[Beowulf] Re: ECC Memory and Job Failures (Huw Lynes)

Robert G. Brown rgb at phy.duke.edu
Mon Apr 27 12:09:05 PDT 2009


On Mon, 27 Apr 2009, Lux, James P wrote:

> There are so many physics things wrong with this, it's hard to know
> where to start.

Well said, top to bottom.

I would make two tiny exceptions.  One is that if high voltage arcs
(even weakly) around insulators under the right circumstances, it spits
out high energy broad spectrum noise that can induce arcs in other
nearby near-circuits with enough cross-sectional area.  Arcs
(intermittant or otherwise) aren't unknown, and even relatively low
voltage/current arcs give off enough energy to cause "static" noise on
radios or (possibly) induce spiky overvoltages in nearby places where
there are large conducting loops.  So "radiation" no, but RF noise
maybe.

The other I haven't actually heard of in the literature, but perhaps you
have.  If there is arcing or a heavy corona in the vicinity of sharp
points exposed to air, it does seem like the immediate vicinity of the
defect might be able to produce e.g. ozone and/or nitrous oxide, both of
them more dense than air, both of them highly oxidative (and hence
capable of damaging tissue).  Granted that epidemiological studies show
marginal results at best -- more anecdotal evidence than anything
reproducible -- I've often wonder if the little signal that might or
might not be buried in the noise might not be due to collateral damage
caused by chemistry locally catalyzed by the power line, not
"radiation".  Toxic molecules created by the large store of free energy,
which might only happen at all at "problem" sites -- places where there
is a defect that permits arcing -- and hence PRODUCE a spotty, anecdotal
signal that disappears on studies involving well-maintained lines or
varies seasonally enough to wash out the primary signal.

Because this alternate explanation for an association between HV power
lines and cancer occurred to me some time ago, I've tried to keep an
open mind on the issue.  Maybe you know of somebody that has looked into
it and eliminated this possibility?  I know that the signal itself is
very weak (although studies that get what SHOULD be statistically
significant results are not unknown, one is reduced to worrying about
methodology).

    rgb

Robert G. Brown	                       http://www.phy.duke.edu/~rgb/
Duke University Dept. of Physics, Box 90305
Durham, N.C. 27708-0305
Phone: 1-919-660-2567  Fax: 919-660-2525     email:rgb at phy.duke.edu





More information about the Beowulf mailing list