Quoted from ForceFlow:Other than a lighting strike, is it even possible to exceed what a 1N4004 is rated to handle?
Considering that US games take 110v from a wall socket, even if something shorted, I can't think of a way for the voltage to get anywhere near 400v.
Inductive kickback from coils and relays is no joke and can easily hit the kilovolt (or tens of kilovolts) range. The diode starts to conduct as the power is cut to the coil, which clamps this down to some mere tens or hundreds of volts for you, and since there's so little current generated by the collapsing field it does just fine. If you ever make the mistake of touching a solenoid or relay when it's firing, even with a diode across it you do get a pretty noticeable kick out of it.
Mostly, though, pinball machines are moving slowly (by electronic standards), and the ramp up and down when a coil or solenoid fires is slow enough that it's easily kept within the 400V rating of the 1N4004. Can be exceeded, especially if a switch goes bad on you and the coil is machine gunning pretty fast. (The faster a coil is fired, especially repetitively, the more kickback you'll see.)
BTW, this is part of the reason for suppression capacitors on switches that tell the MPU to fire a coil. It protects the diode, which is protecting the driver transistor, which is protecting the buffer transistor (by handling current it cannot), which protects the PIA (ditto), which protects the MCU (and so on, and so forth.)