Hacker News Reader

Comment by Animats | Hacker News Reader

The USB-C spec is just terrifying. All that functionality. All those pins in a tiny space. High-voltage high-current power, in either direction. HDMI. Analog audio. Even supports USB. Either end can be the master. There's even a protocol to support a tiny display on the cable to tell you what's going on or what's wrong. And some more features I missed. Yes, the mobile handset industry wants to get down to one connector, but they may have overloaded this one.

Nobody is going to get 100% of this right for some time. Coming soon: lots of questions on Stack Overflow that begin "When I plug a USB-C cable between my X and my Y, ..."

View on HN

Let's also not forget that pure digital USB audio has been around for a long time, and works with any USB connector:

http://www.usb.org/developers/docs/devclass_docs/

The "Audio Class" was standardised in 1998. Many "USB soundcards" have appeared since then, and if the manufacturers really wanted to, they could've just used that standard instead of overcomplicating things by making yet another one.

That's the case already with USB Power Delivery Standard: The dream of "One Wall adapter and one car adapter" to charge all your devices (phones, laptops, battery driven accessories).

Actual situation: not really.

Examples: Apple's MacBook 12 charger delivers non-standard 14.5V and the Dell XPS 13 only accepts 19V. USB PD however allows 5-12-1USB and on top of that the old USB Battery Charging standard with 5V@2A.

Additionaly USB-C allows methods to authenticate peer devices. We will for sure get DRM'ed chargers and whitelisting for some accessories...

IIRC the Apple adapters are variable voltage and will increase up to "18.5V" depending on the load current, but the laptop itself is perfectly happy to receive full voltage --- and slightly more --- all the time.

In fact, laptops that have a 19V adapter will be fine with 20V too, and may actually run slightly cooler because of the lower current draw. I suspect all the funny voltages is just to create differentiation and a sense of lock-in, but the underlying engineering and tolerances on components mean that a 19V adapter may actually be in the range of 18-21V depending on load. It's amusing to find laptop adapters claiming 18.25, 18.75, 19.6, and a bunch of other "why!?" voltages which are technically all in the same tolerance range.

I've written about this before too: https://news.ycombinator.com/item?id=7594383

Some further discussion on MacBook-PSU-related details here: http://www.eevblog.com/forum/repair/replacing-apple-magsafe-...

http://www.eevblog.com/forum/projects/apple-magsafe-power-br... (last post is enlightening)

The real obstacle with the "one adapter for all laptops" is the proprietary DRM-ish ID schemes that a lot of them have, where they may complain/warnin and/or won't charge the battery with just pure voltage input.

Couldn't this DRM-esque stuff be a more innocent reflection of different manufacturers having different protection circuitry on their batteries?

Batteries are behind the charge controller which does have its own way of communicating with and authenticating the battery pack.

The term "laptop charger" is a misnomer; it's just a power supply without any battery charging circuitry.

> Examples: Apple's MacBook 12 charger delivers non-standard 14.5V and the Dell XPS 13 only accepts 19V.

Actually Profile 5 allows anything from 5 to 20 volts so they are both compliant.

I actually wouldn't mind a little white listing.

Folks like Underwriters Laboratory could serve as CAs, issuing certificates for compatibility with different profiles / standards. Then chargers could refuse to throw current at unproven devices.

Of course then you need the ability to boot your charger into some kind of "promiscuous" mode for development or backwards compatibility purposes.

To get REALLY ambitious, your charger could have an Internet connection and monitor a blockchain for certificate information. Some careful protocol design could let users override the DRM at will without letting cheap chinese crap factories ship exploding devices.

> your charger could have an Internet connection

Quoted out of context

> To get REALLY ambitious

Honestly we already have standards for tethering over USB... Still crazy, but it could actually work.

the day when chargers monitor a block chain for certificate information is the day I retire

Yes, but why worry, because adding another new subclass to USB isn't making the situation significantly worse than it already is. Stack Overflow is already full of USB questions.

Because there is already so much confusion about USB 1, 2, 3, 3.1, low speed, full speed, high speed, super speed, super speed+, USB A, micro-B, micro-AB, mini-A, mini-AB, mini-B, USB PD (rev 1 & 2), OTG, MHL, Samsung's 11-pin micro, and even non-USB lightning....

Can it get worse?

> Samsung's 11-pin micro

That’s actually USB 3.0 micro super speed+

All of those however top out at 5V 2.5A.

With this new power spec, that is actually independent of the C plug spec, but unveiled at the same time, can go anywhere from 5V to 20V depending on negotiations between devices.

On top of that the C plug and the data speeds are also interchangeable. So you can get a C plug and 2 speed, or you can get an A plug and 3.1 speed.

It probably could work of one assumes two equally smart devices on both ends, but USB always included shortcuts to allow the cheapest and flimsiest implementation (add a resistor between two pins) by garage shops building Phone chargers for 30¢/piece. And for them, its easy to mess around randomly, until their nonstandard non compliant junk means who-knows-what to a USB-C device.

So, your $300 headphones will pretty certainly work with your $800 iPhone, but plugging them into that $3 (including shipping) charger?

Actually that's a good point. Will we end up needing drivers for headsets if they add any non standard features?

I can't help thinking planned obsolecense when I see these moves.

> There's even a protocol to support a tiny display on the cable to tell you what's going on or what's wrong.

You mean, like the light on the plug on the "computer" end of the Apple magsafe charger cable? Or something more complicated?

Indeed. Rather than have the power spec etc demand a C plug, you can mix the power spec with old A and B plugs. You can even end up with C plugs that only deliver USB2 speeds.

Never mind that C devices actually have to negotiate what side is up, rather than sort it out mechanically.

You mean, nobody will manage to do it exactly as in Lightning?

It is already being done.

Lightning is none of those things.

Reply on news.ycombinator.com