Xiaofan Chen wrote:
> Even with more traditional lower switching frequency buck converter
> (MOSFET + free-wheeling diode), where conduction loss is the
> dominating
> factor, lower duty ratio is still not desired as the free-wheeling
> diode will conduct most of the time and thus have higher loss due to
> the diode
> voltage drop.

That's a good point I screwed up in my previous message.  In a buck
converter, part of the output current comes directly from the input and some
from ground thru the diode.  The ratio depends on the input to output
voltage ratio, with high input to output voltage ratio causing a larger
portion of the output current to pass thru the diode and its voltage drop.
At a low output voltage like 5.6V, even a Schottky can waste 5% of the power
during the current from ground part of the pulse.  If you make sure
switching times are short compared to the pulse width, the high side
switching losses will decrease with higher voltage.  In the end you can
achieve are roughly constant efficiency as I did over a reasonable range.
Since at my lowest input voltage the ratio was already nearly 4:1, the diode
was conducting for about 3/4 of the output current already, so adding the
other 25% didn't make that much of a difference.

> That is why synchronous buck converter becomes more
> and more popular in many applications.

Absolutely, especially with low output voltages.

However let's keep in perspective the original question was what voltage
Dario should bus around his house to power a few distributed devices.
Simple buck regulators that are good enough can be easily designed with 12,
24, or 48 volts input.  I was advocating the higher voltage to reduce IIR
losses in the cable and maximize the amount of power that can be transmitted
for the same size cable.  Over time he is likely to add more and more
devices, so designing now for more power capacity sounds like a good idea.
I envision a 48V off the shelf commercial power supply in the basement
powering the system all the time, then the modules bucking that to 5.6V with
a 5V LDO post regulator.  We may disagree on whether 48V will be less
efficient to use than 24 or 12 volts, but in either case a few percent
difference in the modules isn't really the issue, and would probably be more
than offset by the cable losses anyway.  If he uses off the shelf CAT5
cable, then it will have a maximum current carrying capacity.  At 48V that
same cable will be able to deliver twice the power to the distributed
collection of modules than at 24V.


********************************************************************
Embed Inc, Littleton Massachusetts, http://www.embedinc.com/products
(978) 742-9014.  Gold level PIC consultants since 2000.
-- 
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist