3 ms·
Uh... How was this not obvious from looking at it? Every geek who's put together or taken apart a computer knows the metal frame is grounded, and most geeks hav
by peterwwillis 12y ago
Uh... How was this not obvious from looking at it? Every geek who's put together or taken apart a computer knows the metal frame is grounded, and most geeks have enough basic understanding of electricity to understand you only need one wire for a ground. The switch enables or disables grounding. Since that is literally the only thing it can do, that is what it does.
- myth_buster 12y agoI think we are talking late '60s here. Perhaps we are better off now with accessible electronics.
- kbob 12y agoI agree that it's great to be able to get a SBC for the price of lunch, and that just wasn't possible in the 1970s. But the AI Lab PDP-10s were extremely accessible. I walked right up to them several times, and I was never affiliated with the Lab. (I was an undergrad.) The Lab wasn't exactly open to the public, but they didn't actually keep anyone out either. The PDP-10 itself was mostly discrete logic, so you could see the components and their interconnections just by looking at them. In addition, DEC published full schematics for most of their products. (I never saw the PDP-10 schematics, but it would be unusual if they weren't available.) If that wasn't enough, Alan Kotok, the architect, came by TMRC every couple of weeks and was more than happy to expound on any part of the system. https://en.wikipedia.org/wiki/Alan_Kotok https://en.wikipedia.org/wiki/Alan_Kotok https://en.wikipedia.org/wiki/Tech_Model_Railroad_Club https://en.wikipedia.org/wiki/Tech_Model_Railroad_Club
- snowwindwaves 12y agomost switches don't have one terminal internally connected to the switch body, which then may or may not be connected to ground (or some other potential) depending on what it is mounted on.
- tempestn 12y agoYou're close, but not quite there. The ground plane of the machine was almost certainly grounded either way. But this switch likely introduces a connection to a separate ground, resulting in a ground loop. When you disconnect that loop it's not surprising at all that you would get a bit of a voltage spike in the ground plane, hence the crash. (I was a power electronics engineer a few years ago, before getting into the web.)
- parfe 12y agoHaha HN, how did this post get downvotes? The second I read about "a switch needing two wires" I thought "Unless it's grounded." Interesting story, but hardly "magic"
- mcphage 12y agoIs it common for one of the switch terminals to be electrically connected to the switch's container? That seems like a dangerous idea?
- parfe 12y agoCommon? Hell no! Obvious? Yes. For a switch to work a circuit must be present. Either a neutral line or a ground it is the only way to send information. If a functioning switch has only a single wire feeding into it then it must be reaching ground through other means.
- freyr 12y ago> If a functioning switch has only a single wire feeding into it then it must be reaching ground False dichotomy. You omitted a third option: magic.
- kbart 12y agoI suspected ground potential differences halfway through the article too, but I guess for web "hackers" out there it's equal to magic.