Need a power supply expert 3

[nob1]

Not sure about the sequence of events after you press the ON button. I have a Gigabyte GA-81HXP MoBo with a 2.8 P4, 512Gb RAMBUS. It suddenly decided not to boot. You press the button the fans start up, the power supply light comes on - and then everything goes off. As far as I can discover, as the supplies come up a 'power good' signal is sent from a chip on the MB to the PSU to maintain the supplies. All the supplies seem to coming up (checked with multimeter)but I'm not sure if the 'power good signal' is a pulse or a permanent voltage and if I can measure it or not. Common sense tells me that the PSU is overloaded so I changed the PSU for a 500 watt PSU with the same result. This meant there was a possible power line short to deck so I changed the MB! Same result! What do I do now? Any help would be greatly appreciated.

 

[technome]

Possible short in a peripherial device..hard rive, floppy, CDrom etc. Start machine with just the keyboard, mouse and video card.

Still no good, remove the mobo, place it on paper, power it on. Do you have a spare video card to swap in.

 

[nob1]

Thanks for answering Technome. I've removed all the peripherals with the same result - 'fraid I haven't got a spare graphics card. I'm going to try the paper trick.

 

[nob1]

Tried the paper insulator - still no go. That leaves the graphics card. I'm going to swap it with the card in my wifes machine but I'll have to wait till she comes home or she'll kill me. Thanks again.

 

[felixc]

The power on/off command to the ATX power supply is driven by a circuit that is permanently powered by a "5V backup voltage" from the power supply. This circuit takes the front panel switch input, and it also takes commands from the CPU for the power-off sequence.
Does your mobo have time to issue a "beep" before the power goes off?

 

[Freestone]

In order to prevent CPU overheating and possible damage, some (most? all?) motherboards will signal power shutdown if it doesn't detect CPU fan rotation within a certain range. Perhaps your CPU fan is causing your problem?

 

[G0AOZ]

It won't hurt to power up the mobo without a graphics card inserted. In fact, you could also lose the RAM module as well... So that you end up with NO CONNECTIONS AT ALL other than the PSU, and literally just the CPU and heatsink inserted.

ROGER - G0AOZ.

 

[pcdoc4christ]

Nob:

Be sure to use a static strap, as even a small amount of static electricity may fry the sound card or system board! You may get one cheap at Radio Shack or some mom-n-pop computer store. And try not to rest the sound card on a plastic or metal surface.

I also like to run a humidifier for an hour or so in the room where i'll be working on the PC.

Regards,
Doc

======================================

"Freely you have received, freely give." --Jesus

pcdoc4christ@yahoo.com

 

[edfair]

The power on signal goes from the M/B keepalive circuits to the power supply to turn the remainder of the power on. Once everything is stable there is a power good signal from the power supply that allows the microprocessor clock to start running.

Freestones fan suggestion is good. You could try removing the power on switch wires and use a screwdriver to short the pins to eliminate the switch as potential problem. You can also have problems with the processor not starting up but that shouldn't power the system down.

I would guess the M/B has bought the farm.

Ed Fair
Give the wrong symptoms, get the wrong solutions.

 

[nob1]

I'm overwhelmed with all the answers - Thankyou. Right; this is what I have done so far:-

Isolated the MB apart from graphics card, KB and mouse. Paper underneath to stop any shorts to deck. - Result the same.
Changed the MB. - Result the same.
Changed the graphics card for one from a functioning machine - result the same.

I'm going to try and find the keep alive voltage and measure that.

 

[nob1]

No, there is no beep before it shuts down.

 

[Frank4d]

The 5VSB is a purple wire. It should be around +5 volts (relative to one of the black ones) anytime the PSU is plugged in... whether the PC is running or not.

Shorting the Power-On (green wire) to a black wire will turn the PSU on. If the ATX power supply is plugged into the MOBO, the MOBO is supposed to drive the green wire close to 0.5 volts when turned on by the power switch.

The Power-Good (gray wire) should be around +5 volts (relative to one of the black wires) when the PSU output voltages are OK.

 

[nob1]

CPU fan runs as soon as I press the ON button - then goes off again. The ON button just shorts out the two pins on the MB I believe. How could it cause problems?

This one has really got me. When I changed the power supply, the new one had a plug, which according to the manual supplies extra power, missing. It's a 6 pin inline socket on the MB. The old PSU had this plug so after testing it with the new psu and getting the same result I changed the PSU back to the old one.

 

[edfair]

The extra 6 pin PS connector is P4 specific to keep the current under 5 amps on all connectors. And shouldn't have affected anything for the short duration of the test.

The switch with a partial short would probably have shut it down after 5 seconds which is the built in power down. Less than that is not the switch. My bad.

When you changed the motherboard, did that include the processor? They can short internally and give this same sequence of symptoms. If the processor is bad, looking at it may show some discoloration.

Ed Fair
Give the wrong symptoms, get the wrong solutions.

 

[Freestone]

Fan rotation doesn't rule it out as the culprit as there is a rotational sensor that could be bad, giving incorrect readings. Perhaps someone knows of a way to test this or at least provide a "good" reading to the motherboard so the fan can be eliminated as the cause.

 

[Frank4d]

Fan rotation doesn't rule it out as the culprit as there is a rotational sensor that could be bad, giving incorrect readings. Perhaps someone knows of a way to test this or at least provide a "good" reading to the motherboard so the fan can be eliminated as the cause.

The fan rotation speed is a tachometer output. A multimeter will read it's output as an AC voltage, or as a frequency (usually 1 or 2Hz per RPM depending on the fan) if the multimeter is able to measure frequency. An oscilloscope could be used too... but most people don't own one.

 

[paparazi]

So, just to confirm and try, if you havn't already done so.

Bare motherboard! just CPU with heatsink/fan, one stick of ram, video card (monitor attached) and new power supply including 12V P4 connector if it has one.

Absolutely nothing else connected!! no front panel wires, IDE cables nothing!

Now clear the cmos (3 pin header near the button battery) refere to manual

Then, momentarily touch the two pwr (power) pin out headers with a small flat blade terminal screw driver.

Anything?

Is the reset or power button stuck in?

Is the CPU fan attached to the CPU fan header (NOT SYSTEM FAN HEADER)

Martin



We like members to GIVE and not just TAKE.
Participate and help others.

 

[nob1]

Thanks everyone for your extremely erudite answers. Haven't had time to do the tests suggested by Paparazi yet (Its mother's day in UK)but will let you know as soon as I do.

P.S. power button or reset not stuck in. CPU fan runs then stops (connected to header)

 

[pcdoc4christ]

Turn off that PC & go be with mom!

======================================

"Freely you have received, freely give." --Jesus

pcdoc4christ@yahoo.com

 

[nob1]

I'm back. I've done everything Paparazi suggested. The CPU fan starts up, the PSU fan starts up, the power ON lamp flickers briefly - everything goes off! At a complete loss now. 2 MBs having the same problem I feel is unlikely, but that's what it's beginning to look like. I'll get hold of a new MB and see what happens. The 2 MBs that are causing the problem are Gigabyte 81HXP, I'm unable to elicit any response either by Email or phone from Gigabyte and I'm beginning to wonder if this was another reason they discontiued the board so quickly and not just the RAMBUS situation. Oh well! One lives and learns. Thank you all for helping so patiently. Brilliant site.

Arthur Clarke.