MS-7061 KM4M - Not powering up.

[socketa]

I used this board to test a graphics card that was in another computer.
This motherboard was in storage for 2 years
I turned it on, and it booted up, and windows started;
but it was responding with long delays, and i foolishly assumed that it was the operating system.
So went away and came back in 30mins and it was no longer working.
The PSU voltages are all within spec
Prior to storing it, i replaced the input caps with Nichicon HZ, and it was running good with those, right up till i put it in storage.
There is 3.3V on power switch header, but no startup when shorting to ground, and the southbridge doesn't get warm
No dead shorts on the input rail or the CPU VCORE rail;
but there is 30 ohms to ground/earth on VCORE rail, with no CPU installed
Looking at the schematic (attached)...
i removed the 6 VCORE, OST RLX caps, and they all tested within spec and had good ESR.
Removed (and destroyed, when trying to desolder and lift the legs) the low side MOSFETs.
Removed the diode that is in parallel with that rail;
But there is still 30 ohms to ground
From what i can see, the only reasonable path to ground, that is left, is via the driver chip.
Or is it OK for there to be 30 ohms to ground on VCORE rail?

The two MOSFETS and the two diodes, that i striked out in the schematic are not installed on this particular board, so they are irrelevant.
The components that are circled in green are the ones that i removed

Any suggestions or tips on what to do next?

[piernov]

Re: MS-7061 KM4M - Not powering up.

VCORE also powers northbridge VTT. Low resistance to ground is normal.

[Yousuf]

Re: MS-7061 KM4M - Not powering up.

@OP did you see Vcore or not, sorry got a bit confused. What about other rails for PCH or AMD hub? Details of the processor, PCH (chipset) would help.

[socketa]

Re: MS-7061 KM4M - Not powering up.

I didn't put a mutimeter across VCORE, because there wasn't even a twitch from the PSU fan.
I just checked now, and no voltage registers on the 12V CPU plug when i push the motherboard power button.

From what i now understand, the PSU should respond when the MB power switch is pushed, even if the 4-pin CPU plug is not connected.

VT8237 southbridge VCC3SBY has 3.2V (measured at pin A14 of PCI slot)
the other chip is W83697HF

What rails should i measure?

Here's the schematic:

[socketa]

Re: MS-7061 KM4M - Not powering up.

There is 2.46V at VCC25SBY, and 3.2V VCC3SBY, on at least couple of 'pins' of the southbridge. (i checked at a couple of SMD capacitors).
From what i could see, the power button looks like it's voltage comes from southbridge PWRBTN# pin.
From this info can it be inferred that the southbridge is OK (not blown up)?
I looked at it carefully, and don't see any crack on it

[socketa]

Re: MS-7061 KM4M - Not powering up.

The voltage on the ATTP1 Over Temperature Protection Controller stays pulled up when i push the power button,
and the base to the transistor that switches it, gets no voltage.
So PSON says high (because it's not being grounded by the chip), and therefore the board wont start
There is only two places that the base of the transistor connects to:
1. The Southbridge
2. The ACPI controller

The power that switches the transistor would come from the southbridge?

So is it looking like the southbridge is kaput?
What are the odds that the southbridge got slowly cooked after i turned on the computer, despite the caps and PSU testing in spec?
Maybe the caps were bad initially then they came good - i don't think so though, because if they self-healed then the south bridge wouldn't have cooked.
One of the 6 1800uf caps tests a bit high (2135uf) - which is boarderline (+19%), the others are between 1900uf and 2000uf, and all 6 caps have ESR of around 0.03ohms according to an MESR ESR meter

[piernov]

Re: MS-7061 KM4M - Not powering up.

PSON is driven by SLP_S3# which is an output from the southbridge. Southbridge should pull it high after receiving a trigger on PWRBTN#. It most likely requires RSMRST# to be pulled high previously and RTC clock to be working, and of course the standby rails to be working.
The MS-7 ACPI controller seems to be a proprietary chip, not very clear what it does but it seems to be helping in enabling the various power rails, and I think it takes SLP_S3# as an input for this purpose.

[socketa]

Re: MS-7061 KM4M - Not powering up.

Thanks for the replies
I put in a new CMOS battery

It most likely requires RSMRST# to be pulled high previously

Yes, that has 3.2V when mains power is turned on

and of course the standby rails to be working

Yes, they are working as previously mentioned
There is 2.46V at VCC25SBY, and 3.2V VCC3SBY

RTC clock to be working

No, it doesn't look like that the RTC clock is working - i don't get any significant voltage when i check across each leg of the crystal. (i compared it to the voltage present, about 0.8V, on a similar motherboard)

With the JBAT1 jumper removed, there is 3.1V on pin 1.
When i replace the jumper, it falls to 98mV
3.1V --> 98mV
(this doesn't happen when i compare it to what happens on a similar board, where VBAT = 2.9V with the jumper installed)

When i switch on mains power, with the jumper removed, there is 3.4V.
When i replace jumper, it falls to 253mV
3.4V --> 253mV

The current draw with mains power turned off (voltage is supplied by the battery) is 2.47mA, which seems way too high when i compare it with what's happening on the other board, which only draws 2.3uA
The current draw, with mains power turned on, is 6.5mA

VBAT only goes to the southbridge and the I/O chip
0.008W of power is being dissipated out of the CMOS battery by some component or fault.
0.02W, is being dissipated, if the standby power is turned on

there is 6K ohms from the positive CMOS battery terminal to ground

Where can we go to from here?

[piernov]

Re: MS-7061 KM4M - Not powering up.

Indeed the current draw from the coin cell battery should not exceed 10-15µA.
Make sure jumper is only connected pins 1/2 and not 3 of JBAT1, and measure resistance to ground on pin 2 of JBAT1. Resistance to ground on RTC power rail should be at least a few hundred kiloohms.
If it's low, either C316, U8, or SB1B is shorting it to ground.

[socketa]

Re: MS-7061 KM4M - Not powering up.

Resistance to ground of JBAT1, pin 2, is 38ohms
I left the meter in position, and then shorted all of the pins (except VBAT) of the I/O chip (U8) to ground, and the resistance didn't change,
which means that if the fault is in the I/O chip then there must be 38 ohms to a ground pin that's within it.
I tried the same method with the components that are surrounding the southbridge, but no change in the meter reading.

VBAT is also connected to the southbridge via a 1Mohm resistor, but that tests OK

Also searched for "39R" (and standard resistance values either side of that) on the schematic, to no avail.

I'll remove C316, but might experiment, with hot air, on another board first, since it's sandwiched between a tiny SMD resistor and JFP1, which has plastic that could melt,
or else i will use two soldering irons and someone else with tweezers can remove it.

Will update when done

[socketa]

Re: MS-7061 KM4M - Not powering up.

I used two irons to get the capacitor off,
and it was faulty



So that's good.

What would have likely caused this to happen?, considering:
"Ripple and overvoltage are issues, but MLCC are typically rated for nearly 100% overvoltage and significant ripple."

[piernov]

Re: MS-7061 KM4M - Not powering up.

MLCC can die as a result of mechanical or thermal stress, but they also die seemingly randomly sometimes.
It's much more common to have the southbridge (and sometimes the SIO) die than the capacitor on the RTC power rail, so you are very lucky.

[socketa]

Re: MS-7061 KM4M - Not powering up.

Yes, it was a bit of a surprise

i was going to use a 50V cap because of what someone posed in this forum:
"Generally, with ceramic capacitors, you use capacitors rated for at least 5-10 times the voltage you're gonna work with (100v rated for 12v, 50v rated for 3.3v-5v etc)"

But i came across this, and it seems that voltage rating doesn't have any significant effect on loss of capacitance.
So i guess that it's not for that reason
What am i missing?

[piernov]

Re: MS-7061 KM4M - Not powering up.

No, using a 6.3V MLCC on a 3.3V rail is perfectly fine. I may use a 10V MLCC instead, but any higher than this is unnecessary. MLCCs should already be tested at twice their voltage rating in the factory.

Higher voltage rating won't make the MLCC last longer if it never sees voltage spikes near or above its rating.
On the contrary, higher voltage rating means the cap will be physically longer, which may not be practical in a densely populated board, and it also makes it more likely to be damaged by the board flexing.

Also, an MLCC can start losing a bit of capacitance after many years, but the fact that it is partially shorted is not due to normal aging, it just has failed internally.