ASUS ASUS P5VDC-MX Dead CPU?

[momaka]

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')

That means the traces are going through internal layers. Could be up to 8 on a motherboard like that, so that does make tracing much harder (though not entirely impossible if you are really hard headed with the thing )

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)

You mean pin 69, VBAT?
Pin 68 is "COPEN#" (chassis intrusion detection... that's why I always give the name and the pin # as well .) "COPEN#" should have continuity to ground.

Speaking of which, don't always trust continuity function on your multimeter. Just because you hear a beep doesn't really mean that there is actually a good connection. Most multimeters will beep at less than 20-100 Ohms - and this threshold really varies with different multimeters. Let's say it beeps at 15 Ohms - that is hardly a good connection. Thus, always double-check with lowest resistance scale as well. "Good continuity" would be less than 2 Ohms (provided your meter can measure that low, which it should if it doesn't have dirty dial contacts and is calibrated properly).

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

Double-check this. I don't think VBAT should be connected to anything on the front panel. It's NOT a general purpose (GP) register either (see page 27 on datasheet), so it can't be configured for anything else.

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

Are you using a rechargeable batteries for your multimeter? If yes, that could be the problem too. Before messing with your meter's calibration, first try some good disposable batteries.

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

Sounds good . Just don't insert voltages directly from a power source. Always do it through a resistor. 1 to 10 KOhm should be suitable for this. Preferably 10 KOhm. And I would try that with PWROK2 on the ITE chip, first, as it seems to be getting some strange voltage.

[stj]

Re: ASUS ASUS P5VDC-MX Dead CPU?

ni-cad / ni-mi pp3's are usually only 8.2v

[socketa]

Re: ASUS ASUS P5VDC-MX Dead CPU?

Thanks

Yes, pin 69 VBAT - it's a hard, rectified, short 0.0 ohms

Recharged the new battery overnight, and still get the same reading
So turned one of the pot screws (v1) inside the meter so that it read 5.25V from the output of that linear regulator, which then measured 3 ATX supplies to be under spec
Returned the pot back to it's previous position, where VBAT = 5.7V
Got a brand new non-rechargeable battery, and it made absolutely no difference.

[socketa]

Re: ASUS ASUS P5VDC-MX Dead CPU?

I just put the negative probe on pin 69 VBAT, and get a rectified short to all the other terminals of that chip.
Perhaps that's useful info?

Also tried looking for another similar regulator chip on a good board to use as a verification for the meter, but none of them were linear regulators.

[stj]

Re: ASUS ASUS P5VDC-MX Dead CPU?

what the hell is a rectified short?

[socketa]

Re: ASUS ASUS P5VDC-MX Dead CPU?

i remembered reading somewhere that a diode, by itself, can technically be classified as a rectifier.
Because current only flows when the source is at one polarity, but not the opposite
Made sense to me
But won't use that phrase again, if it's going to raise eyebrows

Anyway, all i was saying is that there is short, and when the leads are interchanged, there is an open circuit

[momaka]

Re: ASUS ASUS P5VDC-MX Dead CPU?

i remembered reading somewhere that a diode, by itself, can technically be classified as a rectifier.

Correct.
However, when you call a diode a rectifier, that usually implies (relatively) high power and/or current... i.e. not used for signaling but to power something.

But won't use that phrase again, if it's going to raise eyebrows

Yes, don't
short -> short-circuit
rectified -> current or signal passes only one way

so you are saying you have a short-circuit only one way... which I think I know what you are trying to say (that you get a reading one way but not the other on your meter?), but it is technically incorrect to say it that way.

Which brings me to this:

Anyway, all i was saying is that there is short, and when the leads are interchanged, there is an open circuit

That shouldn't be possible. If you have a real short-circuit one way, you will have it the other way too. Again, a short-circuit is a relative term. 100 Ohms is not a short-circuit, but if you are dealing with power lines with 100's of KVolts... then yes, it is pretty much a short-circuit. On low-voltage electronics like with what we are dealing here, short-circuit would be defined as less than 2 Ohms resistance, and "good" continuity as less than 5-10 Ohms.

My guess is you are trying to say that you get a reading on your multimeter one way when you measure those test points but not the other? If that's the case, then this is not a short-circuit. You are just seeing a diode in the circuit, that's all.

Or, if you happened to be measuring across some caps of large capacitance, sometimes your multimeter will show low resistance reading for a moment before the caps charge up. If you reverse the leads, the multimeter could even show a dead short-circuit for about a second or so. This is because the caps are charged one way and your multimeter is trying to insert and measure the voltage the other way to determine the resistance. The end result is that the meter gets fooled for a few moments until the caps charge again, this time the other way. So if you do get in a situation like that, then leave the multimeter probes for a bit longer until the measurement settles to a steady value.

I hope all of this text makes sense, as I did do a bit of speedy typing this time. Let me know if not.

[socketa]

Re: ASUS ASUS P5VDC-MX Dead CPU?

My guess is you are trying to say that you get a reading on your multimeter one way when you measure those test points but not the other? If that's the case, then this is not a short-circuit. You are just seeing a diode in the circuit, that's all.

That's right.

To make it very clear (hopefully)
I turn selector on the meter to Ohms
Put black probe on pin 69 VBAT
Put red probe on any of the other pins on the same chip
Each time, the reading on the display is 000.0
The reading remains at 000.0 as long as the the probes remain in place

Now, if i put the red probe on pin 69 VBAT
And put the black probe on any of the other pins on the same chip
The reading displays 0L - or in other words, no load, or open circuit

(and to back up my previous thought process/logic: When i'm swapping the position of leads with my hands, i'm effectively creating an AC generator - even if it's only one cycle, and the output waveform on an oscilloscope that is inserted into the circuit would demonstrate that the current from the multimeter is being subject to a half-wave rectification, and the max voltage would equal the source voltage minus any negligible forward voltage drop across the diode (or whatever is creating this diode-like effect), i.e. a short circuit)

Also have double and triple checked that there is continuity from Power LED pin to Pin69 VBAT, and that there is something like a diode in between.
Short circuit when red probe is connected to positive pin of Power LED pin
Open circuit when probes are reversed

[stj]

Re: ASUS ASUS P5VDC-MX Dead CPU?

your just seeing the diodes or transistors inside a chip with your so-called rectified short.
and your meter is not set/working right!

[socketa]

Re: ASUS ASUS P5VDC-MX Dead CPU?

Ok, so i dug out an old analog hioki meter, and notice that i was getting more than a full scale deflection when taking those readings again, and immediately recognized this as voltage.
No one told me to remove the BIOS battery.
Might have spotted it earlier if the datasheet mentioned the word BIOS for pin 69
And in hindsight, it's pretty obvious now that it's not a good idea to test for continuity with the BIOS battery still connected
Now that i have removed the battery, neither meter displays a short/diode.

[stj]

Re: ASUS ASUS P5VDC-MX Dead CPU?

lol

[socketa]

Re: ASUS ASUS P5VDC-MX Dead CPU?

Yeah, thought that there might have being a chance of getting a lol from that

Also, no one told me to take any resistance measurements either, so this wouldn't have happened if i hadn't got a bit too curious

[stj]

Re: ASUS ASUS P5VDC-MX Dead CPU?

have you got the datasheet for the bridge chip(s)?

you need to trace the battery pad back to the battery holder.
there will be a diode inline - maybe other stuff too.

[socketa]

Re: ASUS ASUS P5VDC-MX Dead CPU?

No, haven't checked that out yet.

Just found another, budget, multimeter
This one displays 5.59V for the 78L05 regulator output after putting the new battery in it
The Hioki display reads spot on 6V
And the original meter displays 5.7V

So now that's 3 meters that are apparently reading wrong

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 measured this again with the meter that i just found and get 1.21V
So now there is a couple of last resort options to try:
Inject a voltage onto that pin, or the Power Ok pin of the CPU PWM controller chip

[momaka]

Re: ASUS ASUS P5VDC-MX Dead CPU?

Ok, so i dug out an old analog hioki meter, and notice that i was getting more than a full scale deflection when taking those readings again, and immediately recognized this as voltage.
No one told me to remove the BIOS battery.

Oops, sorry about that. I assumed you removed it, and I shouldn't have. Then again, if anything, you learned something new here, didn't you?

Might have spotted it earlier if the datasheet mentioned the word BIOS for pin 69

Yes, but the datasheet isn't very specific, unfortunately. Really needs a careful looking. I went through the entire thing twice just to give you the test points above and there is still a lot that I didn't understand.

And in hindsight, it's pretty obvious now that it's not a good idea to test for continuity with the BIOS battery still connected
Now that i have removed the battery, neither meter displays a short/diode.

Good.

Also, no one told me to take any resistance measurements either, so this wouldn't have happened if i hadn't got a bit too curious

Well, don't assume we know everything (we don't... or at least I don't). So with a lot of problems like this it is looking up datasheets and trying to think creatively .

Just found another, budget, multimeter
This one displays 5.59V for the 78L05 regulator output after putting the new battery in it
The Hioki display reads spot on 6V
And the original meter displays 5.7V

So now that's 3 meters that are apparently reading wrong

I don't know what to say here.
Three meters showing that high. Maybe that 7805 regulator really is outputting that high of a voltage. Or maybe it is somehow throwing off the meters.

I guess the best bet would be to find another motherboard with a 7805 regulator (doesn't matter if the board is working or not) and measure its output voltage with the meters again. If you still get a high voltage like that... then those three meters are all wrong... which is highly improbable, but not impossible I suppose. And if they are not wrong, then that 7805 regulator should probably be replaced. Depending on what is connected to it, 5.7 V could be enough to blow some chips. I know the FIC AM37 motherboard (used in many eMachines) had some 5 V logic gates rated only up to 5.5 V. These gates were sitting on the 5 VSB rail. What's worse is that eMachines often used the Bestec ATX-250 12E power supply in those PCs, and that power supply is known to get bad 5 VSB, shooting well up to 9V+. So you can imagine what happened to many of these motherboards.

By the way, is your house very old? Maybe some ghosts are playing tricks on your gear.
Frankly, I've never seen anything like this before (a linear reg being that far off without being bad).

[socketa]

Re: ASUS ASUS P5VDC-MX Dead CPU?

Yeah thought it was all getting a bit beyond it.
So today, i tried my skill/luck on another board - an intel D101GGC with a fan that only twitched, which indicated PSU short circuit protection kicking in.
Pulled motherboard input lytics off, and short was still present (which wasn't a surprise), so found a mosfet that was shorted from drain to gate, and source to gate.
Replaced it with one from from the ASUS P5VDC, and put the input caps back in after checking their capacitance and ESR.
Replaced the bulging output 3300uf lytic
Now CPU fan spins and power LED comes on, but get no voltage output from any of the 4 VRM mosfets, no BIOS beep, and also the board wont power down by pressing the power button.
So i checked for low resistance at the ATX motherboard connector, and found only one - the 5VSB which was 47 ohms to ground.
Then i compared that with a good board and can see that this is probably not done by design.
Any idea what's going down here?
Fried south bridge?

By the way, is your house very old? Maybe some ghosts are playing tricks on your gear.
Frankly, I've never seen anything like this before (a linear reg being that far off without being bad).

Na, house is not too old, and no one here has glimpsed any ghosts recently

i've got a few boards and would love to verify that (possibly odd) linear regulator by comparing the voltage reading with another 8-pin linear regulator on another board.
Have checked on about 5 boards and, after checking the component datasheets, haven't found one similar- looking linear regulator on any of them.
Also had a looksie in a couple of PSUs, but was unable to sight any.
Are they hard to come by?

[momaka]

Re: ASUS ASUS P5VDC-MX Dead CPU?

Yeah thought it was all getting a bit beyond it.
So today, i tried my skill/luck on another board - an intel D101GGC with a fan that only twitched, which indicated PSU short circuit protection kicking in.
Pulled motherboard input lytics off, and short was still present (which wasn't a surprise), so found a mosfet that was shorted from drain to gate, and source to gate.
Replaced it with one from from the ASUS P5VDC, and put the input caps back in after checking their capacitance and ESR.

Should have done some experimenting with that P5VDC before using it for parts.

But I guess it is too late now (unless of course, you put back the MOSFET you borrowed from the P5VDC). Anyways, I guess make another thread for the Intel motherboard, and let's see where we can get with it. Would be confusing to have it in this thread.

i've got a few boards and would love to verify that (possibly odd) linear regulator by comparing the voltage reading with another 8-pin linear regulator on another board.
Have checked on about 5 boards and, after checking the component datasheets, haven't found one similar- looking linear regulator on any of them.
Also had a looksie in a couple of PSUs, but was unable to sight any.
Are they hard to come by?

Shouldn't be on older (socket 478/462) motherboads, especially ECS and Intel.
That said, they may not be in a 8-pin SOIC package like on this motherboard. In fact, ECS often used 7805 regulators in TO-92 packages.
(If these component package numbers don't mean anything to you, just Google them and you will see what they look like.)

If you can't find a 7805 regulator, perhaps the next best thing would be a 1084 linear regulator. Those are very common, especially on older hardware. They mostly come in a TO-263 package (a.k.a. D2PAK or DDPAK). Just be aware that they have different output voltages. A 1084 with a "33" suffix in the name usually indicates 3.3V output, "25" indicates 2.5V, "18" is 1.8V, and I think "A" or "ADJ" is for adjustable (i.e. it is determined by a network of resistors).

[socketa]

Re: ASUS ASUS P5VDC-MX Dead CPU?

Thanks for suggesting those alternatives,
Didn't find any more 5V linear regulators in any package - even checked a few socket 745 and AM2 boards - bit of a bugga
But did find a LM1086 CS-3.3, and an APL1084 on 2 different boards

(Measurements included in attachment)

To me it's looking like there there was something odd going on with that 7805 regulator, because (if the various meter readings of these other two regulators are any thing to go by) these meters are not giving readings that are suggesting that any one of these meters are at least 0.45V out of calibration (Taking 5.7 to be the actual reading, and 5.25 to be maximum that the 7895 chip should output).
Or perhaps there was a stray voltage that was coming from the other direction?

P.S Can't stick the mosfet back on that board because the chip cracked and the leg came off of one when i was removing it, so then i successfully removed the one next to it.

[momaka]

Re: ASUS ASUS P5VDC-MX Dead CPU?

To me it's looking like there there was something odd going on with that 7805 regulator, because (if the various meter readings of these other two regulators are any thing to go by) these meters are not giving readings that are suggesting that any one of these meters are at least 0.45V out of calibration (Taking 5.7 to be the actual reading, and 5.25 to be maximum that the 7895 chip should output).

Indeed.
Makes one wonder if that 7805 was the cause of the motherboard not booting.

That said, now that we have given this regulator more attention, I did notice that there are some free spots where components should have been installed next to that 7805. In particular, one of them looks like it is for a ceramic cap array, and it is right on the input of the 7805... kind of like the tan ones near the keyboard connector in the lower-left in this picture:
https://www.badcaps.net/forum/attach...6&d=1445215441
Perhaps when the board was new, everything was just marginally stable, and now that components have deteriorated over time, the 7805 is not getting good enough filtering on its input. You can see on page 9 of its datasheet that there needs to be a cap with 0.33 uF (minimum) on the input if the regulator is more than 3" away from another power supply filter.

Then again, the output of that regulator likely goes to the ALC653 audio IC (might need to do some tracing there), so I don't see how that can prevent the board from booting. But it does raise a question where else ASUS could have cut corners, possibly making the design on this board only marginally stable.

P.S Can't stick the mosfet back on that board because the chip cracked and the leg came off of one when i was removing it, so then i successfully removed the one next to it.

Is it a MOSFET from the CPU VRM that is in parallel with another MOSFET? If yes, you can sometimes get away with just one MOSFET in there (provided the CPU is not a Pentium 4 or Pentium D or some other power-hungry CPU.) I'd need to see a picture to tell you for sure, though.

[socketa]

Re: ASUS ASUS P5VDC-MX Dead CPU?

Perhaps when the board was new, everything was just marginally stable, and now that components have deteriorated over time, the 7805 is not getting good enough filtering on its input.

Which components would that be? Nearly all of the caps are FL and FJ and none of show signs of bloating

I went went back to the boards that i checked for the 8 pin 5V regulator, and found a GS78L05S regulator.
Link for GS78L05XX

The board that it is on, is an ASUS P5GL-MX
(The chip is on the bottom left hand side next to the PCI slots)

Stuck a heatsink on it, booted the board, and then checked the chip input and output voltages with all 3 meters:
Digitech : 11.78/5.02
El Cheapo : 11.6/4.95
Hioki : 11.75/5.00

Looking at the datasheet, i see Min/Max range of 4.8V - 5.2V
So hopefully this is verification that my meters are good enough.

When i first tested at the input pin, there was a spark at the probe lead, and a "pop" sound (ever got spark off of a multimeter probe when taking a voltage measurement?) - thought that i might have blown the chip, but it still booted into BIOS OK and the chip was still outputting 5V.

Could this spark produce an audible "pop" sound?

The last time that i heard a "pop" sound, it was a south bridge getting fried.

Is it a MOSFET from the CPU VRM that is in parallel with another MOSFET? If yes, you can sometimes get away with just one MOSFET in there (provided the CPU is not a Pentium 4 or Pentium D or some other power-hungry CPU.) I'd need to see a picture to tell you for sure, though.

So there is now two vacant MOSFET positions on the P5VDC-MX VRM area and they are in parallel with each other (i.e. a hard short from drain to drain and hard short from source to source)

But since the MOSFET is on the other board, and caps have being resoldered, i might start a thread for that now; and if it proves to be unfixable, will grab the MOSFETS off of that.