Watching people repair old electronics on Youtube has opened my eyes to the realities of real-world electrical engineering. In short: it’s all about tolerances.
A power supply may have a nominal voltage of 5V, but anything from 4.8 to 5.2 is a-okay. Why? Because your TTL components downstream of that can tolerate that. Components that do 5V logic can define logic zero as anything between 0 and 0.8 volts, and logic one as low as 2 volts. That’s important since the whole voltage rail can fluctuate a lot when devices use more power, or draw power simultaneously. While you can slap capacitors all over the place to smooth that out, there’s still peaks and dips over time.
Meanwhile, some assembly lines have figured out how to aggressively cost-reduce goods by removing whole components from some circuits. Just watch some Big Clive videos. Here, the tendency is to lean heavily into those tolerances and just run parts hot, under/over powered, or just completely outside the published spec because the real-deal can take it (for a while). After all, everything is a resistor if you give it enough voltage, an inductor if the wire’s long enough, a capacitor if the board layout is a mess, and a heatsink if it’s touching the case.
Watching people repair old electronics on Youtube has opened my eyes to the realities of real-world electrical engineering. In short: it’s all about tolerances.
A power supply may have a nominal voltage of 5V, but anything from 4.8 to 5.2 is a-okay. Why? Because your TTL components downstream of that can tolerate that. Components that do 5V logic can define logic zero as anything between 0 and 0.8 volts, and logic one as low as 2 volts. That’s important since the whole voltage rail can fluctuate a lot when devices use more power, or draw power simultaneously. While you can slap capacitors all over the place to smooth that out, there’s still peaks and dips over time.
Meanwhile, some assembly lines have figured out how to aggressively cost-reduce goods by removing whole components from some circuits. Just watch some Big Clive videos. Here, the tendency is to lean heavily into those tolerances and just run parts hot, under/over powered, or just completely outside the published spec because the real-deal can take it (for a while). After all, everything is a resistor if you give it enough voltage, an inductor if the wire’s long enough, a capacitor if the board layout is a mess, and a heatsink if it’s touching the case.
The way I got 100 in a lab once (electrical engineering) was by not using inductances in a frequency filter because their +/- is shit.