Hacker Newsnew | past | comments | ask | show | jobs | submitlogin
New 100W USB 3.0 Spec Can Charge Laptops (techweekeurope.co.uk)
78 points by pwg on July 24, 2012 | hide | past | favorite | 58 comments


I'd be careful with these; it means you can't plug in an untrusted power supply anymore.

Just imagine one of these cables also acting as a USB keyboard, which brings up a new terminal, adds an SSH authorized key and closes the terminal window without you even seeing it.


Since a true machine charger would have its data ports wired to ground, couldn't you make simple adapter that ensured that was the case? Then security wouldn't be an issue?


And Thunderbolt is already dangerous, given that it's essentially just serialised PCI Express (so you've got a free view into RAM).


But you don't typically plug in random, untrusted TB cables -- you do typically plug in random, untrusted chargers.


Thunderbolt's the primary display connector on Macs as of 2011 (and some PCs); so not as common as chargers but not entirely outlandish.


There is security against that, you can limit access for thunderbolt devices to various parts of memory, same as you could with Firewire. Thunderbolt is not anymore insecure than Firewire for instance.


There is the ability to protect memory from malicious devices with IOMMU, but it is often disabled by default. My experience with DMA attacks via Firewire or Thunderbolt is that they work out of the box.

Inception is a nice tool to play with if you want to try out DMA attacks: http://www.breaknenter.org/projects/inception/

You may also want to check out the Volatility forensics framework: http://code.google.com/p/volatility/


Considering that firewire is a popular attack vector that isn't really comforting...


How often are you currently plugging in untrusted power supplies? ThunderBolt can chain a lot of different device types but I don't think it's a huge attack vector to plug in untrusted monitors.


> I don't think it's a huge attack vector to plug in untrusted monitors.

Yet. Give it a couple of years. The monitors are really smart now - many of them running code and have flash memory. The attack won't be directly by the monitor - the monitor will have been infected by malware earlier ... say, by an infected computer ...

Our devices are getting too smart, and our protocols getting too trusty, for our own sake. (Although I'm sure governments and other law abusers like the RIAA are raising a glass to each of these "advancements")


Personally, I'm borrowing someones magsafe at least once a week. A majority of students at my college use MBPs, so they are abundant.


There's already a hacking platform for that. https://hakshop.myshopify.com/products/usb-rubber-ducky Seems like it would be pretty easy to adapt to a USB power supply.


Hopefully they decide to deal with it in the standard and make data blocking (or power-only if you prefer) cables feasible.


Or worse, find a device driver (any device driver, thanks to plug-and-play) with an 0day and traipse around the kernel.


I have done research on this topic, it is a standard, not ratisfied yet, and no way your old usb cables can take 100watt power(voltage/amperage), also wall sockets need standards, which is not there yet.

there are other working powered USB implementations available http://en.wikipedia.org/wiki/Powered_USB


The example you cited doesn't even use the same connector. Furthermore, the support of Intel and Microsoft is basically a sure-shot that we'll see an implementation. There's skepticism, and then there's just silliness.


This is awesome, being able to charge everything from a single power plug is a total dream. But it'd mean losing MagSafe and I can't see Apple going down that route.


Just use an adapter, apple customers seems to love those.


> apple customers seems to love those

Er, Apple designers seem to love those.


> Er, Apple designers seem to love those.

Stockholders should too.


Is there any reason you couldn't have both?

Use MagSafe at home/work because it's a superior solution, but hopefully Apple hooks up one/all of their USB ports as well to power, so you can recharge with a generic USB cable in a pinch.


It's always possible, but I'm not sure would do it. They like their magsafe, they like the form factor, and they're going down the route of sliming it down. If they could pull it off I imagine they would be but thats probably a while off yet.


Wouldn't they just integrate wireless charging into the battery? To eliminate the need for a port?


USB is slowly approaching "single cable" solution. By adding power, USB3 and Thunderbolt are two differentiated solutions moving towards a similar goal.


While they are both "single cable" solutions, their differences give them different uses. Currently there are lots of cheap USB gadgets, and you probably have lots of USB cables floating around your house. Thunderbolt uses active cables, which means the cables contain hardware to code the signal specially for that cable. This allows much better guarantees of throughput, while making cables much more expensive. With this in mind, I think we are likely to see thunderbolt used in more 'dedicated' scenarios, like removable harddrives, displays, and printers. Daisy chaining means you don't need many ports either.

On the other hand, USB cables are cheap, but have less guarantees on throughput. In fact, since many ports could be on one hub, you can't even use all ports to there maximum potential. USB is likely to stay in its current position for the foreseeable future though.


It is very interesting to see the group move to a switchable voltage, expecting to support 5V, 12V and 20V. My first reaction was 'oh no.. going to fry some 5V electronics with 20V' but if all the negotiation is done right - and not too easy to spoof it should keep people from building cheap adapters to go inline, request the wrong voltage and fry things.

The voltages are a nice choice.

5V: backwards compatibility. Everything is 5V.

12V: same as your car - lots of things run on 12V. It is easier to build a step down converter than a step up. so that gives the hackers some options.

20V: It was genius of them to step up to 20V. as others have pointed out, charging a laptop (often 17-20VDC required) is perfect. This is also a good choice for keeping the conductor size small in the cable. We get to 100W (wow).

now if you really want to be forward thinking, maybe look at the way electronics have moved from 12V, to 5V, and then to 3.3V. If they want this to be super portable for future use, maybe put in 3.3V - although I can't think of what it would be used for right off.


3.3V and 1.8V are often used as VCC in embedded processors, and they're the ranges for hard drives & RAM. If you're dealing with those low powered devices, there's usually a dirt-cheap and tiny DC-DC converter on-board. And you're not going to see anything high-powered enough to use these voltages, I don't think.

RE: 12V. Your car's voltage supply has huge margins (ripple, DC offset, etc), but in general - sure.


As the wireless networks are easier to use, technology that was built for data transfer is now used for power. I would call that cable USC (Universal Serial Charger). It looks like a standardization of a power adapter, but I don't think we can call it USB, as the "serial" part will not be really used. Profiles 4 and 5 would need special power sources, so you cannot just plug your laptop in a simple desktop (or another laptop) and get 100W of power, so you'll get a brick adapter with an USB plug instead of a custom plug - like most cellphones. But I would not expect the brick to use the initial USB function (data transfer).

This also means that the cheap Chinese USB power bricks will be even more dangerous.


It would be unbelievably cool to be able to walk into an office with a laptop, and only have to roll out a USB cable from the desk in order to charge it. If there was some way of combining this with an ethernet port, all my prayers would be answered!


It's compatible with normal USB2.0 and 3.0. No reason it couldn't be offering ethernet as well as power. At that point it does indeed sound quite compelling.


That is the basic idea behind docking stations


Yep, and the reason they work so well is that everybody has the exact same model laptop.

I couldn't care less what the standard is, as long as it's a standard and everybody uses it. 400-pin docking station connectors are not standard, and pity the fool who tries to change that.


20v@5a! Nice! I can't wait to see USB ports instead of the 13.8v ports we currently have in cars.

Sadly my 170w laptop still would need a separate adapter.


Remember that your laptop is unlikely to actually draw 170W at any time unless there is a fault (that it its supply's rated maximum, not what it is expected to draw all the time). Most of the time it'll draw a few 10s of W, maybe stepping up as high as 100W+ when at full tilt (CPU & GPU hammering away, cooling fans on full blast, display at full brightness, all disks spinning & heads flapping, active drives or USB-powered cup warmers in every available port, ...). If you have (or have access to) one of those plug-in power meters give it a try without the battery connected: you might be surprised how little power is being drawn most of the time.

A laptop's power system is essentially a UPS: if there is no (or not enough) power coming in from the main source, flip to the battery. This won't be any different for this use: as soon as it wants to draw more than is available it'll see the under-current and switch over. The smoothing capacitors will ensure it can stay alive during the switch-over just as they do when you unplug the charger normally. If it is a little brighter then at times it might be be able to draw as much as it can from the external power and draw the excess from the battery, but I suspect no laptop is currently designed that way (they generally either get more than they need or nothing at all from the external source, so trying to be clever in response to getting half what you need would be over-engineering).


> instead of the 13.8v ports we currently have in cars.

That's a bit too much precision for that standard. It's only ~13.8 V when your car is running and the alternator is charging the battery. When the car is off, it'll be 13 V or slightly less.


And even when it's running, you have to expect up to at least 16V, some of the cars and trucks with larger alternators will run a bit higher, I've seen 15V myself but I've always heard that 16V is what to expect.


Unless your laptop is using 170W all the time, you could still charge it using just 100W. Even a very high powered laptop is probably using <10W in idle.


I used to think that too. But I just measured my ExoPC slate (Tablet, Atom N450, Pine Trail chipset); It draws 0.1W while suspended, 9W idling with screen off, 13W idling with screen on, and 18W with screen on and 100% CPU.


If you really do need 200W, can't you just plug 2 cables in? I've seen quite a few USB devices that have an extra plug to draw extra juice.

wow... just think if every USB3 port needs an extra 100W of potential from the PSU in your computer (plus the overhead of switching effeciency). So if you have 3 USB3 ports that would mean +300W on your supply (to be safe).


A decent number of cars have USB ports (usually for audio purposes but they also supply power). Ford Sync comes to mind here.


I'm curious if USB could be turning into the next electrical socket. Imagine if everything you plugged into a wall outlet used USB. Not only would it probably be safer, but it would probably mean the end of country-specific electrical adapters. Outlets like this are already becoming popular: http://www.amazon.com/Newer-Technology-Power2U-Outlet-Chargi...


Probably not. The 5v ones can't deliver enough power to do much, and if you have 20v you might as well have 120v - you anyway need to convert the voltage to whatever you use internally, and with 120v the wires are 1/5 the size (thickness).


I find that argument suspect when I look at all the laptops that run on 19 volts. Internal adapters are a lot messier for 120v. Most computing things love the 12 volt range, my screen uses 19 volts, a fancy light bulb is either LED and low voltage or CFL and gets no benefit from 120.

Higher voltage is better for distribution and for very high power loads (including heaters). But a lot of things like low voltage.


> Internal adapters are a lot messier for 120v.

Why do you say that? Look at any USB charging plug - very simple inside. You only need "messy" electronics for high power or high (above 80%) efficiency.

Basically you would need 20v -> 5v vs 120v -> 5v - there isn't really a lot of difference. (There is some, but not enough to make changing everything worthwhile.)

Also, the double conversion isn't great (you would need a power supply in each plug), assuming typical 80% power supplies you would end up with end to end 64% efficiency.


>> Internal adapters are a lot messier for 120v. >Why do you say that?

Well, for starters, most for-profit enterprises like to avoid liability. 20 V things are considered "low voltage" for regulatory purposes while 120 V things are considered "high voltage." That has certain implications for how you have to electrically insulate your portable device.

Then you also have to consider the difference between AC and DC. A power supply meant to connect to the wall has to have an energy storage device (usually a capacitor) for the 120 times per second that the wall delivers no energy.

Designing and manufacturing a device to power from 20 VDC (+/- 10% or so) is amazingly easier than one that must slurp 120 VAC.

>Basically you would need 20v -> 5v vs 120v -> 5v - there isn't really a lot of difference

This is completely wrong. 120 VAC is 120 V RMS, which has peaks of +/-170 V. Go to digikey and compare the price of 340V capacitors to 40V capacitors. Then consider the fact that your low voltage capacitors only need to filter relatively high-frequency ripple and occasional step-loads, while your high voltage caps need to supply energy when the wall isn't (so they needs to have much more capacity).


Having a centralized efficient 20v transformer could actually lead to better efficiency overall. A low voltage DC-DC converter can be far above 80% efficiency without any fancy components.

And when I called 120v adapters messy I meant for the kind of thing I see that runs off 20v. Relatively high-powered computers and game consoles and monitors.


A centralized transformer wouldn't work well. First you would have to run a lot of very fat wires in order to serve all the outlets.

But worse, power supplies have an optimum efficiency, usually at around the 75% usage point. If you have a power supply large enough to power everything, it's not efficient enough to power just a few things.

> And when I called 120v adapters messy I meant for the kind of thing I see that runs off 20v. Relatively high-powered computers and game consoles and monitors.

And I still don't see what's messy about it. Those things would anyway need a power supply to convert the 20v to whatever they use. You could just as easily put in a 120v power supply (and use thinner wires).


Those things tend to have bulky external power supplies that transform 120 volts into 20. 'whatever they use' is already 20 volts, with minimal internal components to get any other levels. You'd be able to eliminate a lot of awkward transformers.

It's probably not worth wiring a whole house with fat wires but in theory having nice safe universal plugs without dealing with external transformers would be nice.

I'm sure if someone wanted to design a supply that's efficient at different currents they could. Worst case would be 2-3 different supplies glued together with a management chip.


Expect a new class of cheap USB gadgets in the future. USB powered microwave oven. There's a lot of toasty cheese melting power in 20 volts and 5 amps. There's going to be some patents filed...


Way ahead of you https://www.thinkgeek.com/stuff/41/fundue.shtml :) (try the buy now button)


The duplex power is interesting. Let's say you've got a fully-charged phone (or similar). Would you be able to use it as a battery instead of charging it?



That is a picture of a standard USB 3.0 A to B cable. Ex. http://www.amazon.com/Micro-Connectors-Superspeed-Cable-Feet...


No it's not. Look at the plastic "tongue" in the middle of it - it was pierced and they put the power connectors inside it.


Um, yes it is. How about the fact that that particular image is at least 3 years old?

here is an article from nov 2009 http://www.tuexperto.com/2009/11/05/intel-no-desarrollara-el...

and many more http://www.tineye.com/search/d582fc65885700337f2c3ade01878bc...


Seems you're right.

But that makes me wonder where they are adding the pins, there doesn't seem to be room (unless it's inside the connector, with two rows of pins).


The device end seems really chunky for today's electronics. I can't see that being used to power anything more portable than a laptop (and it's still big for ultraportables) if there is no mini version.


Is the USB 3.0 still held by Intel?




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: