[Cialug] Power Down

[Zachary Kotlarek]

On Apr 24, 2008, at 5:20 PM, Matthew Nuzum wrote:

> On Thu, Apr 24, 2008 at 4:03 PM, Todd Walton <tdwalton at gmail.com>  
> wrote:
>> The beef of the article is to shill for the APC Power-Saving
>> SurgeArrest surge protector.  It's a nifty idea: the you plug your
>> computer into a "master" plug-in, and then your printer and monitor
>> and speakers and all those other gadgets into slave outlets.  When
>> whatever is plugged into the master outlet stops drawing power the
>> strip cuts power to the slave outlets.  So, when you turn your
>> computer off all your other stuff gets turned off too.
>
> I wonder why they don't have a usb plug on them so that when your
> computer hibernates or goes to sleep (and presumably powers down the
> usb port) it detects it and powers down your devices.
>
> I agree this is a serious problem. It seems like it shouldn't be too
> hard to make the average power brick smart enough to limit its power
> usage to say a couple mA when the device attached to it stops drawing
> power or is disconnected.

Actually that's the whole problem -- it is somewhat difficult.

For one thing, your standard sub-1 A, unregulated power brick uses  
about the same amount of power whether or not a load is attached. And  
short of a manually-actuated power switch there's no good way around  
that. Not only would monitoring the load significantly increase the  
complexity and reduce the compatibility of such devices, but it would  
itself require a power supply to power the monitoring equipment. You'd  
have a hard time designing a system that saved more power than it used  
to do monitoring, and even if you did it would likely double the size  
of the power brick.

For larger devices such monitoring is more practical, and is  
essentially what happens -- most DC-powered, soft-power devices have a  
reasonably high efficiency switching power supply for the main  
operational load, and a secondary, always-on power supply that run the  
control circuitry for the former. There's some room for improvement in  
these types of systems, as they already have dual power supplies, but  
it's not an easy task. You still need to constantly supply a small  
amount of power to the control circuitry, and while inefficient,  
resistively regulated power supplies (i.e. power blocks) are still one  
of the best ways to do that.

The fundamental problem is that each device needs its own power  
control circuitry, and along those lines, that it's inefficient to put  
power control circuitry into low-power systems. If you had a  
centralized control system that could turn outlets on and off you'd  
only need a single 500 mA power brick per house (or at least per room)  
to run a single set of sensors IR/proximity/power usage/motion/etc.  
But power control systems are expensive, and most homes are not  
equipped with individually switched outlets, let alone a programmable  
control system for those outlets. And programming those outlets and  
sensors to work correctly together is non-trivial as well, at least  
for non-geeks. I personally change my thermostat settings based on the  
presence or absence of Bluetooth devices, but that's probably asking a  
lot for the average homeowner.

	Zach
-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 1682 bytes
Desc: not available
Url : http://cialug.org/pipermail/cialug/attachments/20080424/2b9c2b64/smime-0001.bin