering irons in the vicinity had to be turned off.
One of those irons did turn out to be very useful in a test fixture I built to generate electrical power line transients. This was nothing more than a chattering relay (it's contacts in series with it's coil) to switch the AC line on and off to the soldering iron. Any time this could be shown to cause bit errors in a system under test we knew the shielding and/or power line isolation needed work. It also served to prove to management that bit errors are indeed caused by power line transients getting into the twisted pair and not necessarily through equipment faults, so they did not demand 100% error free tests.
Great story Glenn. This takes me back to the early days of my career when I was testing satellite modems. Those last couple of data points on the BER vs. Eb/No curve took forever to get. We never had issues with EMI from the soldering irons, but we did run into some interesting environmental influences that were only measurable down at the 10^-11 BER level...