Freeze-ups. Is power supply the only remaining possibility?

  • Thread starter Thread starter Not Here
  • Start date Start date
I do not currently have access to a multimeter, but I did briefly, and
noted these values under very light load.
Or-3.32 Or 3.32 Red 4.94 Red 4.94 Gray 4.67 Prpl 4.97 Yellow 12.5
Or -12 Blu -12 Grn .03 White -5.24 Red 4.94 Red 4.94
Click to expand...

Appreciate my problem. Numbers are from two different sources - one
of unknown quality and the other without loading. If both numbers are
accurate, then a major problem exists with the Yellow wire - varying
from 11.7 to 12.5. Yellow wire voltages should not change that much.
More useful conclusions possible with better numbers.

Not having a meter is like saying you don't have a screwdriver.
Both tools are that ubiquitous as to be sold even in Sears, Wal-Mart,
Radio Shack, Tru-Value Hardware, and K-mart. 3.5 digit multimeter is
a ubiquitous and necessary tool - and so inexpensive.

No need to measure every wire of same color. A reference to the
phrase "voltages on any one of red, orange, yellow..".

Green and gray wires (that don't get increasing loads) are more than
sufficient confirming what was suspected: power supply controller and
power supply are communicating good information properly when loads on
other voltages are small. With other useful numbers, then power
supply is either suspect or exonerated - definitively.
 
Appreciate my problem. Numbers are from two different sources - one
of unknown quality and the other without loading. If both numbers are
accurate, then a major problem exists with the Yellow wire - varying
from 11.7 to 12.5. Yellow wire voltages should not change that much.
More useful conclusions possible with better numbers.

Not having a meter is like saying you don't have a screwdriver.
Both tools are that ubiquitous as to be sold even in Sears, Wal-Mart,
Radio Shack, Tru-Value Hardware, and K-mart. 3.5 digit multimeter is
a ubiquitous and necessary tool - and so inexpensive.

No need to measure every wire of same color. A reference to the
phrase "voltages on any one of red, orange, yellow..".

Green and gray wires (that don't get increasing loads) are more than
sufficient confirming what was suspected: power supply controller and
power supply are communicating good information properly when loads on
other voltages are small. With other useful numbers, then power
supply is either suspect or exonerated - definitively.
Click to expand...


Here are the readings I have using a multimeter, very consistent
whether idling in BIOS setup or under heavy load--Orthos, virus scan
and playing 2 movies from 2 DVD drives.

Red - 5.00 Orange - 3.35 Yellow - 12.58 Purple - 5.02 Green 0.04
Gray 4.72

Graham
 
Here are the readings I have using a multimeter, very consistent
whether idling in BIOS setup or under heavy load--Orthos, virus scan
and playing 2 movies from 2 DVD drives.

Red - 5.00 Orange - 3.35 Yellow - 12.58 Purple - 5.02 Green 0.04
Gray 4.72
Click to expand...

I am troubled that previous numbers were so different. Ignoring
all other data, these numbers say the power supply is well and
healthy; not a reason for hardware slowdown or crashes. Move on to
other suspects.

Looking back at previous discussions. A freeze is reported. By
freeze, do you mean a crash? IOW with Task Manager loaded, does even
Task Manager stop reporting data or does the machine slow down so
greatly and Task Manager shows a massive consumption of CPU time?

You replaced some electrolytic caps. Did you replace them all in
that area? Replace all of same type even if some did not look
buldging? Most failed parts have no visual indication. Some failing
electrolytic caps iimplies all of the same family are probably just as
defective.

BTW, some multimeters will also measure caps. If you do replace
those others, then measure the old ones - to learn if that would
explain freezing.

I don't remember if other standard data sources were examined such
as system (event) logs and Device Manager for hardware conflicts.

And finally, OS crashes are typically limited to a subset that
includes Sound Card, Video Processor, CPU, memory, and motherboard
components only related to those functions. You can ignore or
eliminate hard drive, keyboard, mouse, etc from the suspect list.
Heat is a diagnostic tool. When heated on highest setting, then
diagnostics for those suspects may report a hardware problem. Heat
can make an intermittent hardware problem into a hard, detectable
failure which is why testing a computer in a 100 degree F room with
diagnostics is so informative. Best testing is conducted with
semiconductors uncomfortable to touch but not hot enough to leave
skin.
 
I am troubled that previous numbers were so different. Ignoring
all other data, these numbers say the power supply is well and
healthy; not a reason for hardware slowdown or crashes. Move on to
other suspects.
Click to expand...

My earlier post was reporting voltages as seen by software in Windows.
Now using the multimeter I see that the software doesn't give the same
values as the meter.
Looking back at previous discussions. A freeze is reported. By
freeze, do you mean a crash? IOW with Task Manager loaded, does even
Task Manager stop reporting data or does the machine slow down so
greatly and Task Manager shows a massive consumption of CPU time?
Click to expand...

By freeze I mean the display is frozen, clock not updating etc.
Keyboard and mouse not responding. Need to reboot.
You replaced some electrolytic caps. Did you replace them all in
that area? Replace all of same type even if some did not look
buldging? Most failed parts have no visual indication. Some failing
electrolytic caps iimplies all of the same family are probably just as
defective.
Click to expand...

I did replace all three caps in the cluster which had 2 visibly
damaged caps.
BTW, some multimeters will also measure caps. If you do replace
those others, then measure the old ones - to learn if that would
explain freezing.

I don't remember if other standard data sources were examined such
as system (event) logs and Device Manager for hardware conflicts.
Click to expand...

No conflicts in Device Manager. I'm researching the event logs.
And finally, OS crashes are typically limited to a subset that
includes Sound Card, Video Processor, CPU, memory, and motherboard
components only related to those functions. You can ignore or
eliminate hard drive, keyboard, mouse, etc from the suspect list.
Heat is a diagnostic tool. When heated on highest setting, then
diagnostics for those suspects may report a hardware problem. Heat
can make an intermittent hardware problem into a hard, detectable
failure which is why testing a computer in a 100 degree F room with
diagnostics is so informative. Best testing is conducted with
semiconductors uncomfortable to touch but not hot enough to leave skin.
Click to expand...

I have replaced the heatsink fan just now. The temperature is still
being reported as 70C but haven't had a freeze after 1.5 hours of hard
use but that's probably coincidence.


Graham
 
Here are the readings I have using a multimeter, very consistent
whether idling in BIOS setup or under heavy load--Orthos, virus scan
and playing 2 movies from 2 DVD drives.

Red - 5.00 Orange - 3.35 Yellow - 12.58 Purple - 5.02 Green 0.04
Gray 4.72
Click to expand...

I am troubled that previous numbers were so different. Ignoring
all other data, these numbers say the power supply is well and
healthy; not a reason for hardware slowdown or crashes. Move on to
other suspects.

Looking back at previous discussions. A freeze is reported. By
freeze, do you mean a crash? IOW with Task Manager loaded, does even
Task Manager stop reporting data or does the machine slow down so
greatly and Task Manager shows a massive consumption of CPU time?

You replaced some electrolytic caps. Did you replace them all in
that area? Replace all of same type even if some did not look
buldging? Most failed parts have no visual indication. Some failing
electrolytic caps iimplies all of the same family are probably just as
defective.

BTW, some multimeters will also measure caps. If you do replace
those others, then measure the old ones - to learn if that would
explain freezing.

I don't remember if other standard data sources were examined such
as system (event) logs and Device Manager for hardware conflicts.

And finally, OS crashes are typically limited to a subset that
includes Sound Card, Video Processor, CPU, memory, and motherboard
components only related to those functions. You can ignore or
eliminate hard drive, keyboard, mouse, etc from the suspect list.
Heat is a diagnostic tool. When heated on highest setting, then
diagnostics for those suspects may report a hardware problem. Heat
can make an intermittent hardware problem into a hard, detectable
failure which is why testing a computer in a 100 degree F room with
diagnostics is so informative. Best testing is conducted with
semiconductors uncomfortable to touch but not hot enough to leave
skin.
 
As demonstrated, BIOS voltage numbers (that were read in Windows)
had little relationship to actual numbers. It is called a motherboard
*monitor* - to detect changes. Comparing those numbers to those from
the multimeter, and then doing appropriate math, adjust 'alarm'
threshold unique for that motherboard. Default 'alarm' number is
adjusted so that low voltages alarm occurs when actual voltage is
under 3.23, 4.87, or 11.7. Use the meter to calibrate that
motherboard voltage monitor.

Temperature (ie a new heatsink) is not a solution. Intentionally
increase temperature 30+ degrees F on individual parts so that
diagnostics can report something useful. Others (often with only A+
Certified training) want to fix problems by curing temperature - ie a
new heatsink. We don't want to do that. We want to make the defect
obvious (more heat) and then use something that can identify that
defect (diagnostics). We want computer parts even warmer. Heat is a
diagnostic - not a problem to be solved.

As I recall, this is an Intel CPU. High temperature only causes
throttling - not shutdown. CPU might then have too much work and
'appear' frozen.

A 'frozen' screen can occur due to a crash or due to something
consuming too much CPU time - locking out other processes. Both will
make keyboard and mouse appear frozen. But information on Task
Manager has higher priority - may indicate the difference. We must
determine a crash verses an overloaded CPU. If CPU has too much work,
then high priority tasks (ie Task Manager) will appear very slow but
will still make some data updates.

That short list of suspects is for a crash. That short list is only
part of a longer list for too much CPU time. We must know if it is a
crash or is a CPU with too much work.

If it is a crash, then you want the CPU to be even hotter (also true
for the other suspects on that short list) when executing
diagnostics. Diagnostics run in a 70 degree F room often will not
report the failure.

I am repeatedly bothered by the 70 degree C number from CPU. I was
hoping others could better define where that number was coming from.
It should be from a diode inside the CPU and measured by a peripheral
chip. I would hope an Intel CPU diagnostic (from Intel or third
party) could provided better indications. But again, this is why some
computer manufacturers provide comprehensive hardware diagnostics.

If heating that CPU with a hairdryer, do temperature numbers
increase in but a minute? It should if hardware is measuring
temperature properly. Hairdryer on highest heat applied and removed
should appear in the CPU temperature monitor.

Your problem still exists. Replaced were some visible defective
capacitors. But hardware apparently still failed. Problem remains as
long as we do not first identify the failure. Best way to find that
failure is to heat everything to uncomfortable temperatures that all
good semiconductors love and that bad techs want to cure. Bad
electronics then would be easier to locate especially with
diagnostics. Again, we don't cure a symptom by fixing heat. We use
heat as one diagnostic tool to find defective electronics.
 
Back
Top