I remember sometimes getting a bright flash on the screen, after which it took about a second to go back to its normal geometry, if an ancient CURSOR magazine p

I remember sometimes getting a bright flash on the screen, after which it took about a second to go back to its normal geometry, if an ancient CURSOR magazine program briefly set the "killer bit" to be able to run some code quickly. On those ancient first generation PETs, the killer poke just caused video noise if left on, by removing the "avoid writing screen memory while the display hardware is reading it" lockout. On middle generation PETs - the dynamic RAM, pre-6845, small screen ones, the POKE did nothing. It was the later, big screen PETs - both 40 and 80 column - that were vulnerable. It seems the POKE - which set an input pin to output - actually prevented the vertical deflection from operating. I can see that frying something if left on for a while. This is all from memory so some of the details may be wrong. It was known, but not considered a big deal. But were there a lot of consumer sales of PETs anyway? They seemed to be more of an institutional (school) machine. Consumers really got into it with the VIC20 and of course the C64 later.