Intel cooling fan shuts down power supply
 

George Orwell wrote:
A friend gave me a PC what did not work. It has Pentium
socket 478 and Intel cooler. When switched on, it shuts
down after 1 or 2 seconds. I started removing things, and
it was down to motherboard still stopping it. I put an ATX
power supply tester on instead, and the power supply stays
on. But this is no load. Out of curiosity, I connected the
CPU fan to the power supply and bingo - it stopped. I had
a few spare Intel fans and connected those instead. The
power supply stays on. So I measured the resistance across
the black and yellow wires of the killer fan and it reads
open circuit (well greater than 2 megohm). All the good
fans read 1800 to 2000 ohms with same meter.
Weird eh? A high resistance drags down PS!!!

No, a switching power supply controls voltage by limiting the "on time" power is
sent to the filter system to become (more or less) DC. As load falls (high
resistance) eventually the switching time becomes too small, the output waveform
changes from a ripply DC to flat with spikes, and it shuts down before the
crappy DC breaks something. Connect any load, and the DC gets better.

NOTE: the symptoms sound like a short rather than an open, but many power
supplies will not work feeding just a motherboard, it doesn't draw enough load
at some voltages. A moderately good explanation is at:
http://en.wikipedia.org/wiki/Switchi...ly#Explanation
Again, I am taking your high resistance measurement at face value, even though
you would expect the system to start, then the CPU to shut down with no cooling.