[SGVLUG] Calling the brightest of cal tech and other nerds!

[David Lawyer]

The following is unlikely but this is a strange case.  What about the
possibility of a broken (or high-resistance connection) neutral
conductor in the house wiring so that the return current flows thru
water pipes, etc. and the earth, instead of the copper conductor it's
supposed to flow in.  This can happen if the house is wired with a
jumper going from the neutral white wire to physical ground at the
fuse (or CB) box.  Then the green ground wire (3rd prong) which should
be grounding to the computer case will have voltage on it and there
may be high magnetic fields in the house due to ground currents.

This situation is more common than you may think since it goes on for
years without being detected and I found a few % of houses in my
neighborhood suspected of being like this in a survey I once made by
walking around the block with a special meter made for AC magnetic
field detection.  A few houses had high magnetic fields near their
buried water pipes under the public sidewalk => high current in the
pipes.  In fact, highest of all was current flowing in pipes under
California St. by Caltech (in conduit pipes ??).  Somehow, I never got
around to reporting it to Caltech.  I was going to investigate it in
more detail before reporting it but never did since I had borrowed the
flux meter and had to return it.  If you are interested, you might
check this out on California St. west of Wilson.  The National
Electric Code sort of requires wiring that will result in such
currents.  But they added a phrase saying that you may wire it
differently if magnetic field are a problem.  How many electricians
check out the magnetic fields and then rewire, as permitted by code,
if the fields are too high?  Almost none.

There are alleged heath dangers from long-term exposure to such fields.
You can use a search coil connected to a voltmeter (high impedance on
the smallest millivolt range) to measure the rate of flux change thru
the search coil: curl E = -dB/dt per Maxwell which gives volts (for
one turn of the search coil) = webers per sec flux change.  They say
that long term exposure to over .5 microteslas AC is a hazard. DC is
no problem. To get webers you need to know the angular frequency w
(2 pi 60) where B (flux) =k sin wt and dB/dt = ....  and then solve
for k, the max flux and take the square root to get rms.  A tesla is a
weber per sq meter.

But I really doubt that the above is the problem.  The static spark
current should flow directly to ground from the case and not trip
anything.  So it could be that somehow the case is both shorted to a
sensitive component and not adequately grounded.  The spark current in
the "sensitive component" causes the tripping of the power supply.
This happens when static voltage is applied to the case.  Measure AC
voltage on the case wrt ground from a pipe.  See if the meter voltage
changes as you reroute the meter leads => magnetic fields in the room.
Try grounding the case to a water pipe with automotive jumper cables.

What about heat sensors that can shut down a PC if it gets too hot (or
gets stray current in the sensor (or sensor control chip) from static
electricity?  There are also voltage sensors.  Suppose that your heat
or voltages are almost out-of-tolerance and that a sensor is on
the verge of tripping (shutdown).  Then the slightest stray current
could result in shutdown.  I would also search the Internet for
similar cases.

			David Lawyer