Author Topic: USB 3.0 Power Delivery Specification  (Read 16241 times)

0 Members and 1 Guest are viewing this topic.

Offline PsiTopic starter

  • Super Contributor
  • ***
  • Posts: 9889
  • Country: nz
USB 3.0 Power Delivery Specification
« on: July 24, 2012, 11:38:26 am »
http://www.techweekeurope.co.uk/news/usb-3-0-power-delivery-charging-86983

I'll copy the interesting details in case you don't want to read it.

Quote
The USB Power Delivery Specification enables a switchable source of power.
It will feature six power profiles, and will be able to alter between them on the fly.

Profile 1 - 5V@2A (10W)
Profile 2 - 5V@2A  or 12V@1.5A (18W)
Profile 3 - 5V@2A  or 12V@3A (36W)
Profile 4 - 20V@3A (60W)
Profile 5 - 12V or 20V at 5A (100W)

The high power charging will require special, electronically detectable and very clever “USB Power Delivery” certified cables, which will be backwards compatible with all types of USB 2.0 and 3.0.

100W seems kinda insane to me, but if we get ultra fast charge batteries in our phones then it maybe needed.

One thing that worries me is viruses destroying your hardware or starting fires by compromising the USB subsystem and forcing the port to 20V when the device expects 5V
« Last Edit: July 25, 2012, 05:02:09 am by Psi »
Greek letter 'Psi' (not Pounds per Square Inch)
 

Offline digsys

  • Supporter
  • ****
  • Posts: 2209
  • Country: au
    • DIGSYS
Re: USB 3.0 Power Delivery Specification
« Reply #1 on: July 24, 2012, 11:56:40 am »
Quote
One thing that worries me is viruses destroying your hardware or starting fires by comromising the USB subsystem and
forcing the port to 20V when the device expects 5V
Better start adding a polyfuse + transorb to all USB inputs I guess !! No doubt there'll be a few tears until everyone finds out
Hello <tap> <tap> .. is this thing on?
 

Offline PsiTopic starter

  • Super Contributor
  • ***
  • Posts: 9889
  • Country: nz
Re: USB 3.0 Power Delivery Specification
« Reply #2 on: July 24, 2012, 11:59:33 am »
As long as the voltage request and switching is done entirely in hardware it should be ok
Greek letter 'Psi' (not Pounds per Square Inch)
 

Offline Things

  • Regular Contributor
  • *
  • Posts: 224
  • Country: au
  • Laser Geek
    • NQLasers
Re: USB 3.0 Power Delivery Specification
« Reply #3 on: July 24, 2012, 12:00:42 pm »
It'd make sense to me that the only thing that could change the USB voltage would be the device itself, and the change would be done in hardware rather than software (Although probably still in software, but using a uC or other device not part of the main processor)
 

Offline _Sin

  • Regular Contributor
  • *
  • Posts: 247
  • Country: gb
Re: USB 3.0 Power Delivery Specification
« Reply #4 on: July 24, 2012, 12:12:56 pm »
Ah, this will usher in a new age of pointless USB "peripherals".

Having said that... mini USB reflow oven anyone?

Programmer with a soldering iron - fear me.
 

Offline PsiTopic starter

  • Super Contributor
  • ***
  • Posts: 9889
  • Country: nz
Re: USB 3.0 Power Delivery Specification
« Reply #5 on: July 24, 2012, 12:19:20 pm »
With 100W  (20V @ 5A) you could drop the custom AC adapter and charge over USB instead.

Which introduces some interesting ideas...  Like charging one laptop from another, or the ever so popular energy black hole when you charge a laptop from itself with a loop cable :D
« Last Edit: July 24, 2012, 12:22:45 pm by Psi »
Greek letter 'Psi' (not Pounds per Square Inch)
 

Offline _Sin

  • Regular Contributor
  • *
  • Posts: 247
  • Country: gb
Re: USB 3.0 Power Delivery Specification
« Reply #6 on: July 24, 2012, 12:23:44 pm »
USB soldering iron, just in case I need to repair something while travelling. Like the laptop!

Programmer with a soldering iron - fear me.
 

Offline Rerouter

  • Super Contributor
  • ***
  • Posts: 4694
  • Country: au
  • Question Everything... Except This Statement
Re: USB 3.0 Power Delivery Specification
« Reply #7 on: July 24, 2012, 12:25:09 pm »
actually psi that is something i have to wonder about, would anyone ever think that far ahead to actually prevent a device that is both powered by and supports usb to not stupidly drain itself from a power loop like that :/
 

Offline PsiTopic starter

  • Super Contributor
  • ***
  • Posts: 9889
  • Country: nz
Re: USB 3.0 Power Delivery Specification
« Reply #8 on: July 24, 2012, 12:32:52 pm »
Doubtful, if someone wants to run their laptop flat quicker who are we to argue with them :P
Greek letter 'Psi' (not Pounds per Square Inch)
 

Offline T4P

  • Super Contributor
  • ***
  • Posts: 3697
  • Country: sg
    • T4P
Re: USB 3.0 Power Delivery Specification
« Reply #9 on: July 24, 2012, 12:58:54 pm »
Special cables ...  ::) Go on, i want to see people using AWG 24 wires  ;D ( That's like the largest they use now )
 

Offline madires

  • Super Contributor
  • ***
  • Posts: 7695
  • Country: de
  • A qualified hobbyist ;)
Re: USB 3.0 Power Delivery Specification
« Reply #10 on: July 24, 2012, 08:46:53 pm »
Will we see standalone USB power supplies? Or does my PC power supply need to provide +100W for each USB port? 10 USB ports = 1000W? Doesn't make much sense. Maybe just two 100W USB ports?
 

Online tom66

  • Super Contributor
  • ***
  • Posts: 6679
  • Country: gb
  • Electronics Hobbyist & FPGA/Embedded Systems EE
Re: USB 3.0 Power Delivery Specification
« Reply #11 on: July 24, 2012, 08:59:27 pm »
Why 20V, and not something like 24V or 42V? Both have readily available DC-DC converters. Or, same voltage as POE, 48V DC. Keep it simple...
 

Offline Sionyn

  • Frequent Contributor
  • **
  • Posts: 848
  • Country: gb
Re: USB 3.0 Power Delivery Specification
« Reply #12 on: July 24, 2012, 09:05:14 pm »
its understandable ive never like usb device that use usb to suck power like fans hovers etc
eecs guy
 

Online mariush

  • Super Contributor
  • ***
  • Posts: 4983
  • Country: ro
  • .
Re: USB 3.0 Power Delivery Specification
« Reply #13 on: July 24, 2012, 11:07:23 pm »
... and just as everyone designs power supplies with dc-dc regulators for 3.3v and 5v inside that can only do 20-25a max, they have to come up with 5v@2a for each port...
... and why the hell would you add 20v, how did the reason selecting this voltage? Compatibility with laptop chargers?  Why not 24?

Well, I guess devices won't really all use 2A on each port but still...

In an ideal world I would like to see pc power supplies that would deliver just 12v and possibly 5v to the computer and get rid of the 3.3v rails... seems like only mSATA and 1.8" ssd drivers currently use it. Some motherboards use it with a dc-dc converter to power the ram slots but that's about it as far as I know.

Anyway, I just can't picture 100w on usb ports in the back of a motherboard... wonder if there's gonna be ANOTHER 8pin connector there just for usb.
 

Online tom66

  • Super Contributor
  • ***
  • Posts: 6679
  • Country: gb
  • Electronics Hobbyist & FPGA/Embedded Systems EE
Re: USB 3.0 Power Delivery Specification
« Reply #14 on: July 24, 2012, 11:20:04 pm »
My idea would be to ditch the 5V, 12V, everything, and replace it with a single 3.3V or 5V AUX (standby rail @ 3A max) and a relatively unregulated 30 to 42V DC rail. Reasoning behind this is 30-42V is within the range of a lot of 42V DC-DCs so there's existing parts base, it's reasonably safe (can't shock yourself on 42V, it might tingle a bit but that's it) and being tolerant of unregulated supplies means you can directly power the PC from a 10 cell lipoly pack or an 18 cell (3x12V) lead acid battery. Also you wouldn't need a ridiculous 80A on 12V for 1kW, but a more sane ~35A. In the ideal design, the aux rail would be optional, and if it is not present, the motherboard provides the aux power from the main power, otherwise the aux power is used to run standby functions and turn on/off the main rail.

The only things that really care about regulation are fans and motors, so there'd either need to be speed regulation on each or a 12V DC-DC on the motherboard.

It could also be practical to power computers off a central DC supply, but it would be more likely to be 300V+... I believe some server rooms are doing this now. Obviously you don't want 300V in your computer.
« Last Edit: July 24, 2012, 11:30:53 pm by tom66 »
 

Online mariush

  • Super Contributor
  • ***
  • Posts: 4983
  • Country: ro
  • .
Re: USB 3.0 Power Delivery Specification
« Reply #15 on: July 25, 2012, 12:39:34 am »
The reason why I said 5v and 12v is simply for compatibility with all the devices that already exist out there. I would ditch the 5v as well and keep only 12v but there's too many devices out there using 5v...

I would think 3.3v is too low, the voltage drop on the cables would be big if the cables are long, you would probably need voltage feedback from the power supply to adjust for the losses. 12v is reasonable.

The Open Compute guys (facebook's open source hardware initiative) seem to have about the same idea: http://opencompute.org/projects/power-supply/

They're designing the servers to work on 12.5v dc and their power supplies as you can see in the link above accept 277v AC (+/-10%)  or 48V DC nominal...though they say it's backup power, connected to a 54v battery bank.
 

Offline T4P

  • Super Contributor
  • ***
  • Posts: 3697
  • Country: sg
    • T4P
Re: USB 3.0 Power Delivery Specification
« Reply #16 on: July 25, 2012, 02:31:08 am »
Will we see standalone USB power supplies? Or does my PC power supply need to provide +100W for each USB port? 10 USB ports = 1000W? Doesn't make much sense. Maybe just two 100W USB ports?

Special USB 3 ports, not just normal usb ports.

 

Offline T4P

  • Super Contributor
  • ***
  • Posts: 3697
  • Country: sg
    • T4P
Re: USB 3.0 Power Delivery Specification
« Reply #17 on: July 25, 2012, 02:39:55 am »
ATX Standard was formed so damn long ago, stop complaining

... and just as everyone designs power supplies with dc-dc regulators for 3.3v and 5v inside that can only do 20-25a max, they have to come up with 5v@2a for each port...
... and why the hell would you add 20v, how did the reason selecting this voltage? Compatibility with laptop chargers?  Why not 24?

Well, I guess devices won't really all use 2A on each port but still...

In an ideal world I would like to see pc power supplies that would deliver just 12v and possibly 5v to the computer and get rid of the 3.3v rails... seems like only mSATA and 1.8" ssd drivers currently use it. Some motherboards use it with a dc-dc converter to power the ram slots but that's about it as far as I know.

Anyway, I just can't picture 100w on usb ports in the back of a motherboard... wonder if there's gonna be ANOTHER 8pin connector there just for usb.


Keep all the rails, there's really not going to have a point to throw away all the current voltages, as +12/-12v +5V +3.3V is the most efficient and best tradeoff between cost and complexity
Imagine more DC-DC converters, 3.3V provides power for the 2.5V references on board
http://en.wikipedia.org/wiki/Molex_connector#Motherboard_power_connector_.28Molex_Mini-fit_Jr..29


Try developing a buck converter that is 42V-sub 2V.
 

Offline PsiTopic starter

  • Super Contributor
  • ***
  • Posts: 9889
  • Country: nz
Re: USB 3.0 Power Delivery Specification
« Reply #18 on: July 25, 2012, 04:59:47 am »
They're going to need a socket colour system to denote the max power profile which a USB3 port supports.
Otherwise it's going to be impossible to know which port supports a device without trying it.
Greek letter 'Psi' (not Pounds per Square Inch)
 

HLA-27b

  • Guest
Re: USB 3.0 Power Delivery Specification
« Reply #19 on: July 25, 2012, 05:18:45 am »
Google engineers proposed deleting all voltages from the power supplies except +12V. Their rationale being that boards have their own voltage regulators anyway so eliminating unnecessary rails would make computers cheaper and more energy efficient. According to them 30 to 40% of the consumed power is wasted for voltage conversion.

services.google.com/blog_resources/PSU_white_paper.pdf


USB folks have a lesson to learn from that.
 

Offline T4P

  • Super Contributor
  • ***
  • Posts: 3697
  • Country: sg
    • T4P
Re: USB 3.0 Power Delivery Specification
« Reply #20 on: July 25, 2012, 10:27:46 am »
Google engineers proposed deleting all voltages from the power supplies except +12V. Their rationale being that boards have their own voltage regulators anyway so eliminating unnecessary rails would make computers cheaper and more energy efficient. According to them 30 to 40% of the consumed power is wasted for voltage conversion.

services.google.com/blog_resources/PSU_white_paper.pdf


USB folks have a lesson to learn from that.

Sorry i mean, 12V-5V is pointless. It creates more waste heat for the harddisks
« Last Edit: July 25, 2012, 02:13:34 pm by DaveXRT »
 

Online tom66

  • Super Contributor
  • ***
  • Posts: 6679
  • Country: gb
  • Electronics Hobbyist & FPGA/Embedded Systems EE
Re: USB 3.0 Power Delivery Specification
« Reply #21 on: July 25, 2012, 12:20:18 pm »
Google engineers proposed deleting all voltages from the power supplies except +12V. Their rationale being that boards have their own voltage regulators anyway so eliminating unnecessary rails would make computers cheaper and more energy efficient. According to them 30 to 40% of the consumed power is wasted for voltage conversion.

services.google.com/blog_resources/PSU_white_paper.pdf


USB folks have a lesson to learn from that.

Ya sure, 12V-1.2V is not what you call efficient

Huh? VRMs already use 12V. They are nearly 90% efficient. More efficient than at 5V, due to lower current needed at 12V.
 

Offline PsiTopic starter

  • Super Contributor
  • ***
  • Posts: 9889
  • Country: nz
Re: USB 3.0 Power Delivery Specification
« Reply #22 on: July 25, 2012, 12:22:42 pm »
Yeah, the losses are on the input side.
So a higher input voltage results in less switching loss.
12V -> 1.2V will be more efficient than 5V -> 1.2V

This is why lots of computer motherboards generate the CPU voltage using the 12V or 5V rail instead of the 3.3V rail.
« Last Edit: July 25, 2012, 12:26:08 pm by Psi »
Greek letter 'Psi' (not Pounds per Square Inch)
 

Offline ivan747

  • Super Contributor
  • ***
  • Posts: 2045
  • Country: us
Re: USB 3.0 Power Delivery Specification
« Reply #23 on: July 25, 2012, 03:49:54 pm »
Why 20V, and not something like 24V or 42V? Both have readily available DC-DC converters. Or, same voltage as POE, 48V DC. Keep it simple...

Well, this (if it is successful) will be used in millions of devices, enough to make semiconductor companies tweak their 24V DC-DC converters to output 20V.
 

Online tom66

  • Super Contributor
  • ***
  • Posts: 6679
  • Country: gb
  • Electronics Hobbyist & FPGA/Embedded Systems EE
Re: USB 3.0 Power Delivery Specification
« Reply #24 on: July 25, 2012, 09:10:04 pm »
Why 20V, and not something like 24V or 42V? Both have readily available DC-DC converters. Or, same voltage as POE, 48V DC. Keep it simple...

Well, this (if it is successful) will be used in millions of devices, enough to make semiconductor companies tweak their 24V DC-DC converters to output 20V.

Yes, but using already existing componentry means that it will be cheap and ready from the start, without having to wait months for development of new components and designs.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf