Bad Laptop PCH?

[Spider1211]

Hey guys, I need some help with troubleshooting a laptop.

Motherboard: ASUS GL552GX rev2.0

Issue:
- Turning on but no display and no drive activity.
- Fan starts slow then goes to full power. (It does throttle at time.)
- Laptop turns off after a 30sec to 1min.

Power draw:
Power supply set to 19.7V
Off state: fluctuates between 9 to 12mA
On state: around 1.6A

Measurements:
- Check for any shorts -> none found
- Check voltage rails (all main rails present):
VCore: 1.77V
VGpu: 0.9V
RAM: 1.35V
Solid 3.3V & 5V always

Other measurements:
OFF State:
BUF_PLT_RST#: 0V
H_DRAM_PWRGD: 0V
H_CPUPWRGD: 0V

ON State:
BUF_PLT_RST#: 3.3V
H_DRAM_PWRGD: 3.3V
H_CPUPWRGD: 1.2V

RTCRST#: 3.3V (Both is ON or OFF state)

All steady voltage levels

Clock:
Stable 32.768KHz both in on and off state
Stable 25MHz in ON state


Here is where things get weird:
- Probing with Oscilloscope. (Laptop in OFF state)
-> Constant 5.53Hz on all BIOS pins except ground
-> Constant 5.53Hz on PM_RSMRST#, ME_SUSPWRDNACK,
ME_ACPRESENT (but inverted waveform).

- Probing with Oscilloscope. (Laptop in ON state)
-> All lines that were at 5.53Hz are now solid 3.3V
-> No data transfer observed on BIOS pins. (data lines go straight to 3.3V from
5.53Hz upon pressing power button.



Other observation
- When HDD is connected to motherboard, there is no activity and LED stays solid ON and drive does not spin up (Voltage at connector is present)
HDD was tested independently and works without issue.


What I have done so far with no luck
1. Flash BIOS from another dump -> No luck, still 5.53Hz
2. Replaced BIOS chip
3. Replaced EC and flash from another dump.
4. Erase EC and let device reflash EC (tried with both Original and downloaded BIOS)


Although there is no backlight, screen can be seen going dark when laptop is turned on and changes back to a lighter color when powered down.

Further info will be added if anything was missed. Am I correct in thinking its potentially a bad PCH? Or is it just a weird BIOS issue. I am unable to find a schematic for this motherboard and only managed to find a repair guide and boardview for what was listed as rev1.0 (although it also seem to match rev2.0).Relevant documents/files & bios dumps are attached.

******************************
MOD EDIT: Bios + EC moved to a new thread: https://www.badcaps.net/forum/showthread.php?p=982393

[emestee]

Re: Bad Laptop PCH?

These look suspiciously like PWM pulses, however this is also what you'd see if the EC is cycling resets. The inversion of ME_ACPRESENT may be a hint here - it goes low, reset goes up. You might want to poke whatever's generating it. I'd start with rail resistances to ground.

1.6A must be going somewhere, is the PCH warming up significantly?

[Spider1211]

Re: Bad Laptop PCH?

These look suspiciously like PWM pulses

My thoughts exactly. But it does not make sense that the same signal would appear on seemingly unrelated paths.

However this is also what you'd see if the EC is cycling resets. The inversion of ME_ACPRESENT may be a hint here - it goes low, reset goes up. You might want to poke whatever's generating it.

If EC was cycling resets, wouldn't it not boot at all? The cycling is only visible when laptop is in OFF state (only AC adapter plugged in). Once the laptop is turned ON, all cycling stops and goes to a steady 3.3V.

I'd start with rail resistances to ground.

1.6A must be going somewhere, is the PCH warming up significantly?

No shorts were found. Also, when laptop is turned ON, all rails are present and stable.

1.6A is the total current the board is taking when turned ON. (It actually starts at 1.3A and gradually increases to 1.6A). Which should be normal if all rails are turning on? (and also given its a gaming laptop, I'ld expected a higher current draw than "normal" laptops)

PCH does get hot (I'ld estimate in the ballpark of 50-60C), but does have a heatsink on it.

[emestee]

Re: Bad Laptop PCH?

Looking at the power on sequence, during power on the signals you see pulsing are preceded by SUS_PWRGD. If it (or something it depends on) oscillates, then you may be looking at a failing autostart which should not be happening - check what's happening on the power button as a start. Power buttons often have pull up resistors and a contact jitter capacitor. These may be misbehaving.

You also mentioned that in S5 *all* BIOS pins are pulsing. Do you mean literally all, as in VCC & GND? Pulsing VCC would have implications.

Let's all thank Asus for being so helpful.

[Spider1211]

Re: Bad Laptop PCH?

Looking at the power on sequence, during power on the signals you see pulsing are preceded by SUS_PWRGD. If it (or something it depends on) oscillates, then you may be looking at a failing autostart which should not be happening - check what's happening on the power button as a start. Power buttons often have pull up resistors and a contact jitter capacitor. These may be misbehaving.

The power button is part of the keyboard keys and is integrated in its flex cable. I traced its connection back to its test point on the motherboard which I short using tweezers. That eliminates a potentially bad power button.

Probing the power button test point, I do have a steady 3V there. Both in ON & OFF state.

You also mentioned that in S5 *all* BIOS pins are pulsing. Do you mean literally all, as in VCC & GND? Pulsing VCC would have implications.

Yes, all except GND. VCC pulsates only in S5 state but goes to steady 3.3V(together will all other pins except ground ofcourse) in S0 state. I did check for shorts/damage nonetheless but everything test/looks OK.

[Spider1211]

Re: Bad Laptop PCH?

So I tried powering up the laptop without the BIOS. Although the laptop will obviously not boot, I do have healthy signals on the BIOS pins. No more 5.54Hz nonsense as well as having a stable and steady 3.3V on BIOS Vcc.

I know its definitely not a faulty BIOS chip since I already tried 3 chips (one of which is from known good from another working laptop). Whenever the BIOS chip was replaced, the same 5.54Hz issue came up again.

Is there anything else I can test before replacing the PCH?

Am trying to avoid the option of replacing since I am currently very low on flux and unable to procure any due to current curfew which is going to last another month.

[RJARRRPCGP]

Re: Bad Laptop PCH?

Vcore way out of tolerance!

[piernov]

Re: Bad Laptop PCH?

CPU VCore is fine.

[jasko_jacker]

Re: Bad Laptop PCH?

On these models it is possible that the problem is the PCH, However, if you turn on the motherboard without the battery, checks are performed each time it is turned on (ec) and it is possible that the pch is hotter than normal.. In these cases I use a card (inserted on the ram slot or using the jdebug port if it exists on the card) to understand if the bios works. With regards to pch it is not easy to make a diagnosis especially if a secondary line is damaged.

[Spider1211]

Re: Bad Laptop PCH?

I did a few more tests and am inclining more towards a bad PCH. Grateful if anyone can provide additional information and/or tests that could help narrow down the issue.

Based on the attached boardview, I removed the resistors connecting the BIOS to the PCH and the following was observed:
1. The 5.54Hz issue went away
2. Laptop turned ON by itself upon connecting power. When forced in off state using power button, it just goes off very briefly and restarts and stays on.
3. WLAN indication LED stays on the whole time.

I then proceeded to resolder the resistors one by one and testing at each step.
Results are as follows:

1. R2406 - SPI_CLK_PCH (Goes to Pin 6 of BIOS) -> No issue
2. R2410 - SPI_SI_PCH (Goes to Pin 5 of BIOS) -> No issue
3. R2411 - SPI_SO_PCH (Goes to Pin 2 of BIOS) -> No issue
4. R2413 - SPI_HDIO2_PCH (Goes to Pin 3 of BIOS) -> No issue
5. R2401 - SPI_HDIO3_PCH (Goes to Pin 7 of BIOS) -> Issue as original post comes back.

No issue means no 5.54Hz as initially described, but boots by itself

Power draw at each step:
step 1,2 -> 0.374A @19.7V; VCore 1.70V, VGpu 0.90V (PCH, CPU & GPU barely warm)
step 3 -> 0.923A @19.7V; VCore 1.77V, VGpu 0.90V
step 4 -> 0.780A @19.7V; VCore 1.77V, VGpu 0.90V
step 5 -> same as initial issue in original post

[jasko_jacker]

Re: Bad Laptop PCH?

Can you measure X2101 and +V1_05S_PCH_VCC ?

[Spider1211]

Re: Bad Laptop PCH?

Can you measure X2101 ?

Its a stable 25MHz when the laptop is turned on, please see "clock" in the first post.

Same for X2001 which is a stable 32.768KHz both in ON and Off state.

+V1_05S_PCH_VCC
0V in off state
stable 1.06V in ON state

[Spider1211]

Re: Bad Laptop PCH?

Whats confusing me is the fact that everything seems to be turning on, etc. Especially VCore. Am still fairly new to signal level troubleshooting and looking forward to learning and understanding more. Any help/explanation received is greatly appreciated.

Would a faulty PCH(even if it was partially faulty) that does not seem to be able to communicate with its BIOS chip (none can be observed on oscilloscope) still allow such a scenario?

CMOS battery is removed and residual power is drained before measuring on oscilloscope. Measurements are also taken before battery is plugged in, as well as during a second boot after the battery was first connected and laptop powered ON.

[piernov]

Re: Bad Laptop PCH?

PCH sends out PLT_RST# so at least it's working until that point. Obviously it can be partially dead. But could be CPU too, or even EC (but you took care of that already), or something else random somewhere like a knocked off resistor.

Once PLT_RST# comes up, CPU starts processing instructions from the reset vector, which is not mapped to RAM but rather makes a request to the PCH to read the BIOS. PCH will send a request on SPI bus (depends on its configuration, could be another bus, there are strapping resistors for that IIRC) to read SPI ROM data, flash descriptor to locate BIOS region then first BIOS instruction. SPI ROM replies back and the instruction gets sent back to the CPU, which processes it and asks for the next instruction (increments address). After doing low level init of CPU, BIOS prepares memory controller, PCH and EC, so it'll try to communicate to the EC through the LPC bus connected to PCH.

So basically you should be observing some activity on SPI ROM CS# pin, then some activity on LPC LFRAME# line.
Also before sending out PLT_RST# the PCH will try to read SPI ROM at some point to initialize ME, and read the soft straps. But even if communication doesn't occur it doesn't necessarily prevent PLT_RST# to come up I think (I may be wrong on that though).
You should at least see at some point a falling edge on CS# when the PCH tries to read something from the SPI ROM. When the machine turns on, CS# pin will go high at the same time as the 3V power rail for SPI ROM comes up, then PCH pulls CS# pin to ground when it wants to talk with SPI ROM.

Some machines need the CMOS battery to be present and above 3V to POST (or even turn on), that's rare but I already wasted time on those so always with a good CMOS battery (and without as well).

[Spider1211]

Re: Bad Laptop PCH?

So basically you should be observing some activity on SPI ROM CS# pin, then some activity on LPC LFRAME# line.

You should at least see at some point a falling edge on CS# when the PCH tries to read something from the SPI ROM. When the machine turns on, CS# pin will go high at the same time as the 3V power rail for SPI ROM comes up, then PCH pulls CS# pin to ground when it wants to talk with SPI ROM.

Those are constantly High. LPC LFRAME# is low while laptop is off but goes high and stays high once turned ON. SPI ROM CS# is at 5.54Hz when off and steady 3.3V when turned ON, no communication observed.

Some machines need the CMOS battery to be present and above 3V to POST (or even turn on), that's rare but I already wasted time on those so always with a good CMOS battery (and without as well).

CMOS battery tested OK.


Referring to first post regarding H_CPUPWRGD, is 1.2V a valid level for this signal or is it low?

[Spider1211]

Re: Bad Laptop PCH?

What would be symptoms of a dead CPU?

[Spider1211]

Re: Bad Laptop PCH?

Latest update:

I eventually ended up changing the PCH, but the problem is still the exact same thing! So I guess the PCH was working just fine and wasn't the issue after all?

Any other ideas/suggestions? Do I have a possibly dead CPU as Piernov suggested? I will try requesting another BIOS just in case....

[RJARRRPCGP]

Re: Bad Laptop PCH?

Was the CPU Vcore at 1.77V? That's way too much for Intel, unless it's a Willy!

[Spider1211]

Re: Bad Laptop PCH?

Was the CPU Vcore at 1.77V?

Yes, and still is.

[piernov]

Re: Bad Laptop PCH?

See post #8. That's Haswell.

H_CPUPWRGD at 1V or so should be ok.