Desktop machine works fine a few days then it doesn't start at all.

[alex777]

Hey, guys!

I have a Gigabyte B450M DS3H motherboard that works fine for a few days and then it doesn't start at all; even though I have standby voltage (+5VDC) on front panel header, when I short the PWRSW pins for turning the MOBO on, it doesn't do anything. According to your experience, guys, where do I have to start testing?

Greetings!

 

[GTP]

For the cost of a replacement motherboard why would you even bother starting?
These B450M mobo's were an ok board but they are on the cheaper scale.
Just replace it and move on?

 

[britechguy]

@Barcelona

Absolutely agreed if this is for a client. But if it's for a learning experience, that's a different kettle of fish.

 

[GTP]

Hmmm, I neglected to ask that question...
So @alex777 is this for a client or just learning?

 

[alex777]

Hmmm, I neglected to ask that question...
So @alex777 is this for a client or just learning?

It's like @britechguy said... for learning purposes.

 

[NviGate Systems]

Typically there will be MOSFETs that control power regulation from the ATX connector. So I would start testing for voltages starting from the ATX connector going to MOSFETs. They could be surface mount or DIP, depends on the board.

You aren't going to readily find data sheets, but you can use some detective skills and slowly test the flow of power.


With power removed !!!!! You can also test for a short to ground, it could be an intermittent short that comes and goes. Sometimes the fault is in one of the many layers of the PCB which you can't fix.

If I can make a suggestion, while this kind of learning is good, it might be better to learn more about laptop power distribution, as that could help you fix client machines and make money. People are not putting money into desktops as much, but for some, getting a laptop running again you'd be surprised how they need it.

 

[alex777]

Thank you very much, buddy!
I'll start trying to find where those MOSFETs are, then I'll let you know.

 

[NJW]

According to your experience, guys, where do I have to start testing?

Power supply ripple. Try putting a capacitor between 5VSBY and ground, then press the power button. (Capacitor value: not critical, try between 400 and 700 µF, between 10 V and 25 V rating.)

If that doesn't make any difference, try injecting any good quality 5 V power source between 5VSBY and ground.

I'll start trying to find where those MOSFETs are ...

Look for the coils – they're easier to identify. The MOSFETs won't be far away (don't ignore the other side of the board, but that's more likely on a laptop motherboard).

 

[Larry Sabo]

I have standby voltage (+5VDC) on front panel header, when I short the PWRSW pins for turning the MOBO on, it doesn't do anything.

Do you have 3.3v on the SuperIO chip? See the videos on the YouTube Electronic Repair School channel for tips on diagnostic techniques.

 

[Markverhyden]

Hey, guys!

I have a Gigabyte B450M DS3H motherboard that works fine for a few days

How do you know it works fine for a few days? What OS were you running?

 

[britechguy]

How do you know it works fine for a few days? What OS were you running?

My guess, just from the information presented and how it was presented, is that the hardware in question is in the long term possession of @alex777 and that he's actually firing it up routinely.

The OS really shouldn't matter, at all, if the issue is hardware based and it certainly sounds like it is (regardless of exactly what piece of hardware).

 

[alex777]

Do you have 3.3v on the SuperIO chip? See the videos on the YouTube Electronic Repair School channel for tips on diagnostic techniques.

The SuperIO chip of my MOBO is the IT8686E, and I found only the IT8686 on the Gigabyte GA-Z270X motherboard schematic which I uploaded for you to see. Here are the readings I have :

3VSB (IT_VCCH) pins 64 and 96: 3.3 volts
SYS_3VSB (pin 95): 3.4 volts
PWRBTSW/PANSW (pin 104): 5.1 volts (coming from a 8.22 kohms resistor and PWSW front-panel header)
SIO_18V (VCORE) pin 65: 1.96 volts
RSMRST (pin 114): No voltage
SLP3 (SUSB#) pin 100: No voltage
SLP4 (SUSC#) pin 106: No voltage

 

[Markverhyden]

My guess, just from the information presented and how it was presented, is that the hardware in question is in the long term possession of @alex777 and that he's actually firing it up routinely.

The OS really shouldn't matter, at all, if the issue is hardware based and it certainly sounds like it is (regardless of exactly what piece of hardware).

This may be a labeling issue. To me a motherboard is just that, not a computer. And people don't buy motherboards to plug into power and watch the couple three led's light up and/or blink. I've had plenty of motherboards that power up and stay that way no problem. But add the OS and it's useless. I've even had machines which would BSD on any version of M$ but ran any *nix without a problem.

 

[britechguy]

To me a motherboard is just that, not a computer.

We're just coming at this from different angles based on our own experiences. Nothing wrong with that.

I never think of a mobo as "a freestanding thing" but only as part of a computer as a whole. And that's just because it's the one and only way I've dealt with them in practice.

 

[Stever]

What about the 12 volt line ? Is it there ? Unless all voltages are present then there will not be a Power Good output to turn everything on !

 

[alex777]

What about the 12 volt line ? Is it there ? Unless all voltages are present then there will not be a Power Good output to turn everything on !

Does the Power Good signal enables the rest of the voltages on the power supply? Is that PWRGD the grey cable, right?

 

[Markverhyden]

When I've manually checked a desktop case power supply I'll connect the green wire to the adjacent black to get it to fire up all rails. Then I can check each one with a VOM. But the problem with that is there is no load on any of the rails.

 

[Stever]

Does the Power Good signal enables the rest of the voltages on the power supply? Is that PWRGD the grey cable, right?

https://en.wikipedia.org/wiki/Power_good_signal

 

[glennd]

A quick safety tip - bare hands and big capacitors don't mix. In a typical switch-mode power supply some of those capacitors could hold up to 400V for a long time after the power's removed. That's ouchy territory.

Great fun at parties though!

 

[alex777]

When I've manually checked a desktop case power supply I'll connect the green wire to the adjacent black to get it to fire up all rails. Then I can check each one with a VOM. But the problem with that is there is no load on any of the rails.

In this case is not a power supply problem, since I've already checked the voltages doing that connection between green and black cables; the readings are fine even putting some loads.