Possible Power Supply Problem

[Bob Havens]

I am using a clone PC that my son in law built for me three years ago. It
has a Microstar K7N420 Motherboard and a MS-6373(v1.x) ATX mainboard with
an AMD Athlon XP 1600+ 1.40GHz Processor and 992 MB of RAM. I use Windows
XP-Home Edition Version 5.1.2600 Service Pack 2. I have all the latest
updates. The CMos Battery is the same age as the PC (3 years).

When I press the power button to start the PC the power light will go on
and the hard drive will spin for about 3 seconds and then stop. The power
light remains on. If I then depress the power button for about 5 seconds
the power light goes off. If I then press the power button the above
process repeats itself. After repeating this process several times the
hard drive may or may not continue to run. If it does continue to run the
computer boots up in the normal manner and the PC operates as it should.
This problem started about a month ago and as time progressed it takes more
attempts to get the PC to boot up.

My son in law thinks the power supply is going out. I would like your
input as to the possible cause of this problem and information as to what I
might check to correct the problem.
Thank you,
Bob

 

[w_tom]

Problem could be most anything in a power supply 'system'.
That includes one 'system' component - the power supply. Use
a 3.5 digit multimeter to first learn what is and is not
happening. Get the necessary numbers. Overall concepts are
explained in "Computer doesnt start at all" in
alt.comp.hardware on 10 Jan 2004 at
http://tinyurl.com/2t69q and
"I think my power supply is dead" in alt.comp.hardware on 5
Feb 2004 at
http://www.tinyurl.com/2musa

In your case, first, voltage on the purple wire must be
correct. All voltages must measure within upper 3/4s of specs
provided in those discussions. Then power supply controller
uses green wire to tell power supply to poweron when power
button is pressed. Within a two seconds, voltages on red,
yellow, and orange wires must be in spec. Then gray wire
tells computer that power supply is OK.

You cannot see any of this. Others only assume a power
supply is the entire power supply 'system'. Smart techs get
that meter; discover in only a few minutes if and where a
power supply 'system' has problems.

Without those numbers, then replies here can only be wild
speculation. Get the meter.

 

[Kerry Brown]

Problem could be most anything in a power supply 'system'.
That includes one 'system' component - the power supply. Use
a 3.5 digit multimeter to first learn what is and is not
happening. Get the necessary numbers. Overall concepts are
explained in "Computer doesnt start at all" in
alt.comp.hardware on 10 Jan 2004 at
http://tinyurl.com/2t69q and
"I think my power supply is dead" in alt.comp.hardware on 5
Feb 2004 at
http://www.tinyurl.com/2musa

In your case, first, voltage on the purple wire must be
correct. All voltages must measure within upper 3/4s of specs
provided in those discussions. Then power supply controller
uses green wire to tell power supply to poweron when power
button is pressed. Within a two seconds, voltages on red,
yellow, and orange wires must be in spec. Then gray wire
tells computer that power supply is OK.

You cannot see any of this. Others only assume a power
supply is the entire power supply 'system'. Smart techs get
that meter; discover in only a few minutes if and where a
power supply 'system' has problems.

Without those numbers, then replies here can only be wild
speculation. Get the meter.

I have the meter. It's much faster and easier to just unplug the power
supply and plug in a known good one. At most three plugs on the motherboard.
You don't even have to remove the old power supply to test a different one.
With the meter once you confirm the power supply is dead it's much cheaper
to replace it than try to fix the old one. As a troubleshooting tool a spare
power supply is also cheaper than a good meter.

Kerry

 

[Bob Havens]

My son in law has several extra power supplies that he said he can swap
out to test the PS that I have. He looked at the PC a couple weeks ago and
we decided I would keep using it till it got to the point where it was a
real problem. It is not getting any better. I read the web pages that w
Tom mentioned and it sure sounds like the problem is the PS. If the swap
out shows my PS to be bad I will just get a new one and my son in law will
put it in. I was hoping that he was wrong and the problem would be
something simple that I could fix myself but I don't think this is the
case.
Thanks for your help.
Bob

 

[w_tom]

Swapping power supplies takes far longer than the two
minutes to use a meter. There are more than three wires.
There are other wires to be moved aside. Screws to remove and
replace. Verification to confirm all plugs are properly
seated and properly aligned. Plus swapping adds additional
unknowns into the problem. For example, number one unknown -
is the second power supply really functional? Does it really
provide sufficient power? Is its timing for data exchange
compatible with the rest of the power supply 'system'?

And after swapping power supplies, does one still know if
the power supply controller is functional? No. A marginal
problem can be masked by power supply swapping. Problem
returns later using that new supply.

Generations of experience - from long before PCs even
existed - says use the meter. Meter answers the question -
with certainty - in less than half the time. Good techs never
shotgun a problem as Kerry Brown suggests. Good techs first
learn of the problem. Second step is to fix that problem.
Kerry Brown recommends bypassing step one and going right to
step two. He recommends fixing the problem without ever
learning why the problem first existed. This explains why so
many computer problems are not repaired the first time.

Get and use the meter - and don't look back. The only way
to get additional assistance from the more technically
knowledgeable is to provide numbers from that meter. Kerry
Brown types will simply say, "Try this and try that."

 

[Bob Havens]

Thank you. What you say makes sense. I am sure my son in law has a meter.
I will pass the info on to him since I don't know enough about computers
and electronics to fix it myself.
Bob

 

[Kerry Brown]

Swapping power supplies takes far longer than the two
minutes to use a meter. There are more than three wires.
There are other wires to be moved aside. Screws to remove and
replace. Verification to confirm all plugs are properly
seated and properly aligned. Plus swapping adds additional
unknowns into the problem. For example, number one unknown -
is the second power supply really functional? Does it really
provide sufficient power? Is its timing for data exchange
compatible with the rest of the power supply 'system'?

And after swapping power supplies, does one still know if
the power supply controller is functional? No. A marginal
problem can be masked by power supply swapping. Problem
returns later using that new supply.

Generations of experience - from long before PCs even
existed - says use the meter. Meter answers the question -
with certainty - in less than half the time. Good techs never
shotgun a problem as Kerry Brown suggests. Good techs first
learn of the problem. Second step is to fix that problem.
Kerry Brown recommends bypassing step one and going right to
step two. He recommends fixing the problem without ever
learning why the problem first existed. This explains why so
many computer problems are not repaired the first time.

Get and use the meter - and don't look back. The only way
to get additional assistance from the more technically
knowledgeable is to provide numbers from that meter. Kerry
Brown types will simply say, "Try this and try that."

I am technically knowledgable with 25 year experience repairing pc's. I have
very few returns. It takes less than a minute to hook up a different power
supply. With the meter it takes about twice the time to test for correct
voltages. Uness you have a very good meter with data recording you may as
well just plug in a cheap PCI LED board. To truly diagnose what is wrong you
would have to measure current as well as voltage. In any case all the meter
will tell you is the power supply is not working properly. In today's throw
away economy it is not very cost effective to repair a power supply. It is
cheaper to replace it. Also an inexperienced person with a meter is likely
to measure the wrong thing anyway. A good tech has the knowledge and
experience to realise that sometimes the common sense answer is better than
the technically correct answer. In a classroom your method is right. In the
real world it takes too long and isn't needed. I'm guessing you are an
engineer.

Kerry

 

[w_tom]

The power supply controller - another component of the power
supply 'system' - is located on motherboard.

 

[HillBillyBuddhist]

Also an inexperienced person with a meter is likely to measure the wrong
thing anyway. A good tech has the knowledge and experience to realise that
sometimes the common sense answer is better than the technically correct
answer. In a classroom your method is right. In the real world it takes too
long and isn't needed. I'm guessing you are an engineer.

Kerry

Completely agree.

If I question the viability of just about any component I swap it with a
known good equivalent. Meter testing? Please. To what end? On the rare
occasion that the individual knows how to use a meter *and* interpret the
results *and* the results yield negative, what then? (it was rhetorical)
Answer, swap it out for a known good equivalent. Coulda been more than
halfway there already.

--
D

I'm not an MVP a VIP nor do I have ESP.
I was just trying to help.
Please use your own best judgment before implementing any suggestions or
advice herein.
No warranty is expressed or implied.
Your mileage may vary.
See store for details.

Remove shoes to E-mail.

 

[NobodyMan]

I am using a clone PC that my son in law built for me three years ago. It
has a Microstar K7N420 Motherboard and a MS-6373(v1.x) ATX mainboard with
an AMD Athlon XP 1600+ 1.40GHz Processor and 992 MB of RAM. I use Windows
XP-Home Edition Version 5.1.2600 Service Pack 2. I have all the latest
updates. The CMos Battery is the same age as the PC (3 years).

When I press the power button to start the PC the power light will go on
and the hard drive will spin for about 3 seconds and then stop. The power
light remains on. If I then depress the power button for about 5 seconds
the power light goes off. If I then press the power button the above
process repeats itself. After repeating this process several times the
hard drive may or may not continue to run. If it does continue to run the
computer boots up in the normal manner and the PC operates as it should.
This problem started about a month ago and as time progressed it takes more
attempts to get the PC to boot up.

My son in law thinks the power supply is going out. I would like your
input as to the possible cause of this problem and information as to what I
might check to correct the problem.
Thank you,
Bob

Ask elsewhere. This is not an XP problem.