[Beowulf] DC Power Dist. Yields 20%

[Jim Lux]

At 08:41 AM 8/11/2006, Geoff Jacobs wrote:
>Mark Hahn wrote:
> >> http://www.eweek.com/article2/0,1759,2000867,00.asp
> >>
> >> 20% is a lot, both in terms of consumption and cooling capacity. I'm
> >
> > the article says "up to 15%" at the facility level - did you get 20%
> > by figuring a savings at the rack level, too?
>I got 20% by misquoting, sorry. Still, 15% is nothing to sneeze at.
>
> > afaikt, the article is based on an assumption that everything is powered
> > by an online UPS, and probably that node PSU's are low-performance (say,
> > 65%).  sometimes studies like this ignore fact that incoming power is
> > inherently AC (that is, only start looking at efficiency given DC supply.)
> >
> > I'd love to see DC gain more traction - PSU's are certainly one of the
> > flakier components in our systems, though per node (HP DL145G2), they
> > only contain 2 of 14 fans (or of 18 moving parts).  I don't know whether
> > there's a reason to think many small AC-DC PSU's would be less efficient
> > than a couple really big ones (factoring in the cost and inefficiency of
> > DC power distribution).
>Do your PDUs receive 220 VAC or ~400 VAC?
>
>They're talking about a 380 VDC distribution grid. On it's face, this
>infrastructure would be at least as efficient as a supply of 270 VAC.
>Unfortunately, the article doesn't lean too heavily on hard facts wrt
>their claims.


380V is fairly high.. 1.4*240V is only 340V or there abouts..

In any case, 300+ Volts for DC distribution is fine to work with.

However, there are some subtleties with DC power distribution.  For 
instance, mechanical switches and relays need to be much larger and have 
provisions for extinguishing the arcs that are drawn (fast moving 
electrodes, magnetic quenching, etc.). AC has a zero crossing every 5-8 
msec which limits the time of the arc.  Likewise, fusing and overcurrent 
protection in general is a bit trickier. (that's why 30VDC fuses are bigger 
than 240 VAC fuses)


>I really just have a physicist's understanding of the forces at work
>here. At this point, perhaps we should get Jim Lux to educate us?
>
> > I'd certainly be interested in a distribution system (whether AC or DC)
> > that avoided so damn many plugs and sockets and breakers and PDUs.
> > I guess I'm more enthused about servers becoming lower-powered, and also
> > quite interested in better ways to dissipate the heat than raised floors
> > and traditional chillers...
>Heh heh. Water cooled racks?
>
> >> curious how long it would take before DC supplied racks become cost
> >> effective.
> >
> > well, there's already a standard DC supply - to the motherboard.
> > one impediment might be that it's got +12, -12, 5, 3.3 and probably
> > a couple others.  if it were just a matter of providing lightly
> > regulated 12V, life would probably be a lot simpler.  and I'm not sure
> > MB's would be much more complicated, either, since the current
> > main consumer, the CPU, already has a fairly flexible and high-power
> > onboard dc-dc converter.  (I wonder how efficient it is, typically...)
>
>Well, a DC-DC power converter is pretty straight forward. I expect part
>of the trouble with switching power supplies is those high frequency
>mosfets, which you wouldn't have (I expect) with a DC-DC converter.


Sure you would.. Actually, these days, you might use an IGBT, depending on 
the switch rate..

Switching PSUs all rely on rectifying the AC supply to generate a DC bus 
voltage, that is then converted to the DC voltage you want.  There's a half 
dozen or so circuit topologies, but they all basically rely on turning the 
DC on and off quickly and then doing something useful with it.  The two 
basic strategies are: You turn it into high frequency AC, then run that 
through a transformer, and rectify and filter it to low voltage DC.  THis 
saves the iron and copper cost of a high power transfomer (since the core 
mass is inversely proportional to frequency... 20kHz or 100kHz saves a 
bunch over 60Hz)

The other strategy is to take the HV DC, and pulsewidth modulate it, then 
filter it, to make a lower voltage (the classic "buck" converter).  That 
is, take 100VDC, turn it on and off with a 5% duty cycle, low pass filter, 
and get 5V.


Jim