I got to have my own version of this while I was in university about ... 15 years ago? My family had a small computer configured as a media center, running MythTV and connected to the living room television. It was built of the cheapest parts available on Newegg, and sometimes got cranky enough to need a reboot.

One day my dad was trying to reboot it and it wasn't cooperating. He'd flip the power switch to off, wait a bit, then back to on, and the fans would spin up but nothing else would happen. After seeing this repeat a few times, I gave it a shot -- power switch off, wait a bit, power switch on. It booted up happily.

The trick was that the power supply took a while to fully shut down after losing mains power. If it regained power too early, it wouldn't start back up again. The only people in the house who could reliably reboot that machine had ears young enough to hear the high-pitched whine of the capacitors, to judge whether the reboot had finished and power could be restored.

In college, I was frequently invited to movie nights in one of the lounges because the TV there was ancient and I was one of the few people who could reliably turn it on. It involved hitting the reset button at a specific point of the warm up for a specific duration. Apparently no one thought to just ask for a new TV until years later.

I had a computer in the pentium 3 days that eventually would not turn on while upright. Laying the case flat (as you would when opening it up to root around looking for the problem) would allow it to turn on again. Once on it could be set upright and keep running till next boot up.

This was only discovered after several months of intermittent frustrating debugging.

I can see why you'd think of it as a version of this story, but it's also a version of "Tom Knight and the Lisp Machine" (http://www.catb.org/~esr/jargon/html/koans.html#id3141171).

There is a lot a poorly designed hardware which behaves this way. The trick is to wait 10 secs before turning it on again.

There is no such thing as ground. Especially for high speed signals. An unterminated wire to ground is an antenna.

Older machines tried really hard to have a solid ground reference. Busbars. Copper sheets brazed together. Stuff like that. Lots of headaches with potentials between different grounded points. Eventually, everything with any significant length went differential, so ground potential didn't matter so much. More electronics, fewer headaches.

I know it's not exactly the same thing, but the elecric feeling I get from my Macbook (I think it's some kind of capacitive coupling, some people can't feel it but some can), and the spark between the magsafe power adapter and the computer when I plug it in are always disconcerting. I have to trust that Apple know what they're doing...

The electric feeling is actually caused by EMI reduction measures: per regulations, a Y2 class capacitor must be connected between live and earth in your laptop's power supply. If you connect your charger to an earthed outlet (like you should), the earth is going to absorb the resulting high-impedance supply. If not, the resulting voltage will be present on your computer's metal chassis, which must be connected to the earth per other regulations on electrical safety. Of course, since the capacitor has a high impedance (reactance, to be precise), no meaningful current can go through it, so the "seemingly mains-voltage" on your chassis isn't dangerous at all (as long as everything is working as intended)

The reason for that is that the adapter is not really grounded, but due to EMC requirements there has to be some AC path to ground. This is done by connecting the output ground to midpoint of input EMC filter and thus the output ground is weakly capacitively coupled to half of the line voltage.

Most higher powered laptop power bricks have three prong plug for exactly this reason.

Yeah I'm using a Macbook Pro power adaptor and an Apple Cinema display. Both have earth pins (UK plug). Somehow their 'grounds' disagree enough to ionize air!?

Are you plugged into the same outlet? Different outlets could be on different circuits which can be enough to cause a ground difference.

Yes, the plugs are next to each other. And a mains tester is happy with the mains output.

Oh! I know about this! It's the "fuzzy" feeling you get when you touch a MBP when it's plugged in!

It only happens when you're using the two-prong "duck bill" on the power adapter itself. If you use a power cord with a three-prong plug instead of the two-prong duck bill, the feeling isn't present.

This is a known issue with ungrounded electronics in general.

Apple Macbooks are particularly bad because the enclosure is metal and the standard plug (the interchangeable part of the charging brick) that is sold in most parts of the world is not grounded.

In my country (in Europe) a Macbook charger is sold with a 'type C', (AKA EuroPlug) which is not grounded. The extended cord that comes with the charger is a 'type F' plug, which is grounded. [0]

For me, this results in always having to lug around the extended cord with the charger, as the non-grounded variant makes my Macbook unpleasant to use.

[0] https://www.worldstandards.eu/electricity/plugs-and-sockets/

In the end, you can always buy a new macbook and that's a plus ;-)

I know this feeling, noticed it since at least the aluminium PowerBook G4 I had once.

I remember reading this a long time ago and feeling the magic...

Now I work in RF and there's no magic at all. It's the expected outcome. The explanation at the bottom of the page is probably (roughly) correct.

Anyone have a year for this? It says 1994 at the bottom but that's just the addendum.

Related:

A Story About ‘Magic' - https://news.ycombinator.com/item?id=27740579 - July 2021 (1 comment)

Hacker Folklore: A Story About ‘Magic' - https://news.ycombinator.com/item?id=21385180 - Oct 2019 (1 comment)

A Story About ‘Magic' (1994) - https://news.ycombinator.com/item?id=9102110 - Feb 2015 (60 comments)

A Story About ‘Magic' (1994) - https://news.ycombinator.com/item?id=7472328 - March 2014 (1 comment)

If I remember correctly, this story appeared in the original Hacker's Dictionary (published 1983). I don't have a copy of that book any more, but someone who does might be able to confirm.

I do have a copy of the 1983 book, and this story is present. Like the online copy linked to here, it also begins with "Some years ago ..."

The original was posted to USENET by Ed Nather (utastro!nather) on May 21, 1983.

The author is Guy Steele (GLS in the story).

Some more stories about more impossible bugs: https://www.reddit.com/r/linux/comments/2p6qy5/4_impossible_...

My PC sits in a place behind my table which isn't easy to reach, and I have my keyboard and mouse connected to the built-in USB switch in my monitor. Because the monitor turns off by itself to save power, there's no quick way to wake the PC from suspend mode (e.g. by pressing a key). However, I've accidentally found that turning off my desk lamp wakes the PC up! So just toggling the lamp and off will do the job. I guess it's somehow because the PC and lamp are connected to the same extension cord. The only inconvenience is that I sometimes forget to turn the lamp off before putting the PC to sleep at night.

It's pretty reliable - at least more reliable than the wireless IKEA light button in the kitchen...

I'd love to see the guts of that switch.

For a lark I once put an AC power supply driven from a bunch of tiny hearing aid batteries into the outer shell of a 9V block battery. A casual DC voltage measurement would yield a big fat zero but a small bulb would light up. Magic!

Hearing it was wired to the ground I thought, “maybe the computer isn’t properly grounded by it’s usual means and this switch fixed the ground”.

Yeah - this was a fix for a disconnected ground somewhere.

That was an awesome read!

I belive you are confusing GLS (Guy Steele, the 'I' in the story) with Eric S. Raymond, the, eh... curator of The Jargon File.

The quote in your comment is from Raymond who is indeed quite racist.

Accepting arguendo that your claims are correct: What do they have to do with the magic / more magic grounding switch?

Assuming this is supposed to be about ESR, this page isn't written by him, since it's signed by someone else…

https://ucr.fbi.gov/crime-in-the-u.s/2019/crime-in-the-u.s.-... https://www.census.gov/quickfacts/fact/table/US/PST045221

White people commit the majority of workplace violence in the US. The Justice dept. has breakdowns of violent crimes by ethnicity if you want to see the details.

Thanks

Edit: I looked it up and the mentioned statistics about overall violent crime is actually true. OTOH cherrypicking only violent crime stats and not mentioning white collar crime stats is racist.

Trotting out a race correlation and not mentioning another correlation (poverty) is also racist.