New Gigabyte Machine is dead!

[roger]

Boy! Do I need some help.

My prior posts will show that I have just built a system consisting
of:

1) MOBO - Gigabyte GA-F@A88X-D3H FM2
2) CPU - AMD A10-5800K Quad
3) RAM - Corsair Vengeance Pro 2X4GB
4) Case - Rosewill Galaxy 3
5) PSU - Coolermaster GX 650W 80 Plus
6) W7

It has been running nicely for a few weeks now. But this morning,
when I powered-up, nothing!

So far I have:
1) disconnected the power switch connector from the mobo pins, and
tried to 'trigger' the machine by the usual shorting of the two
power-switch pins. Nothing!
2) pulled a PSU from an old machine and connected its 24-pin power
connector to the Gigabyte mobo, but discovered that that PSU doesn't
have a connector for the Gigabyte mobo 8-pin so-called 'ATX-12V'
header. Nonetheless, I shorted the Gigabyte pins again to see if any
response. Nothing!
3) connected just the mobo power connector from the new PSU to the
24-pin power header on the old machine mobo, and tried to power up the
old machine. Nothing!

So now I am sitting here, wondering what this incompetent pretend nerd
can do next. I need to determine what component in the new machine
has gone bad so I can return it. I am hoping its the PSU. Can
someone suggest something I can try to prove or eliminate that? Any
help wud be appreciated.

Thanks

Big Fred
4)

 

[roger]

I repeated my testing that I just described, and now I get power to
the old machine from the new Gigabyte's PSU. Ergo, I guess the PSU is
okay. Now I need to ask - how can I determine if my problem is the
AMD CPU or Gigabyte MOBO? Or the Corsair Ram for that matter? They
came from different sources.

Thanks

Big Fred

 

[Paul]

I repeated my testing that I just described, and now I get power to
the old machine from the new Gigabyte's PSU. Ergo, I guess the PSU is
okay. Now I need to ask - how can I determine if my problem is the
AMD CPU or Gigabyte MOBO? Or the Corsair Ram for that matter? They
came from different sources.

Thanks

Big Fred

It's the motherboard. Or rather, it could be.

*******

The steps in the process are:

1) Plug in power supply with switch off. Power supply makes
no power at all.

2) Flip switch to "ON", on the back of the computer.
Power supply makes +5VSB (powers USB devices, motherboard
wake on LAN logic, some motherboard chipset stuff).
If +5VSB is overloaded, then you'll have zero volts
on the +5VSB rail. The fans don't spin, and the power
supply relies on convection cooling while making +5VSB
for you. So in this state, the fans aren't telling you
anything.

3) When the motherboard has a working +5VSB source, it
begins to "listen" for wakeup events. The motherboard
monitors the front-mounted power button. A momentary
contact on the front switch, causes a logic 0 pulse
which is latched on the motherboard. That's why you
can poke the two header pins on the FP_header with
a screwdriver and get the motherboard to start. The
pulse is latched.

4) The motherboard then drives the PS_ON# signal to the power
supply. The fans begin to spin. This is independent of
whether a CPU is even plugged in. Unless something extraordinary
happens, you should at least be able to get a "twitch" out
of the system fans, when pushing the front power button.
And that's because the motherboard latches the pulse on
the front button, and turns it into a steady level on
PS_ON#.

So things we would suspect:

1) Too much loading on +5VSB, causes the power supply to
shut off +5VSB. Remove external USB devices, coffee mug heater,
iPad charging, USB reading lamp etc.

2) Motherboard is not getting pulse from front power switch. Substitute
another switch for testing. You can use the computer case
reset button, as a temporary replacement for the power button.
(But you already checked for this, by using the screwdriver
tip on PWR button pins, as a test.)

3) Motherboard is getting pulse from the front power switch,
but is refusing to drive PS_ON# to logic 0. This sometimes
happens because the open collector transistor isn't working
on the motherboard. Or, the power supply can fail in such a
way, as to "overpower" the transistor on the motherboard, preventing
a logic low from being asserted. That would be "too strong
of a pullup on the power supply". You checked for that,
by swapping power supplies, and seeing the same symptoms.

The fact you've tried another power supply, suggests the
motherboard is at fault. If you were the tech in a shop, you'd
be using your multimeter right now, to verify +5VSB, and also
to look for logic 1 and logic 0 levels on PS_ON# (to see if
the motherboard responded in a useful way). Both those
signals are on the main power cable. You can stick a meter probe
between the ATX PSU wire and plastic connector shroud, where it
plugs into the motherboard. You can touch metal inside the
main connector, while the system is running, to take
measurements.

The PSU can be overloaded two ways. If you overload +5VSB
(draw more than 2 or 3 amps), the power supply just shuts
off the +5VSB. There is no visible or audible indication
of a problem if that happens. When you use an Asus motherboard,
a visible symptom is the green LED is not lighted on the
motherboard surface. (The Asus green LED is tied directly
to +5VSB for monitoring.) Other brands, don't waste $0.10 on a
green LED, so you need to use a multimeter to verify
+5VSB is present. My Asrock doesn't have the LED.
Alternately, you can use a USB reading lamp LED, to verify
that some +5VSB is present. But bottom line is, a multimeter
can tell you exactly what the voltage is on +5VSB.

If you overload the +3.3V/5V/12V main rails, the power
supply has the ability to ignore overloads for the first
35 milliseconds of operation. That time exists, to give
the capacitors on the system time to charge up. The result is,
the fans on the computer may twitch for the 35 milliseconds
that the power supply allowed output to appear. So if you
push the front power button, and you see the fan blade
"twitch", that tells you there is an overload on
3.3V/5V/12V. And you might use your clamp-on DC ammeter
with peak detection enabled, to verify the condition
and figure out which rail is shorted.

Your symptoms say motherboard, but I'd be using the
multimeter, set to volts, to review the status signals
on the main power cable. Power_good, PS_ON#, +5VSB,
are all of interest, both before and after you press
the front button.

*******

If you had a bare motherboard, fresh out of the box,
you can do the following test case.

1) Connect main ATX power cable to motherboard.
2) Connect POWER switch from computer case, to PWR
pins on FP_header.
3) Switch on the power supply at the back.
4) Push the power button.
5) Note that the PSU fan starts to spin.
6) Push the power button again, and the power
supply should switch off. Fan stops spinning.
7) Switch off supply at the back, when test is complete.

The motherboard doesn't need components connected to it.
The CPU socket has VID pin encoding on it. And if
the CPU is missing, the VID pins go to the all 1's
state, which tells the VCore regulator to make
zero volts of output. And that keeps VCore switched
off automatically. So the motherboard can still be
operated with no CPU and no RAM, and be used as a
simple means to get the PSU fan spinning. As a means
to test that the power supply is listening to PS_ON#.

That would be a complete test of the mechanism.

If you just want to test the power supply all
by itself (not connected to computer), you jumper
PS_ON# to COM. That's the green wire as far as I
can remember. You can get the wire colors from an
offical ATX PSU spec from formfactors.org. Jumpering
PS_ON# to COM, is a way to send a logic 0 to
the power supply ("switch on").

A final word of warning. There are some power supplies,
that like a minimal amount of loading when you operate
them. The voltages won't be in spec, unless the
minimum loading is present. I don't happen to own
any supplies like that, so I don't need to put a
couple hard drives as dummy loads on the PSU. But
it's something to keep in mind with older power
supplies. It would be a particular problem
for a $20 supply without OVP, as the outputs could
rise slightly above the normal 3.3/5/12V levels.
Such supplies may have two rows of rating numbers
on the label, with one row being the "minimal loading"
numbers. That would be a hint, that when lightly loaded
(testing PSU all by itself), some dummy loading should
be attached.

Paul

 

[roger]

It's the motherboard. Or rather, it could be.

*******

The steps in the process are:

1) Plug in power supply with switch off. Power supply makes
no power at all.

2) Flip switch to "ON", on the back of the computer.
Power supply makes +5VSB (powers USB devices, motherboard
wake on LAN logic, some motherboard chipset stuff).
If +5VSB is overloaded, then you'll have zero volts
on the +5VSB rail. The fans don't spin, and the power
supply relies on convection cooling while making +5VSB
for you. So in this state, the fans aren't telling you
anything.

3) When the motherboard has a working +5VSB source, it
begins to "listen" for wakeup events. The motherboard
monitors the front-mounted power button. A momentary
contact on the front switch, causes a logic 0 pulse
which is latched on the motherboard. That's why you
can poke the two header pins on the FP_header with
a screwdriver and get the motherboard to start. The
pulse is latched.

4) The motherboard then drives the PS_ON# signal to the power
supply. The fans begin to spin. This is independent of
whether a CPU is even plugged in. Unless something extraordinary
happens, you should at least be able to get a "twitch" out
of the system fans, when pushing the front power button.
And that's because the motherboard latches the pulse on
the front button, and turns it into a steady level on
PS_ON#.

So things we would suspect:

1) Too much loading on +5VSB, causes the power supply to
shut off +5VSB. Remove external USB devices, coffee mug heater,
iPad charging, USB reading lamp etc.

2) Motherboard is not getting pulse from front power switch. Substitute
another switch for testing. You can use the computer case
reset button, as a temporary replacement for the power button.
(But you already checked for this, by using the screwdriver
tip on PWR button pins, as a test.)

3) Motherboard is getting pulse from the front power switch,
but is refusing to drive PS_ON# to logic 0. This sometimes
happens because the open collector transistor isn't working
on the motherboard. Or, the power supply can fail in such a
way, as to "overpower" the transistor on the motherboard, preventing
a logic low from being asserted. That would be "too strong
of a pullup on the power supply". You checked for that,
by swapping power supplies, and seeing the same symptoms.

The fact you've tried another power supply, suggests the
motherboard is at fault. If you were the tech in a shop, you'd
be using your multimeter right now, to verify +5VSB, and also
to look for logic 1 and logic 0 levels on PS_ON# (to see if
the motherboard responded in a useful way). Both those
signals are on the main power cable. You can stick a meter probe
between the ATX PSU wire and plastic connector shroud, where it
plugs into the motherboard. You can touch metal inside the
main connector, while the system is running, to take
measurements.

The PSU can be overloaded two ways. If you overload +5VSB
(draw more than 2 or 3 amps), the power supply just shuts
off the +5VSB. There is no visible or audible indication
of a problem if that happens. When you use an Asus motherboard,
a visible symptom is the green LED is not lighted on the
motherboard surface. (The Asus green LED is tied directly
to +5VSB for monitoring.) Other brands, don't waste $0.10 on a
green LED, so you need to use a multimeter to verify
+5VSB is present. My Asrock doesn't have the LED.
Alternately, you can use a USB reading lamp LED, to verify
that some +5VSB is present. But bottom line is, a multimeter
can tell you exactly what the voltage is on +5VSB.

If you overload the +3.3V/5V/12V main rails, the power
supply has the ability to ignore overloads for the first
35 milliseconds of operation. That time exists, to give
the capacitors on the system time to charge up. The result is,
the fans on the computer may twitch for the 35 milliseconds
that the power supply allowed output to appear. So if you
push the front power button, and you see the fan blade
"twitch", that tells you there is an overload on
3.3V/5V/12V. And you might use your clamp-on DC ammeter
with peak detection enabled, to verify the condition
and figure out which rail is shorted.

Your symptoms say motherboard, but I'd be using the
multimeter, set to volts, to review the status signals
on the main power cable. Power_good, PS_ON#, +5VSB,
are all of interest, both before and after you press
the front button.

*******

If you had a bare motherboard, fresh out of the box,
you can do the following test case.

I did.

1) Connect main ATX power cable to motherboard.
2) Connect POWER switch from computer case, to PWR
pins on FP_header.
3) Switch on the power supply at the back.
4) Push the power button.
5) Note that the PSU fan starts to spin.
6) Push the power button again, and the power
supply should switch off. Fan stops spinning.
7) Switch off supply at the back, when test is complete.

As I said in my second post, the PSU seems okay to me because it fired
up a second computer I have. However, the PSU from that computer did
not fire up the new computer. I tried your above sequence. I see
no fan twitching at all. Not on three case fans not the CPU fan.

The motherboard doesn't need components connected to it.
The CPU socket has VID pin encoding on it. And if
the CPU is missing, the VID pins go to the all 1's
state, which tells the VCore regulator to make
zero volts of output. And that keeps VCore switched
off automatically. So the motherboard can still be
operated with no CPU and no RAM, and be used as a
simple means to get the PSU fan spinning. As a means
to test that the power supply is listening to PS_ON#.

That would be a complete test of the mechanism.

If you just want to test the power supply all
by itself (not connected to computer), you jumper
PS_ON# to COM. That's the green wire as far as I
can remember. You can get the wire colors from an
offical ATX PSU spec from formfactors.org. Jumpering
PS_ON# to COM, is a way to send a logic 0 to
the power supply ("switch on").

I'll remember that. But I really do not think the PSU is bad. I wish
it were.

A final word of warning. There are some power supplies,
that like a minimal amount of loading when you operate
them. The voltages won't be in spec, unless the
minimum loading is present. I don't happen to own
any supplies like that, so I don't need to put a
couple hard drives as dummy loads on the PSU. But
it's something to keep in mind with older power
supplies. It would be a particular problem
for a $20 supply without OVP, as the outputs could
rise slightly above the normal 3.3/5/12V levels.
Such supplies may have two rows of rating numbers
on the label, with one row being the "minimal loading"
numbers. That would be a hint, that when lightly loaded
(testing PSU all by itself), some dummy loading should
be attached.

This was a brand new 650W'er.

Paul

Thanks

Big Fred

 

[Paul]

(e-mail address removed) wrote:

I'll remember that. But I really do not think the PSU is bad. I wish
it were.

Big Fred

There is one final thing you can do. But it's dangerous
for the computer, for its long term health.

If you jumper PS_ON# to COM while the system is
powered, the power supply *will* then start. Since
PS_ON# uses open collector logic, it's safe to
connect PS_ON# to COM. The components are not
stressed by doing so. This used to be called
"wired-OR" logic back in ECL days.

If you do that, you lose:

1) Ability to turn on with front power button.
System is always running, until you switch off at the back.
2) Ability to turn off with front power button.
3) Ability of system to protect against overheat on THERMTRIP.
If your CPU fan stops, the CPU could overheat.
4) System in wonky state, because it is powered when
the system doesn't think it is powered. To return
sanity to the system, use your RESET button.

So if you were on a desert island, could not
send back the motherboard for warranty repair,
could not buy a new motherboard, using the
PS_ON# to COM jumper trick, can give you a
running PC. That's a means of correcting for
a burned out PS_ON# transistor on the motherboard.

To implement such a change, with minimal hacking
to valuable components, buy a 24 pin ATX PSU extension
cable. You can get them for around $8 or so, and
they're a foot long. Strip the insulation off
the green wire and a black wire. And that's where
you can run a connection (or a SPST switch) from
PS_ON# to COM.

(Page 37, green to black - not all ATX 24 pin extension
cables use standard coloring... be careful)

http://www.formfactors.org/developer/specs/ATX12V_PSDG_2_2_public_br2.pdf

( Non-standard coloring! Count pins, to find the right ones. )

http://mt.aa-img.com/images/00/94/10094_525348_TCB-10094.jpg

Paul

 

[Ken]

I did.


As I said in my second post, the PSU seems okay to me because it fired
up a second computer I have. However, the PSU from that computer did
not fire up the new computer. I tried your above sequence. I see
no fan twitching at all. Not on three case fans not the CPU fan.




I'll remember that. But I really do not think the PSU is bad. I wish
it were.

This was a brand new 650W'er.


Thanks

Big Fred

Remove and clean your RAM. It would be even better if you could test
the RAM in another MB. If I am not mistaken, even the simplest of
actions from the MB require a working RAM and processor.

 

[roger]

Paul -

You said:

The motherboard doesn't need components connected to it.
The CPU socket has VID pin encoding on it. And if
the CPU is missing, the VID pins go to the all 1's
state, which tells the VCore regulator to make
zero volts of output. And that keeps VCore switched
off automatically. So the motherboard can still be
operated with no CPU and no RAM, and be used as a
simple means to get the PSU fan spinning. As a means
to test that the power supply is listening to PS_ON#.

That would be a complete test of the mechanism.


So, I pulled the CPU, and re-supplied power to the PSU. The case fans
all run! I guess that means the problem is the CPU, eh?

Thanks

Big Fred

 

[roger]

What on earth is going on?

I had removed both the CPU and its fan before, and thought - maybe its
the fan. So I replaced the CPU without the fan and tried the fan
trick. Same thing - the case fans ran. So then I replaced the fan
and tried again - the machine is now working just fine. So now I
don't know what to do. Except give both the CPU and MOBO back to the
vendor of course.

Any opinion?

Thanks

Big Fred

 

[Paul]

Remove and clean your RAM. It would be even better if you could test
the RAM in another MB. If I am not mistaken, even the simplest of
actions from the MB require a working RAM and processor.

But it's not switching on. The RAM can't do anything at this
point, because 3.3V/5V/12V are unavailable. If the fans don't
spin on a system, there's no power to speak of.

But your idea got me thinking about another potential issue.
That's a flat CMOS battery. Some motherboards won't work,
if the CMOS battery is down to zero volts. That really shouldn't
stop a motherboard, but there are models which are known for this.
Almost as if the SuperI/O (with hardware monitor and VBAT input),
has some say in the matter (gates off PS_ON#).

Paul

 

[Paul]

What on earth is going on?

I had removed both the CPU and its fan before, and thought - maybe its
the fan. So I replaced the CPU without the fan and tried the fan
trick. Same thing - the case fans ran. So then I replaced the fan
and tried again - the machine is now working just fine. So now I
don't know what to do. Except give both the CPU and MOBO back to the
vendor of course.

Any opinion?

Thanks

Big Fred

All I can say at this point is, good work on the diagnosis

I'm not aware of anything that would cause those symptoms.
You changed power supplies, leaving plenty of opportunity
for any latched up circuits to clear. If that was the
kind of fault present, installing the second PSU should
have worked immediately.

There really should not be any "control path", between the
CPU socket and the PS_ON#. You should be able to unplug the
CPU, and still use the front power button to turn the PS_ON#
signal on and off.

So I don't really understand, how a bad contact on the CPU
could do any of this.

The only thing left, is the idea I put in a post a couple
minutes ago, which is a flat CMOS battery (CR2032) might be
able to keep the motherboard in an off state. Certain brands
of motherboards do that, and maybe it's the way the SuperI/O works
in those cases.

It could be, that the CR2032 is marginal (below about 2.4V),
and you may find the system runs one day and stops running
the next. Once it's started, it'll probably keep running.
But if you switch it off at night, if the CR2032 is flat,
it might not start the next morning.

I prefer to test those coin cells in-circuit with my multimeter.
Because it's hard to get them out of the socket without
breaking something.

Put the multimeter on volts. Connect the black lead to
a shiny conductive point on the computer chassis (an
I/O screw will do). Touch the red lead from the multimeter
to the top of the battery. A new battery reads >3.0V.
A marginal battery, is anything < 2.4V. Some motherboards
won't start, once the battery gets down to zero volts.
There is nothing in any documentation, to explain why
that would be the case. Perhaps a sneak path through the
SuperI/O, where it monitors VBAT. Some SuperI/O chips
have a VBAT input, which is one of the inputs to the
analog to digital converter of the hardware monitor portion.

*******

I don't know if I'd be in a rush to return the hardware
just yet. You still don't know for sure, whether it's
something flaky in the install (shorting to bottom of
motherboard), or it's an actual motherboard hardware
failure. The fact that you got it started, puts
a whole new complexion on the issue.

Paul

 

[roger]

All I can say at this point is, good work on the diagnosis

I'm not aware of anything that would cause those symptoms.
You changed power supplies, leaving plenty of opportunity
for any latched up circuits to clear. If that was the
kind of fault present, installing the second PSU should
have worked immediately.

There really should not be any "control path", between the
CPU socket and the PS_ON#. You should be able to unplug the
CPU, and still use the front power button to turn the PS_ON#
signal on and off.

So I don't really understand, how a bad contact on the CPU
could do any of this.

The only thing left, is the idea I put in a post a couple
minutes ago, which is a flat CMOS battery (CR2032) might be
able to keep the motherboard in an off state. Certain brands
of motherboards do that, and maybe it's the way the SuperI/O works
in those cases.

It could be, that the CR2032 is marginal (below about 2.4V),
and you may find the system runs one day and stops running
the next. Once it's started, it'll probably keep running.
But if you switch it off at night, if the CR2032 is flat,
it might not start the next morning.

I prefer to test those coin cells in-circuit with my multimeter.
Because it's hard to get them out of the socket without
breaking something.

Put the multimeter on volts. Connect the black lead to
a shiny conductive point on the computer chassis (an
I/O screw will do). Touch the red lead from the multimeter
to the top of the battery. A new battery reads >3.0V.
A marginal battery, is anything < 2.4V. Some motherboards
won't start, once the battery gets down to zero volts.
There is nothing in any documentation, to explain why
that would be the case. Perhaps a sneak path through the
SuperI/O, where it monitors VBAT. Some SuperI/O chips
have a VBAT input, which is one of the inputs to the
analog to digital converter of the hardware monitor portion.

*******

I don't know if I'd be in a rush to return the hardware
just yet. You still don't know for sure, whether it's
something flaky in the install (shorting to bottom of
motherboard), or it's an actual motherboard hardware
failure. The fact that you got it started, puts
a whole new complexion on the issue.

Tell me about it! Now every time I turn my 'beast' on, it insists on
going through a recovery procedure that takes some 15 or so minutes.
I'm betting it is the W7 outgrowth from Chkdsk. Great!

I agree on the hardware return.

Thanks

BF

 

[roger]

But it's not switching on. The RAM can't do anything at this
point, because 3.3V/5V/12V are unavailable. If the fans don't
spin on a system, there's no power to speak of.

But your idea got me thinking about another potential issue.
That's a flat CMOS battery. Some motherboards won't work,
if the CMOS battery is down to zero volts. That really shouldn't
stop a motherboard, but there are models which are known for this.
Almost as if the SuperI/O (with hardware monitor and VBAT input),
has some say in the matter (gates off PS_ON#).

Paul

But, but, but - this was a brand new configuration, with I would hope
a brand new battery. Same with the RAM. Nonetheless, if this goes
on, I will replace the battery and maybe mess with the RAM.
This all beats anything I have ever encountered.

I have never said, but I now am some 83, on death's door, and have a
long main-frame, mini-frame, and micro-frame resume. Hell, I even
had several TRS80's. Remember 'assembly' and 'machine' language?
DOS? Fortran? PL/1? COBOL? Maybe I am just 'getting old'. Should I
resign? My cardiologist would like that. '60 minutes' is coming
on. Where the Hell was 'Meet the press' when I needed it this AM?
But I digress....

Thanks again - I appreciate your intelligence and knowledge.

BF

 

[miso]

Interesting thread. I've had a few Gigabyte mobos fail and I'm sour on
the brand. I've had their CPU fans fail, but I kind of assume they just
outsource those, and lots of fans fail.

I had a 5V SB problem building a computer once and yes, it was an Asus
mobo. The led told me what the problem was. [I recall having to put in a
different power supply to fix the problem, but the details have long
left my memory.] Asus mobos can be a little finicky at times about RAM
and such, but in general that is my preferred brand unless you are
building a server. [Some ECC RAM features are missing on Asus boards.]

You can buy power supply testers. Simple LED ones about $10. With
metering, about $20. I doubt they put much of a load on the supply, but
they are good for a quick go/no-go.

http://www.coolmaxusa.com/productDe...&subcategory=powertester&category=powertester

Coin cells are now made in China crap. They used to be made in Japan.
Get used to them failing or not lasting long. I'm pretty klutzy, but I
don't find removing the coin cell all that hard.

These shitty Chinese coin cells are a real nuisance for people who
aren't technical. A lot of people don't want to open a PC case.

These days, at 83 you might get to build yet one more PC. My dentist has
a lecture that we now live three lifetimes if we are lucky. Life
expectancy when we were barely a notch above the apes was 30 years. Now
you can make it to 90 with luck. I saw Bob Newhart on Big Bang Theory.
Not too shabby for 83. The one that freaks me out is Arsineo Hall. When
I heard he was back on TV, I figured I would see a young Nelson
Mandella. Hall looks good for 57.

 

[roger]

On top of every thing else, last night my machine went dead again.
And, pulling the CPU and its fan had no effect like before. That
makes me think that the cause of what happened before is still unknown
as is a solution. Then suddenly as I was trying to power up, the
machine started. It is still running now, some 12 hours later.

I probably should just ship the mobo, and maybe the CPU back to the
vendors, but I bought the MOBO from MWAVE.COM, who so far has ignored
my phone calls and e-mails on that subject. The CPU came from
Amazon, who has said to ship it back for full refund. Thank you
Amazon. So much for Mwave, and for that matter, Gigabyte. I passed
being aggravated a long time ago.

Thanks for response.

Big Fred

Interesting thread. I've had a few Gigabyte mobos fail and I'm sour on
the brand. I've had their CPU fans fail, but I kind of assume they just
outsource those, and lots of fans fail.

I had a 5V SB problem building a computer once and yes, it was an Asus
mobo. The led told me what the problem was. [I recall having to put in a
different power supply to fix the problem, but the details have long
left my memory.] Asus mobos can be a little finicky at times about RAM
and such, but in general that is my preferred brand unless you are
building a server. [Some ECC RAM features are missing on Asus boards.]

You can buy power supply testers. Simple LED ones about $10. With
metering, about $20. I doubt they put much of a load on the supply, but
they are good for a quick go/no-go.

http://www.coolmaxusa.com/productDe...&subcategory=powertester&category=powertester

Coin cells are now made in China crap. They used to be made in Japan.
Get used to them failing or not lasting long. I'm pretty klutzy, but I
don't find removing the coin cell all that hard.

These shitty Chinese coin cells are a real nuisance for people who
aren't technical. A lot of people don't want to open a PC case.

These days, at 83 you might get to build yet one more PC. My dentist has
a lecture that we now live three lifetimes if we are lucky. Life
expectancy when we were barely a notch above the apes was 30 years. Now
you can make it to 90 with luck. I saw Bob Newhart on Big Bang Theory.
Not too shabby for 83. The one that freaks me out is Arsineo Hall. When
I heard he was back on TV, I figured I would see a young Nelson
Mandella. Hall looks good for 57.

 

[miso]

On 11/27/2013 3:50 AM, (e-mail address removed) wrote:

I used mwave in the past with no problems, but I never tried to return
anything. ;-) I had two Gigabyte mobos get bloated caps, and they
claimed to use Japanese caps.

I have been avoiding these no name vendors ever since a computer I built
ended up using a CPU from a gang that killed a security guard in a
robbery. I bought it from Texas, it drop shipped from California with
the guys name on the box. Then I read about his arrest in the murder.

New Egg and Amazon are killing the competition, but they aren't killing
people. You do have to watch with both New Egg and Amazon in that they
front other companies. This is clearly stated when you purchase. So
while a device that comes out of a Amazon or New Egg warehouse didn't
come from street crime, you don't know about those mystery 3rd parties.

To credit New Egg, they fight patent trolls. Amazon just pays the hush
money. So New Egg at least has a spine. Mother Jones magazine did an
article on what it is like to work in an Amazon warehouse.

http://www.motherjones.com/politics...-free-online-shipping-warehouses-labor?page=1

To entice reading the article, here is her description of the job:
"My brief, backbreaking, rage-inducing, low-paying, dildo-packing time
inside the online-shipping machine."

 

[micky]

Boy! Do I need some help.

My prior posts will show that I have just built a system consisting
of:

1) MOBO - Gigabyte GA-F@A88X-D3H FM2
2) CPU - AMD A10-5800K Quad
3) RAM - Corsair Vengeance Pro 2X4GB
4) Case - Rosewill Galaxy 3
5) PSU - Coolermaster GX 650W 80 Plus
6) W7

It has been running nicely for a few weeks now. But this morning,
when I powered-up, nothing!

So far I have:
1) disconnected the power switch connector from the mobo pins, and
tried to 'trigger' the machine by the usual shorting of the two
power-switch pins. Nothing!
2) pulled a PSU from an old machine and connected its 24-pin power
connector to the Gigabyte mobo, but discovered that that PSU doesn't
have a connector for the Gigabyte mobo 8-pin so-called 'ATX-12V'
header. Nonetheless, I shorted the Gigabyte pins again to see if any
response. Nothing!
3) connected just the mobo power connector from the new PSU to the
24-pin power header on the old machine mobo, and tried to power up the
old machine. Nothing!

You should get a voltmeter so you don't have to spend time in such silly
trials. No offense meant. Spend 5 dollars or more at Harbor Freight
and about 20 dollars most other places. At other places, don't get the
smallest cheapest meter, but one step up from that.

The voltmeter will include an ohmmeter so you can test the on/off
button.