[Techtalk] OT but I'm desperate.... Windows....!

[Alvin Goats]

> > > It is true that a monitor is usually the most power hungry component of a
> > > computer, however it is drawing its power from the mains, so doesn't
> > > affect the computer
> 
> > To some extent that is true. However, as I recall from an earlier
> > message, all of your hardware is running off of one outlet. The power is
> > shared between outlets, so a power consuming object on one plug is
> > sapping some power from the other plug.
> 
> Uhm
> 
> no.  sorry.  I think not.

I guess you've never experienced the dimming of lights when a power
hungry item is turned on. The voltage doesn't drop that much, but the
current sharing is disturbed with the surge. If there is sufficient
current draw on an overtaxed line, then a voltage drop will occur. In
most homes, this isn't so noticeable. 

However, my airconditioning is on a seperate line. When the compressor
and blowers kick in for the cooling cycle, there is sufficient current
drain to dim the lights 1 or 2 seconds. Once the electric motors are
started and moving, the current drain is reduced. 


> 
> > This can still lead to a brown out type condition (before tripping
> > circuit breakers/blowing fueses).
> 
> If the voltage drop caused by additional load on the same circuit is
> sufficient to cause problems, then I would suggest that the house wiring
> needs to be looked at in some depth.


Now, I've not said anything about the voltage dropping, just "brown
out". Brown outs severe enough to cause a voltage drop is the classic
form. But a power drop (voltage x current), is also refered to as a
'brown out'.
> 
> Mains voltage is very well regulated, and has a lot of oompf behind it to
> back it up.  The voltage drop across properly installed cables when
> addition load is applied should be pretty much negligible, and the voltage
> droop caused by the additional load should be easily corrected by
> regulating transformers after a short delay.

The voltage is well maintained in most of Europe, Australia, Canada and
the US. Current draws and current spikes are not so well regulated. And
the voltage is RMS, so a spiking condition is averaged out by the root
mean square. For some critical semiconductor manufacturing equipment, it
has been necessary to add line buffering/filtering equipment to
eliminate the millisecond 'brown out' caused by line spikes from such
things as a power drill being used across the street at a construction
site. 


> The time period that the mains voltage droops for doesn't matter anyway,
> since if the power supply is correctly regulating its output rails then it
> should have no effect, in fact a well designed power supply should be able
> to cope with no power coming into it at all for at least half a cycle of
> 50 Hz or 60 Hz, usually much more (the hold over time).  Power supplies
> should be able to cope over a range of voltages at least as far apart as
> 100 and 120 V or 220 and 240V.  I believe (though I'm probably wrong) that
> the acceptable range of voltages to find on a mains outlet in the USA is
> 90-130V, and in the UK/europe 210-250.  There is a percentage range
> defined, though I can't remember it off hand.

Unfortunately, not all power supplies are well designed. For a power
supply that can survive voltage spikes, current surges, brown out
conditions, can still deliver the proper output voltages regardless of
input, requires the expensive military power supplies everyone likes to
complain about the cost of. However, the power supplies for our PC's are
typically a cheaply made, over priced switching power supply of
borderline design. I've had several die and dissected them. Several were
covered in solder flux (which should be removed), capacitor voltage
ratings barely above the regulated voltage, components with large drift
characteristics with temperature, and so on. If you have a good supply,
congradutlations! Many are not. 


Again, the voltage regulation of the power line is well maintained, but
does not account for what is going on after the power is generated. When
management is brow beating you for test failures or line stopages that
you have to put an oscilloscope on the power line to find the cause
of...well, you get to learn what the power companies control and don't
control. So the guy who patched his drill on a buffered and filtered
line caused spikes that wiped me out. It was sporadic since it only
occurred when he started his drill...
  

> As to the original problem, if it crashes every time you go to print, I
> would be rather suspicious.  Though the error pasted in another email
> would seem to suggest that windows is experiencing its equivalent of a
> kernel panic.

This I concur with. Windows has a bad tendancy to do things to it's own
software device drivers. Particularly after some of the other statements
from the victim about the bsod error codes. 

Alvin