steve d. podleski wrote:
> Arno.
>
> Thanks for helping.
>
> How can I tell if the PSU fan is bad without having the PSU installed?
>

The fan monitor has a "minimum RPM" value. When the RPMs
drop below the minimum value, the readout goes to zero.

I have a motherboard, where the minimum measurable value is
1800RPM. If the fan spins at 1801 RPM, the readout says 1801.
If the fan spins at 1799, the readout says 0. Your minimum
could be different than this.

This problem is caused by a counter overflow in the monitor chip.
There is a divider in the monitor chip, that can be adjusted, so
that lower fan RPM values can be read. A program like Speedfan
from almico.com may be able to make the lower fan values readable.
(At least, up to the limits of the adjustment in the monitor chip.)

The motherboard BIOS seems to be ill equipped to set this up
properly. The motherboard BIOS can set the threshold quite high.
Speedfan uses an autoranging algorithm, where the monitor chip
is adjusted down, until a value can be seen. If the fan *is*
actually zero, this can still be detected. When the monitor
is dialled as low as possible, it can still properly read out
a zero RPM fan as being zero.

A PSU can be tested, by connecting PS_ON# to COM. It is best to
have a small amount of load on the PSU, so that it can regulate
properly. An old disk drive, for example, draws 5V @ 1A and 12V @ 0.5A
and is better than nothing, as a dummy load. When PS_ON# is connected
to an adjacent COM pin, the fan will start to spin. If the fan
doesn't spin, you'd need a multimeter to figure out what is working
or not working. (The supervisor voltage on the PSU, is called +5VSB.
The power supply cannot run the rest of the outputs, unless the +5VSB
is operating first. If the +5VSB fails or is overloaded, there is
no chance of the other voltages being generated.)

Watch it with the old Dell supplies! They can fry an ordinary computer.
Especially if they have that "extra" connector. Look for Dell pinouts
on the web for more details, before using one.

Paul