News for hackers. Stuff that matters.

A comment on USB Type-C is still a mess. Parent comment

Re:USB Type-C is still a mess by vertexFarm on Monday June 11, 2018 @01:11PM (#17286273)
Forgive me for sounding extremely ignorant. I know this is not my area of specialty and I admit I'm about to sound stupid. I want to admit this right off the bat.

Can't you just hook up the VCC and ground from any USB charging IC to a type-c socket? It's all +5v right? I don't understand why the type of connector dictates how fast a cable should be useful for charging. That seems like it should be up to the simple gauge and length of the wire and thus how much current it can handle.

From what I've heard re: that chart in OP's link that shows the different fast charge rates for different manufacturer's wall warts and cables, most of that is caused by proprietary ICs doing sneaky things like throttling the charge rate when it discovers it isn't connected to an official Huawei cable or whatever through some resistor value going to ground from one of the data lines, or some similar method. Couldn't the IC in the device read the resistance / inductance of the cable and determine through that how much current it could handle? I feel like this is all bad faith shenanigans through proprietary cable DRM. When charging is concerned, a conductor is a conductor assuming there's enough thickness to prevent overheating, right?

[ Parent | on HN ]
Re:USB Type-C is still a mess by mikeash on Monday June 11, 2018 @01:24PM (#17286402)
You can’t make that assumption at the end. These cables are as thin as they can possibly be to carry the maximum current they’re rated for. That maximum current could be 0.5A or it could be 5A. Try to pull 5A through a cable rated for 0.5A and you’re likely to burn the customer’s house down. You can’t really measure the resistance of the cable from one end. You can try to guess by watching for a voltage drop as you pull more current, but that’s extremely error prone. Because of this, all of the components are supposed to signal what they can handle, and if they don’t, they fall back to the minimum current known to work.
[ Parent | on HN ]
Re:USB Type-C is still a mess by thfuran on Monday June 11, 2018 @01:32PM (#17286476)
Wait, so a USB-C cable cannot be a completely passive component?
[ Parent | on HN ]
Re:USB Type-C is still a mess by digi_owl on Monday June 11, 2018 @02:01PM (#17286725)
Yes and no. There is no IC in there, but there are small segment in the cable spec that talks about resistors on one of the lines that signal the type of cable available.

So you put one resistor in for a low rated C to C cable, another for a higher C to C rated, and yet another for an A to C converter (indicating that the C end should fall back to the older battery charging spec).

Guess how many fly by night cable manufacturers get those wrong...

the C to C spec charging, or Power Delivery, spec is a whole different barrel of stinking fish. Where you can have everything from 0.5A at 5V to multiple As at 20V running over the same C to C cable.

This is akin to the Qualcomm quickcharge scheme that run 9V over existing micro-USB cables and ports (thus allowing higher wattage without cranking up the amps).

[ Parent | on HN ]
Re:USB Type-C is still a mess by kevin_thibedeau on Monday June 11, 2018 @04:29PM (#17288260)
Power Delivery is not just for USB-C. It was defined for use with USB 2.0 and up on any of the legacy connectors.
[ Parent | on HN ]
Re:USB Type-C is still a mess by digi_owl on Monday June 11, 2018 @05:15PM (#17288661)
While true on paper, i wonder if it will ever be implemented as such. And frankly it may well be yet another black mark on the mess that is USB "C" (or 3.1 or whatever).
[ Parent | on HN ]
Re:USB Type-C is still a mess by barbegal on Monday June 11, 2018 @01:53PM (#17286672)
The resistor in the USB-C cable to identify its current carrying capacity is still a passive component. And of course you can omit this resistor but your device will fall back to the minimum current.
[ Parent | on HN ]
Re:USB Type-C is still a mess by andromeduck on Monday June 11, 2018 @01:56PM (#17286690)
They can be passive - thus the resistor to incidicate the cable's rating.
[ Parent | on HN ]
Re:USB Type-C is still a mess by theatrus2 on Monday June 11, 2018 @01:20PM (#17286372)
Type-C offers charge rates up to 100W, which means the voltage is ramped to 20V.

There is an entire protocol to negotiate power delivery (including who charges who) http://www.usb.org/developers/powerdelivery/

[ Parent | on HN ]
Re:USB Type-C is still a mess by digi_owl on Monday June 11, 2018 @02:02PM (#17286738)
And the latter part have caught some people unaware, where a battery pack and a phone will ping pong who is charging who until both are dead.
[ Parent | on HN ]
Re:USB Type-C is still a mess by chimeracoder on Monday June 11, 2018 @02:35PM (#17287051)
> And the latter part have caught some people unaware, where a battery pack and a phone will ping pong who is charging who until both are dead.

Or, if you're really unlucky, it'll attempt to charge your wall. No, I'm not joking: https://twitter.com/chimeracoder/status/941820122406064128

[ Parent | on HN ]
Re:USB Type-C is still a mess by organsnyder on Monday June 11, 2018 @02:58PM (#17287336)
My phone has offered that option when I plug into the car. While the car could in theory benefit from the charge (it's a plug-in hybrid), I doubt the phone would be able to provide much of value.
[ Parent | on HN ]
Re:USB Type-C is still a mess by Johnny555 on Monday June 11, 2018 @04:43PM (#17288393)
Sounds like a good way to steal power from work -- charge up your phone at the office and supply power to your wall outlet when you come home.

:-)

[ Parent | on HN ]
Re:USB Type-C is still a mess by digi_owl on Monday June 11, 2018 @02:55PM (#17287301)
Ok, that's it. Murphy does not need to verified any more. And i need a really really stiff drink.
[ Parent | on HN ]
Re:USB Type-C is still a mess by djsumdog on Monday June 11, 2018 @03:30PM (#17287694)
That is hilarious. At least it gives you the option.
[ Parent | on HN ]
Re:USB Type-C is still a mess by vertexFarm on Tuesday June 12, 2018 @09:06PM (#17299449)
Gotcha, I was wondering if the VBUS was dynamic. Thanks for clearing that up for me and linking to the documentation, I appreciate that.
[ Parent | on HN ]
Re:USB Type-C is still a mess by tadfisher on Monday June 11, 2018 @01:24PM (#17286399)
That's a tiny part of the USB 3.1 Power Delivery spec. The bulk of the issue lies in the negotiation between sink and source, which is fraught with complications.

PDF link: http://www.ti.com/lit/an/slva842/slva842.pdf

[ Parent | on HN ]
Re:USB Type-C is still a mess by Karrot_Kream on Monday June 11, 2018 @01:21PM (#17286378)
There's usually a lot more involved in charging a device, not to mention that USB-C can charge at 20V instead of just 5V. Charging speed is often based on current, but you also want to make sure you don't discharge too much current too quickly for various reasons (such as the battery capacity, and to prevent surges, etc). If you're curious, there are teardowns that go over chargers that discuss everything involved.
[ Parent | on HN ]