Re: Testing a system that won't boot

R wrote:
> On Sat, 12 Jul 2008 05:06:17 -0400, R <> wrote:
>
>> This system was working for a couple weeks. Needed to reboot during
>> software install. It shut down and never came back up.
>>
>> Symptoms:
>> Motherboard LEDs come on when power supply is plugged in.
>> System does nothing when power switch is pressed. No beeps.
>>
>> Configuration:
>> Antec Neo 480 power supply
>> Asus P5K-E/Wifi-AP
>> Intel Core 2 Quad Q6600 2.4Ghz
>> 4GB RAM (2 x 2GB)
>>
>> I've checked to make sure that the power switch does close contacts
>> (with Fluke meter), and that the power switch/reset connector is
>> seated.
>>
>> Since motherboard LEDs light, the power supply is doing -something-
>> but it could still be a supply problem I suppose.
>>
>> Anyone know of other tests?
>
> Just left the system running for a few minutes and noticed the
> familiar smell of electrical smoke. I believe it's coming from the
> Antec NeoPower 480 supply.
>
> I've had the supply for quite a while. Just got the new Asus P5K-E
> motherboard and Q6600 CPU. I'm wondering about the chances that the
> supply just burned out, and that the motherboard/CPU are OK.
>
> The supply should have had foldback protection, so I'm not sure why a
> motherboard load or short would have taken it down. Still...
>
> The question seems to be whether to try a new supply, or whether that
> may be toast after first power-up.
>
> Random speculation welcome, especially if you've seen this type of
> thing.

My random speculation would be, the Antec died quietly, and left the
motherboard alone. Slap another power supply on it and try again.

I test power supplies here, before I use them. What I did, is built
a load box, with a small load on each of the rails. There is an
80mm fan to cool the resistors and a switch, to turn the power supply
on and off (via PS_ON# and COM). While the power supply is delivering
about 100W or so total, I measure the voltages with a multimeter. And
let it run for a couple hours, just to see if it'll behave properly or
not. It isn't a Tomshardware style test, as I'm not aiming to overload
the power supply, and make smoke come out. It's just intended to prove
the supply is well behaved, before connecting it to a motherboard.

Paul

R wrote:
> On Sat, 12 Jul 2008 06:32:17 -0400, Paul <> wrote:
>
>
>> My random speculation would be, the Antec died quietly, and left the
>> motherboard alone. Slap another power supply on it and try again.
>
> Woke up thinking that today. The Antec's been in use for a while, and
> I wouldn't be surprised if its fan died. I think it's about a week
> over warranty so that would be logical.
>
>> I test power supplies here, before I use them. What I did, is built
>> a load box, with a small load on each of the rails. There is an
>> 80mm fan to cool the resistors and a switch, to turn the power supply
>> on and off (via PS_ON# and COM). While the power supply is delivering
>> about 100W or so total, I measure the voltages with a multimeter. And
>> let it run for a couple hours, just to see if it'll behave properly or
>> not. It isn't a Tomshardware style test, as I'm not aiming to overload
>> the power supply, and make smoke come out. It's just intended to prove
>> the supply is well behaved, before connecting it to a motherboard.
>
> Sounds like a good idea for testing the current supply too.
>
> I do have some big resistors, and was considering testing the +5
> anyway. I wasn't going to run it hot though...just enough load to
> make sure the supply turned on. I was thinking about hanging a 4 ohm
> off the +5 (only 5+ watts).
>
> What value resistors are you currently using in your load box?
>

There is an example here. Note that some of the links on this page
would be dead, as there are likely later versions of the catalog
pages now.

http://groups.google.ca/group/alt.co...b?dmode=source

Paul

R wrote:
> On Sat, 12 Jul 2008 14:33:34 -0400, Paul <> wrote:
>
>> R wrote:
>>> On Sat, 12 Jul 2008 06:32:17 -0400, Paul <> wrote:
>>>
>>>
>>>> My random speculation would be, the Antec died quietly, and left the
>>>> motherboard alone. Slap another power supply on it and try again.
>>> Woke up thinking that today. The Antec's been in use for a while, and
>>> I wouldn't be surprised if its fan died. I think it's about a week
>>> over warranty so that would be logical.
>>> ...
>>>
>>> What value resistors are you currently using in your load box?
>> There is an example here. Note that some of the links on this page
>> would be dead, as there are likely later versions of the catalog
>> pages now.
>>
>> http://groups.google.ca/group/alt.co...b?dmode=source
>
> Thanks Paul,
>
> Turns out that the motherboard and CPU are OK (I think). I replaced
> the Antec Neo 480 with a Cooler Master Real Power Pro 650W. Strange
> thing though---it does seem to stall frequently at powerup. I haven't
> tracked down the reason yet. Perhaps one of the individual lines in
> the Cooler Master does not meet the system's current requirement.
> Can't imagine why that would be.
>
> When it stalls, the front panel lights come on, there's a beep, but no
> Post on-screen. I've got it to boot, probably just by retrying--I'm
> not sure what could be intermittent.
>

If there was no beep at all, on the failed startups, then you'd
conclude that the processor wasn't able to execute any code.
But since it beeps and then stalls, it must be running into
trouble, at a point after it has managed to program enough
hardware to make the beep sound. It could be a peripheral
that is getting stuck.

A PCI port 80 POST debugging card, would be one way to watch
the progress the BIOS is making, but those cards are not really
good value for money. The two digit codes are not well documented,
and on occasion when someone asks about a code, I find it listed
as "reserved", so can't tell them anything about what it means.
And the codes are "progress" codes, rather than "error" codes,
which also isn't as useful.

There are a few motherboards made today, that ship with the
two digit display right on the motherboard. For those, you
don't have to buy a debugging card, and get it "for free".

I sometimes wish Asus would put the Vocal Reporter system on
their products again, as at least that would give an error report,
and a bit more feedback about the startup process. At least
an error report, requires less deductive reasoning, than a
progress report. The feature would probably cost them a
dollar or two.

Paul

On Jul 13, 6:17*am, R <R...@nospam.com> wrote:
> Turns out that the motherboard and CPU are OK (I think). I replaced
> the Antec Neo 480 with a Cooler Master Real Power Pro 650W. Strange
> thing though---it does seem to stall frequently at powerup. I haven't
> tracked down the reason yet. Perhaps one of the individual lines in
> the Cooler Master does not meet the system's current requirement.

Normal for a perfectly good supply to fail in some systems. Also
normal to have a defective supply boot a computer. You have the Fluke
meter. Learn why. Learn what you have AND provide 3 digit numbers so
that others can be more helpful.

A two minute procedure is "When your computer dies without
warning....." starting 6 Feb 2007 in the newsgroup alt.windows-xp
at:
http://tinyurl.com/yvf9vh
Connector chart to locate each color:
http://www.hardwarebook.net/connecto.../atxpower.html

In your case, most important numbers are from any purple, orange,
red, and yellow wires after power up and with the maximum computer
load. Those numbers must exceed 3.23, 4.87, and 11.7. If gray wire
is also OK (greater than 2.4 volts), then the power supply system
(more than just a power supply) is exonerated. Move on to other
suspects.

If power supply controller does not provide an "OK", then CPU cannot
execute to discover and examine other computer components - no beeping
can occur. Ignore things such as RAM, disk drive, etc. These are not
relevant yet. First a power supply controller must let the CPU
execute. First voltages (sufficient to spin disks and light lights)
must meet those numbers. No other way can obtain a definitive
answer. Get numbers using that Fluke meter.