Zeners are pretty useless as overvoltage devices. They have enormous capacitance if they have any power handling capability, and this makes them useless as protection on high speed signal lines. What’s more, a USB-C can provide 3A, so in the worst case you would need a 15W Zener to hold the voltage line down. In the space available, you would be lucky to get a 250mW Zener, which means you would either melt the solder or burn out the Zener.
I spent some years working part time in surge, EMP and lightning protection, and it is hard stuff. Basically there are no good solutions which are small, fast and economical, which is why the world has moved to keeping everything over 20V to the mains adaptor, and is trying to get down to 5V for everything.
It probably doesn’t help that the USB-C port is intended to be used to charge the laptop, so the amount of power you are dealing with is potentially fairly high, and even if the port successfully takes a bullet and saves the rest of the system, you’ve still lost a pretty significant chunk of what makes the computer functional.
The Chromebook Pixel has 2 ports, so it could have limped along had the incident not damaged other components; but not all systems have that many.
Developing the occasional bad USB port is far less serious when it’s not supposed to be the one-port-to-rule-them-all.
I’ve heard that Dell pioneered that practice–at least for its home machines
Oh, I’m well aware. The newer variation on this is for the Chinese sellers to put stock in the country they are selling, and not registering for Tax or running a local business. When the stock is imported the stock value is under declared to reduce their import duty.
The questionable goods are then sold direct from fulfilment centres, the funds go straight back to the sellers home country, or some funnelled through a local company as a front. See http://www.vatfraud.org for more on this.
Don’t get me started on how eBay treat their small business sellers!
Yup. The cable had killed the USB controller, a chip that also managed keyboard initialisation and reading temperature sensors.
Which in turn meant that the Chrome OS’s Verified Boot tech could no longer verify the embedded controller, meaning that the controller might’ve been compromised in some way. And so it would only boot into recovery mode.
During the Y2K run up in which IT so successfully averted disaster that many people thought the problems were overrated - still I think our finest hour as an industry - I heard of a company that had several hundred Dells of the same model. The first one tested had a Y2K real time clock but IT were suspicious and checked a few more. They quickly found that there were at least two RTC variants in the batch, and a number of them were not Y2K compliant.
That assumes you have simply switched power polarity. If the connected circuit has a battery, it is possible the voltage across the diode may get high enough to latch the ICs into S/C mode from the battery supply, whereupon the protection components won’t save you. A large Schottky diode may give good enough protection but its i2t has to be sufficiently bigger than that of the fuse to ensure the fuse blows rather than the Schottky melts. Like I say elsewhere, this is surprisingly difficult stuff.
Gigabit ethernet uses coupling magnetics for isolation. Fiber optics can run to tens of gigabits in a single-wavelength version. Should be therefore possible.
A wonder what some adaptive equalization can do.
What about a fuse and an overvoltage-triggered crowbar, e.g. an SCR?
Or 3.3V, or 1.7 or even 0.8…
This is such a terrible idea I have no words to describe it. Power cable is really one of those things that it pays to have a dedicated device for
Probably as bad an idea as the AV sound guys who use regular SJO AC extension cords and Edison connectors as speaker extension cables. Nothing wrong with the SJO - they actually work pretty well - but the Edison connectors is just asking to have your speakers wired into mains current.
I knew some guys who did this. It is also an awful idea.
Sounds fun. How far did the speaker coils fly?
Only as far as the opposite wall…
Yet my quick charge 2.0 phone charger is a move in the opposite direction, using standard USB to charge compatible phones at 9V.
Let’s just say that the USB-IF’s people have been hitting the ‘universal’ aspect of ‘USB’ pretty hard lately; and not always with good results.
Voltage clamping on USB-C’s all out the window then? That’s horrible. Here I had expected some connectors to be wider than others and otherwise distinguished, but someone had to make it all-bump flowable or something. Still blaming both the Pixel and FaultChargeCo for err.
Maxim’s had 40W Zener-with-follower for a while in their discrete ports. Saw one on an industrial controls design, but nary a luxe mobo.
Some speakers can take 120VAC, 60 Hz just fine.
Hmm…I think I’ll just upgrade my entire signal chain to mains level instead of “line level”…
USB isn’t isolated typically so it’s definitely possible to fry your mobo through USB if you’re careless. I don’t see a bad USB cable doing this though unless you somehow created a ground loop to mains power or something.
Fibre optics, yes of course. And you’re right, I have forgotten GbE. 
Perhaps that’s the future of USB: a cable with power and two optical fibres, and that’s it. But we’ll need to find a really flexible fibre first, owing to the people who will insist on wrapping the cable round the PSU.
