This whole article can be summed up with the one sentence: >The big change next year will be a new USB PD (Power Delivery) standard, which brings much more flexibility and ten times as much oomph: up to 100 watts.
The remainder of the article seems to epicly overstate the benefits. I found the reaches out to multi-building solar panel connections especially "out there" and that's coming from someone who _loves_ solar panels. It's just that solar panels and the talk about "poor neighborhoods with no access to AC mains" are veerings straight off the deep end.
The reason data centers use DC is to keep the space-consuming transformers that go from AC to DC away from the dense, expensive racks of equipment. Most homes don't have that constraint. Yes, it's possible we'll have DC jacks alongside AC jacks in our homes, but will that really change our lives significantly?
There were posts about switching power supplies (and Apple's dubious claims to have invented them) on HN a few days ago. The consensus is they save space, but was that really worthy of being called a revolution? Only in the strictest most limited definition of the word.
Finally, w/r/t to the Internet of Things. This is where the writer's glib worship of updated standards is the most violent. _USB is not an internet network_. It requires a host. USB is not suited for it. Ethernet will win. Ethernet always wins.
Ethernet's magnetically isolated transcievers are a bit pricey, though. Recently I've been wondering about rallying the Arduino crowd behind RS485 or CANBUS for this sort of connected cheap device network.
That (CANBUS and RS485) works ok as long as everything agrees on what "ground" is. Using differential signaling (as both those standards do) does give you some noise immunity and does help with the ground issue somewhat.
The problem is that ground loops can easily generate 10's of volts at the highest impedance point, and there will be ground loops if you try to use this for "smart homeish" things. Most bus drivers and receivers aren't meant for voltages that high--many drivers/receivers have absolute maximum ratings only a few hundred millivolts below ground.
Isolating everything isn't a conceptually difficult thing to do, but you'll spend just as much on optoisolators as you would have on ethernet magnetics.
Ethernet is just part of the picture, allowing communication between homes but not necessarily between devices. Of course, since both California and Britain mandate ZigBee in their smartgrid projects you could say that ZigBee has already won, but that's at a level opaque to consumers.
This whole article can be summed up with the one sentence: >The big change next year will be a new USB PD (Power Delivery) standard, which brings much more flexibility and ten times as much oomph: up to 100 watts.
The remainder of the article seems to epicly overstate the benefits. I found the reaches out to multi-building solar panel connections especially "out there" and that's coming from someone who _loves_ solar panels. It's just that solar panels and the talk about "poor neighborhoods with no access to AC mains" are veerings straight off the deep end.
The reason data centers use DC is to keep the space-consuming transformers that go from AC to DC away from the dense, expensive racks of equipment. Most homes don't have that constraint. Yes, it's possible we'll have DC jacks alongside AC jacks in our homes, but will that really change our lives significantly?
There were posts about switching power supplies (and Apple's dubious claims to have invented them) on HN a few days ago. The consensus is they save space, but was that really worthy of being called a revolution? Only in the strictest most limited definition of the word.
Finally, w/r/t to the Internet of Things. This is where the writer's glib worship of updated standards is the most violent. _USB is not an internet network_. It requires a host. USB is not suited for it. Ethernet will win. Ethernet always wins.