[Beowulf] Re: power usage, Intel 5160 vs. AMD 2216

David Mathog mathog at caltech.edu
Mon Jul 16 09:14:47 PDT 2007


Jim Lux wrote:

> Whoa there cowboy!!  You're risking death, dismemberment, and 
> egregious excitement from the destruction of your datalogger.  The 
> entire insides of the Kill-A-Watt are floated at line 
> potential.  

There's two sides to the line potential, and one is a lot worse than
the other.  I expected big voltages on the analog
side but nothing too extreme on the more delicate digital side.

Now I'm curious.  Being not quite as suicidal as my previous post would
suggest, with the Kill-A-Watt unplugged, I checked for
direct connections between the 6 wires and the 3 socket
connectors using the highest resistance setting on a multimeter. 
Result: one of the 6 wires was shorted to the neutral (at least it
read 0 down to the 2KOhm setting - this was the wire one in from the
"J6" side), there was no electrical connection at all between
the ground and any of the wires, and there were an assortment
of resistances from the hot pin to the other 5 wires.

I didn't test it plugged in, but from this it looks like 
the voltages on those 6 wires are referenced to line Neutral.
Admittedly under a large enough load the Neutral could differ from
ground, and the difference could be enough to make the data logger
really, really, unhappy if it internally shorts the measurement
"ground" lead to actual ground, but unless the socket has been wired
with neutral and hot reversed, it's probably not ever going to
be more than a couple of volts difference.  That's enough to kill the
data logger, probably not enough to kill the operator.

> 
> It's not just a analog current and voltage measurement that you can 
> log, because you need to measure power factor, and that requires 
> digitizing the waveform (or at least fancy electronics to measure 
> instantaneous I*V)..

Right.  I _did_ say that it made a lot more sense to just buy the
$130 data logging version.

Regards,

David Mathog
mathog at caltech.edu
Manager, Sequence Analysis Facility, Biology Division, Caltech



More information about the Beowulf mailing list