Hacker News Reader

Comment by jcrawfordor | Hacker News Reader

Couldn't all of the protections on current delivery be implemented in the interface on the other end of the cable, though? This is the approach taken by basically every other standard for negotiated power delivery. Even things like 802.3af which manage to deliver 48v at nearly an amp and with the complicating case of many valid devices not really tolerating any current delivered.

View on HN

The other end being USB-A power adapters? There’s a vast number of those of varying quality. Some have zero logic and just output 5V/1A. No way they could be trusted for this purpose.

call this a “safety in depth” approach

No I think the GP meant the port on the iPhone as the “other end”. In other words, why not have all of the protection logic built into the phone?

just as an example, imagine you end up putting a fuse in the cable. If the fuse blows, then the cable can be easily replaced. A chip in the iPhone.... not so much.

Though likely there's something in the phone _as well_ as the cable.

If the cable is dumb, it can’t switch the +5v power coming from a USB supply. If it can switch it, it’s not dumb.

So we should buy expensive cables and cheap power supplies?

Those cheap power supplies are ubiquitous, so it's more about Apple building their cables to exist in the built environment than dealing with the fallout of consumers busting up their expensive phones by charging them on whatever's lying around.

Guarantee there was a consumer study where they compared reactions between these scenario.

The manufacturers knows that if it’s their product in contact when a user is shocked, they will get some blame even if it’s “really” the sketchy power supply’s fault. And since some cables have these safety features already, anyone making cables without them will appear extra negligent when bad things happen.

The world is unfortunately not very efficient.

There was already a huge installed base of cheap power supplies

The other end of the cable is generic USB and can often be a naked, logic-free 5V supply.

I’ve cut the ends off way too many USB cables just to expose those red and black wires.

And surprisingly many cables even have just those two wires inside of them and can't carry any data

Isn't direct current component ignored by default just because Ethernet devices are galvanically isolated? I thought even if you applied some DC voltage to Ethernet without proper PoE negotation, nothing would happen at all.

Reply on news.ycombinator.com