Slashdot Mirror
New USB Specification Promises 100W of Power
Blacklaw writes "The group behind the USB 3.0 specification has announced a tweak which could lead to impressive new devices, including large-format displays, printers, and even laptops that are powered entirely from a USB port."
Awesome. I'll finally be able to implement those high powered "negative reinforcement" keyboards I keep dreaming about.
Netbook battery life drops to an average of 12 minutes.
Burns: We're building a casino!
McAllister: Arrr. Give me 5 minutes.
So now, I can hook my computer to my car to jump-start it!
You know? Long ago, Apple made a display that was powered through the display cable. It worked but it was not popular in the end as they stopped doing it. So they are talking about bringing it back again?
I can see power enough to power some devices, but 100W?
You know, whatever USB standards come out, it should work equally well on a battery powered laptop and a desktop as well. People will get confused and frustrated when they buy a fancy new USB 3.0 display unit only to find they can't take it with them on the road because it doesn't work well with their laptop and the tiny travel power adapter they use while on the road.
I just hope it becomes a standard because PoE was nowhere to be found.
The mark up for PoE switches was/is spectacular, because the marketing guys told them to price that feature at "just below the cost of hiring a union electrician to run a dedicated AC line next to the wall plug". Which, it turns out, is a heck of a lot of money.
The marketing people forgot about extension cords. So, most of the real world uses extension cords instead. Whoops. PoE was a cool idea, but too sabotaged to ever make it.
"Science flies us to the moon. Religion flies us into buildings." - Victor Stenger
PoE, for whatever reason, is absolutely dead in the consumer space; but it is alive and kicking in corporate gear. Not quite 100%, of course, because a PoE switch necessarily costs more than a non-PoE one, and wasting PoE ports on desktops and docking stations doesn't make any sense; but some gigantic portion of the corporate world's APs, IP cameras, access control devices, and similar low-power-and-networked junk are PoE powered...
Even with USB2, there was the persistent problem that certain applications(notably 2.5 inch external drives) were right on the edge of what the spec allowed. Some machines played fast and loose, and everything worked fine, some played to spec, and the device wouldn't spin up, or the bus would freak out, or whatever. Despite USB's formalized, standardized, power-request mechanism(100ma on connect, negotiate in units of 100ma for up to 400 more...), the, er... 'inventive'... nature of the peripheral ecosystem always created some uncertainty: Some devices just requested 500ma at all times, to avoid possible brownouts, leading to more spec-compliant busses freaking out about lack of power even when actual draw was well within safe limits, some devices (fans, LED goosenecks, humping dogs) just grabbed the +5 and ground rails and hoped for the best, without any negotiations. Some hubs report themselves as self-powered(and thus good for 500ma per-port) even when they were bus powered(and thus only good for 400ma across however many ports they had). Some others were self-powered; but with wall-warts that could only deliver 500ma to a number of ports smaller than the number available(7-port hubs with 1amp adapters, I'm looking at you...)
This new standard seems like it would simply be a polite codification of this confusion. Particularly at low voltage, 100watts is nontrivial current(and nontrivial power generally, most non-DTR laptop bricks are less than that...) Many PCB layouts would burn a trace trying to deliver that, and you can bet that your garden-variety 10-USB-ports boring desktop isn't going to ship with 1000watts of PSU headroom...
This will mean that, in effect, devices will be able to demand up to 100 watts in a 'compliant' way; but the capabilities of USB ports on the market will vary enormously by device. A laptop with an 85 watt power brick is hardly going to be good for 100watts out of a port. Worse, it might be good for 50 when lightly loaded and fully charged; but only 5 when charging its battery and flogging its CPU... Having a device that only intermittently functions is near worthless, even if it is all entirely standard... A desktop might ship with the ability to push a single port to 100; but then it will either have to beef up its traces significantly, or have the always-confusing-to-dumb-users-and-people-fumbling-behind-desks '1 special blessed "high power" port, and 9 identical-feeling-but-low-power ones' configuration. Fan-fucking-tastic...
While a bit more power on the bus certainly would expand the number of viable, bus-powered use cases, I'm just not sure that such a high 'standard' number can ever be usefully 'standard'. Hooray, it is now officially standard for specialized devices to shove 100watts across a USB bus. Doesn't change the fact that it won't work in 90+% of ports, and will probably burn a fair percentage of cheaper cables. Unless they come up with some sensible set of "tiers", so that people actually know what works with what, this seems like it is going to end in a mess of nominally-USB powered docking stations with wall warts and mini-B connectors, at best.
While its comparative obscurity, and the general lack of bus-powered devices made it less of an issue, Firewire flirted with this problem in its early days: Both available power and available voltage on a given 6-pin port were widely variable: A desktop could, if it so chose, be pumping out 24 volts and reasonably credible wattage. One of the(almost exclusively Apple) laptops with a 6-pin port might be limited to a handful of watts at whatever voltage its battery was set to provide. In practice, much firewire gear just skipped bus power entirely(despite the fact that charging over firewire would have been a very popular consumer camcorder feature, if today's flip-likes are anything to go by), the mixture of widely variable power availability, and the 'i.link' or just 'IEEE1394' connectors entirely without bus power pretty much doomed the widespread availability of bus-powered peripherals. USB's pitiful 2.5watts was rather limiting; but at least you could reasonably assume that it would be there...