Computer won't start

M

Metspitzer

If the fan "twitches", then you know the primary is intact, and something
is defective on the secondary. To "twitch", there has to be some
transformer action, between the primary and secondary. The switching
transistors on the primary, have to run for a short period of time,
to provide enough energy through the transformer, to make the cooling fan
on the CPU to twitch. The power supply has capacitors on the secondary side,
and they'd need to be charged high enough, to cause the fan motor to start
moving. So it's going to take thousands of switching cycles on the
switching transistors (maybe 35 milliseconds worth), to get that far.

*******

With respect to your testing the contacts on the AC side of the
supply, you'd need to understand how the multimeter does its test,
to understand the results.

For example, look at the primary here, and tell me under what
conditions the input will conduct. Hint - two diode drops into
a capacitor. Does the multimeter provide a high enough voltage
to cause two diodes in series to conduct ? You really need a
good understanding of how the multimeter works, before you
can understand the results. Some measurements are worthless,
because they can be ambiguous. (Double hint - just because
it reads open, doesn't mean its open. It's a function of
what the meter is doing. To remove the ambiguity, you
remove the power supply cover, and ohm out the fuse directly
if you can find it.) In this example, there is that 560K
on the primary side in parallel, and perhaps you can see
that one.

http://www.pavouk.org/hw/en_atxps.html

An ohmmeter is "a voltmeter, in parallel with a
current source with a limited compliance voltage".
And in English, it means the ohmmeter has two "corner
conditions". Into a dead short, the current source won't
put more than the current limit value into the circuit
(like, 1.000 milliamps of test current or 100.0 microamps
of test current). And when open circuit, the current source
voltage level won't rise to higher than the compliance
voltage level. (Perhaps that is 1 volt on a low power ohms range,
for example. My old Simpson analog meter, could apply as much
as 9 volts under similar conditions. My Simpson is the one
that's going to "find the fuse" for me.)

http://www.allaboutcircuits.com/vol_3/chpt_3/2.html

"Beware that some digital multimeters equipped with a "diode check"
function may output a very low test voltage (less than 0.3 volts)
when set to the regular "resistance" (ohm) function: too low to
fully collapse the depletion region of a PN junction. The philosophy
here is that the "diode check" function is to be used for testing
semiconductor devices, and the "resistance" function for anything
else. By using a very low test voltage to measure resistance,
it is easier for a technician to measure the resistance of
non-semiconductor components connected to semiconductor components,
since the semiconductor component junctions will not become
forward-biased with such low voltages."

That's an explanation of so-called "low power ohms" versus
"high power ohms" type measurements. The difference is the first
one, doesn't use a compliance voltage value high enough to turn
on diodes.

A great learning experience, is to connect two multimeters
together. For example, put one of them on "ohms" and the
other on "volts". Or, put one on "ohms" and the other
on "current" (after changing the holes the test leads go into).
The things you read out, should help you better understand
the stimulus the meter is applying to a circuit. And at
least explain, how two meters might reach a different
conclusion.

Paul

I didn't know there was gonna be a test.

Thanks Paul
 
P

Paul

Metspitzer said:
I didn't know there was gonna be a test.

Thanks Paul

Electronics is really crazy that way. You'd be
surprised how many measurements you can do, which
are perfectly useless :)

Paul
 
P

PC Guy

Metspitzer said:
The correct answer was Power Supply. When I changed it with
another power supply, it started.

A new or different power supply might have appeared to fix the
situation, but if I were you I'd check the large electrolytic capacitors
on the motherboard.

These will be the caps located near the CPU. Anywhere from 6 to 12 of
them depending on the motherboard.

I've replaced the caps on about a dozen motherboards over the past
year. Also have fixed some OTA digital TV convertor boxes with the same
problem.

Various consumer electronic products made between 2005 through 2007 used
capacitors sourced from a particular factory in china, and those
products are failing in large numbers in the past few years.

Samsung is on the hook to fix or replace upwards of 7 million TV's.

The problems with motherboards tend to show up as really unstable
operation, crashing and strange blue and black-screen errors upon
startup.

If you're handy with a soldering iron you can fix these yourself. Buy
new capacitors from Digikey for about $1.50 each.

The caps will have some score marks on the top (will look like an X or
K). Look to see if the metal has opened up along the score marks, or if
the top seems to bulge a little.

These caps will be about 1/2" in diameter and 3/4" tall. 1000uf and
1500uf, 6v or 10v rating.
 
M

Metspitzer

Bump on the CPU fan would indicate that the power supply is doing a
crowbar shutdown because something is drawing too much current.

This can be on the motherboard from shorted capacitor(s), internal
shorts on CPU or graphics chip, or shorted line drivers on the serial
lines. Or just a failure in the power supply itself or on an optical
drive.

Generally have found that failures protected by crowbar have left the
hard drives unaffected.

YMMV
I did try another PS and it did not fail. Woudn't the problems you
mention also effect the replacment PS?

Thanks
 
M

Metspitzer

i had a very similar situation a few years back. It turned out to be
the power supply, but at the same time, it turns out that I had a huge
load of viruses which seemed to affect the computer. If I were you, I
would also try to switch out the hard disks you have and clear the BIOS
if your Mobo has that capability. My uncle ran a bunch of diagnostics
on the system, and it turned out that the BIOS was rewritten by a virus.
Ever since, I have gotten Norton's latest software, hell or high water.
I will do that.
Thanks
 
L

Loren Pechtel

I did try another PS and it did not fail. Woudn't the problems you
mention also effect the replacment PS?

Yes. All of what he mentioned would cause the same failure on any
power supply.
 

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments. After that, you can post your question and our members will help you out.

Ask a Question

Top