ASUS ASUS P5VDC-MX Dead CPU?

[socketa]

Re: ASUS ASUS P5VDC-MX Dead CPU?

The Multi-Phase PWM controller is not receiving power good signal
So the upper and lower mosfets are not being activated

Just rechecked the voltages to ground of that mosfet, labelled H
G 2.7V
D 3.3V
S 1.2V (not the 2.75V that was stated earlier)

So is that source voltage normal enough to power the NB chip (if indeed that is it's function)?

What component would stop the power good voltage from reaching the PWM controller - Does the Power Good rail pass through the NB first?

[momaka]

Re: ASUS ASUS P5VDC-MX Dead CPU?

I'm assuming that you mean pin 10 (not pin 9) of the RT9245A chip.
Got no voltage on this pin after power button is pressed

Ah yes, Pin 10, sorry.

Yup, missing voltage on that pin would cause the CPU VRM to not start up. Now the question becomes why.

Just rechecked the voltages to ground of that mosfet, labelled H
G 2.7V
D 3.3V
S 1.2V (not the 2.75V that was stated earlier)

So is that source voltage normal enough to power the NB chip (if indeed that is it's function)?

Yes, 1.2 V is a lot more like it. All of my socket 775 boards have 1.2-1.3 V for that. I think it is also used for CPU<->NB communication lines.

i'm still thinking about that cooked fet.
maybe it's blown or something it's powering is shorted.
it really isnt normal to see that kind of heat mark.

Oh no, on ASUS and AsRock boards of that era, it really is perfectly "normal". If I had a penny for every ASUS and AsRock board I saw with overheated MOSFETs, I'd have saved enough now to buy a good new PC . It is quite common really. I'm not implying that it is good, though.

ASUS just made a poor design there. This should probably have been a buck-type regulator (i.e. same stuff as the CPU... inductor and 2 FETs + controller). Instead, they just put a FET for linear regulation. Just look at the voltages on that FET: 3.3 V coming in and 1.2 V coming out. That is a 2.1 V drop! At 1 A of current, the MOSFET will dissipate 2.1 Watts, which is more than enough to make it run hot.

What component would stop the power good voltage from reaching the PWM controller - Does the Power Good rail pass through the NB first?

That is a good question. I think it's quite possible the PGOOD signal is generated by another chip. Either NB or SB, or perhaps something else that controls the power to the board. At this point, you will have to start doing some tracing. Hopefully, that trace doesn't go to some inner layer so that you can follow it. If not, then worst case scenario you try to insert proper voltage signal on PGOOD and see what happens. But let's not resort to that now.

I think you should start by getting a datasheet for each of the major ICs on the board. Maybe not the LAN and Audio chips, but definitely the monitoring chip. I'll see if I can find out which one that is, too.

*EDIT*
Okay, from some images online, I think the hardware monitor chip should be either that ITE chip on the bottom-left of your board or the ICS chip on the top-right corner (next to the RAM slots). Can you post the part numbers of those for your board, just so that we can verify?

[socketa]

Re: ASUS ASUS P5VDC-MX Dead CPU?

ITE
IT8712F-A
0551-IXS
ZN2FPRL

ICS
561180A
0552
9P933DFLF

(If i make the last F upper case, and don't add this sentence, then all of that text, except for the first I, goes lower case after saving the post - there might be a bug in the webpage code)

[socketa]

Re: ASUS ASUS P5VDC-MX Dead CPU?

Can't find a datasheet for the ICS chip
Found a ITE datasheet for ITE chip (minus the "-A" part of the part 'number')
http://pdf1.alldatasheet.com/datashe...1/IT8712F.html
In that PDF, i was hoping to find words like: "Power Good signal to VRM controller" - but no

Just for fun, i pushed the power button and measured voltage to ground on:
Vcc pins 4, 35, and 99 - all measured 5.24V
Vin3/ATXPG pin 95 - 5.24V

[momaka]

Re: ASUS ASUS P5VDC-MX Dead CPU?

Can't find a datasheet for the ICS chip
Found a ITE datasheet for ITE chip (minus the "-A" part of the part 'number')

Same here. I got nothing for the ICS, even with only partial strings of the number you gave. (Speaking of which, maybe check that number again? I know "5" and "S", "B" and "8", or "O" and "0" can get confused sometimes)

http://pdf1.alldatasheet.com/datashe...1/IT8712F.html
In that PDF, i was hoping to find words like: "Power Good signal to VRM controller" - but no

Yes, that would be nice, but those chips are highly versatile, so they don't get in specifics like that usually.

That said, look on the diagram of page 41. Check voltage on pins "PWRON#" (pin 72) and "PSIN" (pin 71).

According to datsheet description, PWRON# is a request output. So when you press the power switch on your PC, the LPC first shorts "PSON#" (pin 76) to ground to turn on the PSU. Then, it sends a signal from PWRON#, likely to the ICH, to request a power-up sequence. The request is then returned through pin 71, PSIN. At least that's what I am guessing those pins do. Of course, your board may be designed to use those pins differently (perhaps PSON goes straight to CPU VRM or some other controller). Nonetheless, measure the voltage on them. Perhaps we may be able to get a bigger picture of what is going on.

Also check the voltage on the following pins:
85, "RSMRST#"
32, "PWROK1"
78, "PWROK2"
95, "ATXPG"

I *think* RSMRST# is the pin that tells the LPC (or the ICH?) when the system is ready again after a reset. Maybe something is holding the PC in reset state. The other ones should be self-explanatory, if they are wired to what their names say they are, of course.

Just for fun, i pushed the power button and measured voltage to ground on:
Vcc pins 4, 35, and 99 - all measured 5.24V
Vin3/ATXPG pin 95 - 5.24V

I know this is not the problem... but does your PSU really output 5.24 Volts? That's really only 0.01V away from the maximum allowed on the 5 V rail. I'm sure your PSU is okay, since I know you had some not-so-nice experiences with bad PSUs. Perhaps your multimeter batteries need to be changed . Low battery in a multimeter usually makes it measure voltages a bit higher than what they are.

*EDIT*
I forgot to add...if you see any components connected to those pins that I listed above, check them as well. It would be nice if the problem is only a single shorted transistor or open resistor.

[stj]

Re: ASUS ASUS P5VDC-MX Dead CPU?

maybe it's the power-good from the psu, but it would be buffered by a transistor or a resistor/diode combination.

[socketa]

Re: ASUS ASUS P5VDC-MX Dead CPU?

Ok, a toothpick came in handy for that exercise

Am pretty certain that that part number is correct - i even checked it under the microscope

Yes, that 5.24V is what the PSU is supplying, (it's another ISO brand PSU, and this time i checked that it was within the tolerance before using it, the 5VSB is 5.07 V, which is cool) - i'm using it because it's the only 24 pin that i have that's lying around

Voltages on those pins are:

71 - 3.4V
72 - 0.00V
85 - 3.45V
32 - 3.37V
78 - 1.22V
95 - 5.25 V

[socketa]

Re: ASUS ASUS P5VDC-MX Dead CPU?

So, looking at the voltages on those pins, why would the ICH (North Bridge?) be returning a voltage to PSIN, when it has not received a voltage from PWRON#?

[stj]

Re: ASUS ASUS P5VDC-MX Dead CPU?

a "#" usually means it's active-low.
so putting it to ground means "on"

[socketa]

Re: ASUS ASUS P5VDC-MX Dead CPU?

"if you see any components connected to those pins that I listed above, check them as well."

All the resistance readings i got are texted in the attachment on this post.
The component next to Pin 90 VREF gives open resistance reading on the multimeter
I get no resistance readings on the other surrounding components
Also noticed that there is continuity between pin 65 and 68

Thought to check the temps again after about 5min
Ambient = 17.5 Deg C
SB: 21
NB: 24
CPU (without heatsink attached): 28

[momaka]

Re: ASUS ASUS P5VDC-MX Dead CPU?

Am pretty certain that that part number is correct - i even checked it under the microscope

Well, then I guess there is just no datasheet for it on the net. I hate it when that happens. Best hope is the CPU VRM PGOOD signal does not go through that. I guess you can check by putting one multimeter probe on PGOOD where the signal enters the CPU PWM controller and then use your other multimeter probe to test if there is a connection between any of the pins on that ICS chip.

In fact, I think you should also do that with the ITE chip.

a "#" usually means it's active-low.
so putting it to ground means "on"

Good info. I didn't know that. That explains some of the voltage above.
Speaking of which...

71 - 3.4V
72 - 0.00V
85 - 3.45V
32 - 3.37V
78 - 1.22V
95 - 5.25 V

Okay, so it seems "PWRON#" (pin 72) is sending a power-on request signal to ICH (or whatever else it is connected to), and it is getting a signal back on pin 71, "PSIN", meaning system recognized the request to power-up.

85, "RSMRST#" - I checked again, and according to datasheet (page 32 for mine or anyone else's reference), that is the power-good signal of VCCH (i.e. if VCCH is there, this pin will be low[?]). Low threshold according to datasheet is 3.5V +/- 0.2V and high threshold is 4V +/- 0.2V. I'm not sure if that 3.46 you got is good or not. Might a bit too close to the threshold. But let's keep that in the back of our head for now and continue.

95, "ATXPG" - good voltage there.

32, "PWROK1" - good voltage there.

78, "PWROK2" - not sure if that one is bad too. This is type DOD8/DIOD8 buffer (page 37), which according to datasheet (page 177) has a input voltage deadzone between 0.4V and 2.2V - that is, input voltages to that pin between 0.4V and 2.2V are invalid. You have 1.22 Volts, which could indicate some sort of trouble.

So I guess the next step would be to measure the voltages on all of the above pins again, but this time without powering ON the motherboard. Just plug in power supply in the wall (so that there is 5 VSB going to the motherboard). I want to see what changes there.

All the resistance readings i got are texted in the attachment on this post.
The component next to Pin 90 VREF gives open resistance reading on the multimeter
I get no resistance readings on the other surrounding components

The component next to pin 90 is a ceramic capacitor. In fact, all of those tan components are ceramic capacitors. They should always read open (infinite resistance) outside of the circuit. In circuit, though, they are almost always in parallel with other components (especially power rails), so you can get a low reading and think that it is bad when it isn't. (Speaking of which, you put a "0" next to that component, indicating short-circuit, NOT open-circuit. So what do you mean? Or rather, what is your multimeter showing you?)

As for the resistances you got on the other ceramic caps - is that with the PSU connected to the motherboard? It certainly seems like it to me. 50 Ohms is very typical dummy load resistor on the 5V/5VSB rails.

Check those black resistors too, just to make sure one hasn't gone open-circuit or something. They probably should be in the KOhm range. Get the code that it is written on them and look it up in an online SMD resistor calculator to see what values those resistors should be and what you get. If you measure lower resistances than what is printed, then that is probably okay. But if you get a higher resistance than what is printed on the resistor, it could be bad.

Also noticed that there is continuity between pin 65 and 68

Pin 68 is "COPEN#" - that is Chassis Intrusion Detection. When pulled to ground it means there is no intrusion. There should be a jumper on the motherboard for that, where you can remove the jumper and connect it to a case switch to detect if the side of the case gets open. When that function is not used, manufacturers put a jumper to pull Chassis Intrusion to ground. See page 44 of the motherboard's manual.

So yes, the reading you got (almost no resistance to ground) is good.

Pin 65 is "DSKCHG#" - according to datasheet (page 35), "This signal is active low. It senses whether the drive door has been opened or a diskette has been changed."
I don't know if this should be low or not, but it related to FDD, so it shouldn't prevent the board from booting.

Yes, that 5.24V is what the PSU is supplying, (it's another ISO brand PSU, and this time i checked that it was within the tolerance before using it, the 5VSB is 5.07 V, which is cool) - i'm using it because it's the only 24 pin that i have that's lying around

You don't really need a 24-pin PSU. 20-pin PSU is fine as long as it is of good quality.

[socketa]

Re: ASUS ASUS P5VDC-MX Dead CPU?

No continuity from Pin 10 (Powergood) of CPU PWM controller, to any pins of those other two chips.

Pin - V(with computer powered on) - V(only PSU switch on)
71 - 3.4V - 0V
72 - 0.00V - 0V
85 - 3.45V - 3.43V
32 - 3.37V - 0V
78 - 1.22V - 0V
95 - 5.25V - 0V

The "0 Ohm" in the pic is not a short (sorry 'bout that - need to pay more attention) the meter reading was 0L - which means open.

Yes, those Ohm readings were with the PSU plugged in
So Unplugged it and remeasured (seems like there must be some non-conductive layer on the SMD resistors - as i had to give em a good poke to get a reading)

Then i get to measuring the last resistor (on the bottom left corner connected to pin 67 VCCH) which initially read open circuit, so i prodded it a bit to try and get continuity with the probes, and it popped out of it's solder grips and went under the PCI slot, so i fished it out with a toothpick and it looked unrecognizable so i stupidly threw it into the rubbish bag - which i emptied out on a piece of white paper few minutes later, and could not find anything that looked like a SMD)
From the quick glance that i took before tossing it, it looked like it had somewhat disintegrated.
The measurement of this resistor was 97ohm with the PSU connected - now the resistor is missing, so i plugged the PSU back in and get a reading of 97.5 ohms across each side where the resistor was.
Then unplugged it from the mainboard and get a reading of open circuit (0L)
So i'm starting to think that this resistor was open!
Now focusing on the PSU - measured the resistance across the 5V rail and ground and it was 50 ohms.
Then measured the restance across 5VSB at it was 97.5 ohms!

I can see that that resistor was connected to pin 67 (+5V VCC Help Supply)
Would the +5V VCC Help Supply be another name for 5VSB?

There is some corrosion on four corner legs of the ITE chip next to the (now missing) resistor and what looks like darkening of the trace

The black resistors going downwards from the 01C then across to the 015 or 510 are as follows:

Number - Calculated value - Actual
01C - 10K - 6.1K
18C - 15K - 6.4K
18C - 15K - 8.2K
688? - ? - 4.2K
303 - 30K - 13.5K
01C - 12.4K - 8.1K
510 - 51 - 51

So all are lower (rather than higher)

[momaka]

Re: ASUS ASUS P5VDC-MX Dead CPU?

So Unplugged it and remeasured (seems like there must be some non-conductive layer on the SMD resistors - as i had to give em a good poke to get a reading)

Again, those tan components are NOT resistors. They are SMD ceramic capacitors.

Then i get to measuring the last resistor (on the bottom left corner connected to pin 67 VCCH) which initially read open circuit, so i prodded it a bit to try and get continuity with the probes, and it popped out of it's solder grips and went under the PCI slot, so i fished it out with a toothpick and it looked unrecognizable so i stupidly threw it into the rubbish bag

No worries. It is just a ceramic capacitor used for filtering. Board will still work without it in many cases. But, if you really want to replace it for peace of mind, you can take one from another board. Judging by the size, I would guess it is 0.1 - 0.47 uF. If you have capacitance on your multimeter - great! Pull some caps and see if you can find something in that range. If not... it doesn't really matter anyways. Just pull any tan ceramic cap of the same size and put it there. The value of that ceramic cap is hardly critical.

The measurement of this resistor was 97ohm with the PSU connected - now the resistor is missing, so i plugged the PSU back in and get a reading of 97.5 ohms across each side where the resistor was.
Then unplugged it from the mainboard and get a reading of open circuit (0L)
So i'm starting to think that this resistor was open!

Re-read the last few parts of my post above. Again, those things are capacitors, not resistors. Ideally you won't get a reading across them (i.e. open circuit). But because of other components on the board, that won't happen. And this especially won't happen with the PSU plugged in the motherboard. That's why I said to disconnect it, because it can foul your readings.

I can see that that resistor was connected to pin 67 (+5V VCC Help Supply)
Would the +5V VCC Help Supply be another name for 5VSB?

Well, pin 67 (VCCH) just needs to be connected to a 5V source according to datahseet. That can be 5 VSB or 5 V rail on the PSU. Since you seem to be getting the resistance of the 5 VSB on the PSU, then this pin most likely is connected to 5VSB. A better (and more proper) way to check is to stick one multimeter probe where the 5VSB wire goes on the ATX connector (disconnect PSU from board for this) and test if you get low resistance or good continuity to VCCH on the LPC. That will confirm the connection.

Either way, none of that is too important because VCCH is 5V when you measured it before, so you need not worry about it.

There is some corrosion on four corner legs of the ITE chip next to the (now missing) resistor and what looks like darkening of the trace

Hmm, I see it too. Does the LPC get hot when the PC is OFF (with PSU plugged in the wall) or ON?

The black resistors going downwards from the 01C then across to the 015 or 510 are as follows:

Number - Calculated value - Actual
01C - 10K - 6.1K
18C - 15K - 6.4K
18C - 15K - 8.2K
688? - ? - 4.2K
303 - 30K - 13.5K
01C - 12.4K - 8.1K
510 - 51 - 51

So all are lower (rather than higher)

Means they are probably okay then.

Anyways, I'll have a look at the voltages you posted above tomorrow and try to make sense of them. It is getting late here (past 2 in the morning).

[socketa]

Re: ASUS ASUS P5VDC-MX Dead CPU?

The LPC doesn't get hot
With PC off, it goes from 17.7 to 18.5 deg C
With PC on, it goes to about 22 deg

[momaka]

Re: ASUS ASUS P5VDC-MX Dead CPU?

Okay, so the voltages that really worry me are "PWRON#" (pin 72) and "PWROK2" (pin 78).

PWRON# doesn't seem to change states at all. Unless it is programed to do something else (which according to datasheet page 37, is possible), I don't think it should stay at zero volts.

The most worrisome thing, though, is the voltage on PWROK2. It either has to be below 0.4 V or above 2.2 V. 1.22 V is right in the deadzone - i.e. the IC can arbitrarily interpret it as either "1" or "0" ... and if it's interpreting the wrong value, that could well be causing the motherboard to halt POST.

I think you need to trace where that pin goes, if possible. I know some ASUS motherboards had problems with a leaky transistor (or poorly designed circuit?) on the case intrusion switch detection causing the motherboard to halt shortly after POST. I wonder if ASUS fudged something with the design on this motherboard as well. As things get old, component values change a little due to aging, and maybe enough to make something not work if the circuit wasn't designed well to take proper tolerances into consideration.

Which reminds me... can you also see what the pins for front panel PWR LED are connected to? Maybe trace them out too? You said in post #1 PWR LED has 1.4V across its pins. I wonder if PWROK2 is perhaps supposed to pull down something in the front panel connector but is failing.

Lastly, there is a 5 volt 78L05 regulator (8-pin TSOP) to the bottom-right corner of the ITE chip. Check what voltage it is taking in and putting out.
Here is the TI datasheet:

Pin 1 should be Vout and Pin 8, Vin.

By the way, any history on the motherboard? Did you get it like this or was it working and then stopped while you were using it?

[socketa]

Re: ASUS ASUS P5VDC-MX Dead CPU?

Picked up the board recently (it came with a lot that included the same model board that i previously destroyed with the bad PSU).

The leg PWROK2 (pin 78) is not attached to any visible trace - you can see that in the previous photo (14 pins up from the bottom pin 64)

I examined the traces in the vicinity of the PWR LED - there is short trace on the reverse side that terminates at a dot
And on the front side (on the other terminal), i get continuity to a solder dot on "F_Panel", which also visibly goes nowhere else

The voltages on that 78L05 regulator are 5.7 and 11.87 respectively

[momaka]

Re: ASUS ASUS P5VDC-MX Dead CPU?

The leg PWROK2 (pin 78) is not attached to any visible trace - you can see that in the previous photo (14 pins up from the bottom pin 64)

It could be (and likely is) going behind the ITE chip and then into a via (i.e. a "dot", as you refer to it ). Yes, it takes a long time and some nerves to follow small traces like that, but that's how it is with more complex problems.

I examined the traces in the vicinity of the PWR LED - there is short trace on the reverse side that terminates at a dot

So, if you flip the board, does a trace continue from that "dot"? If not, then you set multimeter to lowest resistance scale and start probing nearby components until you get the lowest resistance reading your multimeter can show - that's when you know you've found where the trace re-appears again.

The voltages on that 78L05 regulator are 5.7 and 11.87 respectively - so it looks like it's doing it's job well enough.

5.7 Volts? Okay, either your multimeter definitely needs new batteries or it's not calibrated properly.

Linear regulators are pretty spot-on when it comes to their output voltage. The 78xx series have 5% tolerance. So you shouldn't see more than 5.25 V there at all. I am pretty sure the regulator is good, though.

Anyways, if bad comes to worse, and you hit a dead end or just don't care about troubleshooting the board further, perhaps we could "play" around with it a little and inject voltages (through a resistor, of course) at various locations (like PWROK2 and the CPU VRM PGOOD signal) to see what happens.

[stj]

Re: ASUS ASUS P5VDC-MX Dead CPU?

5.7 is too high,
is the battery good in your meter?

damn - beaten by seconds!

[socketa]

Re: ASUS ASUS P5VDC-MX Dead CPU?

OK, i checked with an unused battery and got exactly same result, so i'm going to use that linear regulator to calibrate the meter, and will give the battery a good charge just in case both batteries are low
So it looks like it's roughly 0.7V out of whack.
Yes, in post #1 i mentioned that there was 1.4V across front panel PWR LED pins - and that's still the case.
Further checking (with the black probe to ground) shows that there is 5.1V on the positive pin and 0.6V on the other pin (regardless of whether the board is powered up or not), but when i put the lead across both pins i get 1.53V (can i say wtf?)
When the board is powered up, that figure drops to 1.48V.

So, the -ve pin does not go directly to ground (i verified this by comparing with a good motherboard: MS-KMVM)
Since there is 0.6V on one pin, and 5.1V on the other, that should give a reading of 4.5V - Not 1.4V

So, i thought to see what happens on another, good, board that i have (MS-KM4M)
and repeated the same measurements

With PSU plugged in:
0.025 on positive pin
0.025 on negative pin
0V between both pins

With the motherboard powered up:
5.09V on positive pin
0.25V on negative pin
4.84V between both pins

Which (unlike the previous board) makes sense, because 4.84 + 0.25 = 5.09

I'm essentially only trying to get this board fixed as a challenge/learning experience - so if this is nearing "bad comes to worse" i'm happy to tinker with it, as you just suggested, and as in an earlier post, with regards to manually directing Power Good the CPU PWM Controller

[socketa]

Re: ASUS ASUS P5VDC-MX Dead CPU?

I told a lie - both pins terminate on the other side of the board as dots with no traces (forgot to reverse the image in my 'head')
However, i've found out that the positive pin of the Front panel Power LED has continuity with pin 68 VBAT of the ITE Chip.
And there must be something like a diode in between - because current only flows one way (with the positive probe of meter in continuity mode on the Power LED positive pin)
So, if it's a diode, current can only travel from the the LED pin to the ITE pin
Without powering up the mobo, there is 3.4V at pin 68
After powering it up : 3.22V