Hacker News Reader

Comment by TeMPOraL | Hacker News Reader

At which point do we change assumptions about safety of USB cables though?

Myself and most people I know always considered USB connection to be safe-ish - that is, you can keep the cable connected on the supply side, and have the receiver end just lie on the desk on the floor, and the worst that could possibly happen is some tiny sparking if the stray end touches something conductive in a very unlucky way. But the more power I see pushed through these cables, the more I start to look at them as live wires hooked to mains power.

Additionally, such wattage sounds like a serious fire hazard if the cable is damaged, which means the cables themselves need to be handled with care. Something that wasn't the case with typical USB charging until recently.

View on HN

I mean, I share that worry, but isn’t the “-PD” part of “USB-PD” a negotiation step?

I have very little worry that my carpet can accidentally negotiate 60w and up.

I _am_ worried about cheap USB-PD devices that forgoe this negotiation as it is complex and expensive to implement.

There's the cheap devices but even more worrisome are cheap, possibly not up-to-spec cables.

If you buy a fake cable that pretends to have display-port alternate mode or something but doesn't, meh, you're out 10 bucks. But if it pretends to be able to carry 200 W but isn't and burns your house down instead, you might be out a tad bit more...

Should be detectable though, at least in many circumstances.

USB-PD has bidirectional communication, so at least in theory both ends could know how the cable is performing by comparing voltage and current measurements at either end.

If the cable drops too many Watts, the load can be disconnected.

Now you will run into WiFi type problems where your cable gives you different charge speeds depending on the season and time of day. Service will degrade and you may not notice the problem and may never figure it out in the end.

Beats burning down the house though... Also in case of laptops, mobiles etc, the device can report the error.

Is this actually a standard, or would that have to be implemented like some sort of "extension"? And if so, wouldn't that break down the "universal" nature of the beast? As in a Dell laptop won't be able to communicate with an HP power brick (unless said extension becomes at-least de-facto standard)?

As I've commented elsewhere, I wonder if, at that point, there still is that much of a point in sticking with USB-C for what seem like fairly specific applications.

I mean, if we're at the point of carting around such a monster of a laptop, will an additional, dedicated, power-only connector make that much of a difference? Don't get me wrong, I'm actually somewhat tempted by such a thing, but just because it would be easy no move around, not for actually carrying on my back all day every day.

It's not part of the current standard[1] as far as I can see.

There are some mitigations though, like the cables must identify what they can handle (else the high current modes can't be used), and both ends needs to verify the plug and cable etc.

In theory though I don't see why they couldn't include this in the standard.

[1]: https://usb.org/document-library/usb-power-delivery

Your carpet cannot negotiate a high power, but two devices might negotiate a high power while not being aware that that the cable has a cut in it and is in direct contact with the carpet.

Yeah that's the problem. I've had a bent usb c (from a pretty good brand) literally melt into my hands. I'm so glad I was using the phone at the time, I don't know what could've happened if I wasn't there. The dent seemed pretty "small" too, and it was way less damaged than a lot of lightning/micro USB cables I've seen and used in the past.

What I don't understand is how the cable doesn't have a way to detect a short circuit. I'd imagine that a 250 watts capable cable would have more safety features hopefully though.

I don't think the non-USB-c cables have any more safety features to prevent that. I'd be surprised if the USB-C ones do.

The original USB standard was 5V 0.5A, which wasn't going to start any fires.

Oh, I realize that I left out important context. I didn't mean relative to other USB cables, but rather relative to other non-standard laptop power cables. This goal is to replace those high power cables, and those did not have significant safeties against damaged wire.

> I _am_ worried about cheap USB-PD devices that forgoe this negotiation as it is complex and expensive to implement.

Yes. Moving critical safety limits into software? https://en.wikipedia.org/wiki/Therac-25 Let's hope they're all up to the challenge ...

How would power negotiation be implemented without software?

>At which point do we change assumptions about safety of USB cables though?

It's a good question. Generally speaking, DC is considered a shock hazard at or above 60V, but OSHA recognizes 50-60V as being potentially hazardous. It's certainly an arc hazard when disconnecting as the article notes. And 5A at just about any voltage will start a fire in case of a short circuit.

I'm not 100% sure, but I believe a USB-C cable connected to a USB-C charger is cold by default until something is connected. Don't try this at home either, but it should also be perfectly safe to plug a USB-C cable into two different wall chargers.

Reply on news.ycombinator.com