Weird but interesting issue with DELL 3542 (intel mb 13269-1)

[Spider1211]

Hey guys, need help fixing a weird issue with DELL 3542 with board cedar intel mb 13269-1

Schematic from here: https://www.badcaps.net/forum/showthread.php?t=52924

Backstory:
Laptop came in as completely dead. I opened it and removed the top half of the casing (as well as battery) and tried to power it again and it turns on and post. I turned it off, unplug charger, plugged back in and nothing. No boot, no nothing. I tried resetting CMOS but no dice.


Power supply draw:
Voltage: 19.40V
Amp draw: 0.006 - 0.007A

Amp draw stays within this range for all tests done:

Quick initial tests done:
-> Voltage on power button: 3.56V*more on that in the "weird" section further below
-> Check if power button press is being detected both on switch side as well as KBC side: all OK
-> Removed power and checked resistance on all coils: all OK
-> Replaced RAM with known good: same issue and still no boot

Proper testing using schematic power sequence:
-> All good from step -7 up to step -2. I am however missing RSMRST#_KBC

Further digging:
-> Since S5_Enable is present, I check for stable voltage (both with meter and DSO) on PU4503 (TPS51225)
3D3V_AUX_S5 (from 3V LDO) = 3.56V and stable
5V_PWR_2 = 5.13V and stable
3D3V_PWR = 3.31V and stable
5V_PWR = 5.11V and stable
Enable signal is stable
Feedback resistor network is fine

-> Check PSL logic:
- PSL_IN1# (ACOK# from charging IC) =0V
- PSL_IN2# = 3.56V (3.56V > 0V > 3.56V when PB pressed)
- PSL_OUT# = 0V

Check Q2402/3, both tests OK. 3D3V_AUX_KBC present.
ECRST# =3.56V

-> Checked RTC section
- +RTC_VCC = 3.1V
- RTC_AUX_S5 = 3.48V
- D2501 tests OK
- RTC_DET# = 0V
- RTC XTAL = 32.768KHz and stable
- RTCRST_ON = 0V
- RTC_RST# = SRTC_RST# = SM_INTRUDER# = 3.45V
- PCH_INTVRMEN - Not sure of its location on board

-> Checked BIOS
- Check for proper voltage being supplied = 3.56V
- Check data activity with DSO = Present
- Check BIOS resistors = all OK
- Check proper trace continuity from BIOS to SIO = all OK

Concentrating more on RSMRST#_KBC:
- Check for short on that line: ~11KOhm
- Check R1702 = 1KOhm
- Check Q1701 = Tests OK
- Check R1704 = 0Ohm
- Remove R1704 and check possible short on PCH_DPWROK and PM_RSMRST= No short.


Parts I changed:
1. PU4503 (TPS51225)
Reason: Since 3V LDO seemed a bit high (reached 3.75V at one point).
Outcome: No change

2. Q2402
Reason: Although it tested OK, this mosfet is a common failure of this model of motherboard
Outcome: No change

3. KBC24(SIO - NPCE285PA0DX)
Reason: RSMRST# is asserted by SIO. Conditions for its release seems OK.
Outcome: No change
Note: KBC was taken from a known good working device, so I know for sure it is OK.

4. Reflashed BIOS with known good firmware:
Reason: I read a somewhere that SIO needs valid data from BIOS chip before releasing RSMRST#
Outcome: No change
Note: I also replace BIOS chip itself with a known good chip and flashed with known good firmware.

5. PU4404 (BQ737 - Charging IC)
Reason: Possible power issue (following weird event as explained further down)
Outcome: No change

I did a few other tests but it'll be too long to post. BUT!

Here is where things gets very weird:

At this point am running out of things to test and out of frustration I kept cycling power on my bench supply in quick succession (ON - Off - ON with about 0.5 second between On and Off) and to my surprise, the board springs to life with a healthy current draw about 450mA.

Press the power button on the board to turn it off and idle current falls to 0.010A, press power button again and board boots normally! Connected display and it posts with no issue

To check if it was not a fluke, I disconnected my PSU and connected it again, pressed power button andddd nothing -_-. So I tried the power cycling again and it worked! Although at time you need to cycle the power as many as 15times.

And yes, once the board turns on atleast once, RSMRST# is asserted. Also, what I initially though was a higher than usual 3V LDO voltage (3.56V) drops down to a stable 3.30V.

So I tried a few combination:
1. Remove CMOS battery, cycled power until it boots. Once booted, I try a few ON and Off cycles using motherboard power button, and each time the board boots up just fine. It stays ON as long as the power button is not pressed, so its not like it only ON for a short while either.

2. Put CMOS battery and repeated same process as step 1. Again, it works just fine.

Conclusion after a few different tests:
- Whether or not the CMOS battery is present does not affect if the board turns on again after 1st boot. As long as power is not removed from the board after the 1st boot, the board boots (if its off) or turns off (if its ON) every single time the power button is pressed.

I suspect it might be a power related issue, but I have no idea what I should check next. Sorry for the long rant, but I really want to go down this rabbit hole to understand the cause (for learning purposes).

Any help is GREATLY appreciated!!! Thank you all in anticipation.

[mcplslg123]

Re: Weird but interesting issue with DELL 3542 (intel mb 13269-1)

@spider1211, you have already checked whatever is required for troubleshooting. However, have you observed PS_ID when it is turning on perfectly and when it is not?? No corrosion or liquid spill under or near SOC??

[mcplslg123]

Re: Weird but interesting issue with DELL 3542 (intel mb 13269-1)

LID Switch status in both scenario's??

[Spider1211]

Re: Weird but interesting issue with DELL 3542 (intel mb 13269-1)

No corrosion or liquid spill under or near SOC??

The board is actually spotless, barely even has dust on it. SOC area looks very clean with no traces of debris, moisture or corrosion.

I had another same model working laptop (without Discrete GPU though) and compared diode reading on BIOS chips and both look similar (I initially thought it could be an issue with the BIOS traces going to PCH).

LID Switch status in both scenario's??

LID Switch in both scenarios is the same, 3.56V when not working and 3.33V when working.

PS_ID as measure on pin 99 of SIO and charger connection to board is as follows:
When laptop not working:
Charger connection = SIO = 3.45V

When laptop working fine (first boot after power cycling):
Charger connection = SIO = 3.26V

When laptop is turned off using laptop power button:
Charger connection = SIO = 3.34V

When laptop is turned ON again using laptop power button:
Charger connection = SIO = 3.26V

Other observation:
When the laptop is initially off and refusing to boot using power button, I can hear like a faint coil whine coming near the GPU. The whining noise goes away once the laptop is booted.

Since the laptop is working fine once booted, I doubt its a GPU issue.

But I was thinking of doing the following since am out of idea at this point:
1. Resolder the coils in case it might be a dry joint.
2. Resolder switch IC and components close to it.
3. Do a UMA conversion (just testing purposes).

What do you guys think? What else can we test?

[mcplslg123]

Re: Weird but interesting issue with DELL 3542 (intel mb 13269-1)

Coil whine near GPU-i'll suggest to try uma option first.

[Spider1211]

Re: Weird but interesting issue with DELL 3542 (intel mb 13269-1)

Tried UMA but no change. Am I doing the conversion correctly?

How I did the conversion:
1. Remove GPU Coils, all mosfets and switch IC
2. Swap resistor from R2005 to R2008

Is there anything else I need to do?

I also noticed that I can still hear a faint coil whine even with the GPU supply disabled.

Until the above is confirmed, I reverted back to discrete and checked if laptop is still working after power cycling (It is working fine).

Further test:
1. Remove 3.3V and 5V coils (PL4502 & PL4501 respectively)
2. Injected 3.3V on 3.3V rail (system side) and current draw is 0.001A
3. Injected 5V on 5V rail (system side):
-> Initially current draw is almost 0, but after a few cycles current draw jumps to ~420mA and PCH gets warm.

-> Remove power and apply again, same conditions observed as described above (initially 0 the jumps to 420mA after a few cycles)

4. Solder back 3.3V and 5V coils and try to boot laptop normally (to check if I did not kill anything). Laptop boots and posts after the usual power cycling.

My conclusion:
Possible leaky mosfets somewhere down the 5V rail?

[Sametbey]

Re: Weird but interesting issue with DELL 3542 (intel mb 13269-1)

"How did I transform:
1.Remove GPU Coils, all mosfets and replace IC
2. Change the resistance from R2005 to R2008 "

According to the document I have, it should be sufficient. Can you process the measurements on the photo by taking a high resolution photo of the mainboard.
for example.

[mcplslg123]

Re: Weird but interesting issue with DELL 3542 (intel mb 13269-1)

I don't have any further ideas@spider1211. may be soc dry solder issue.

[Spider1211]

Re: Weird but interesting issue with DELL 3542 (intel mb 13269-1)

@sametbey, I'll try to do it, but as mentioned, all voltages are present once laptop is booted.

@mcplslg123, I initially thought the same thing but the fact that it boots everytime after the first boot (even if you let it cooldown) is leading me to think otherwise. I tried to keep the board as long as possible since I really want to get to the bottom of this but will need to return it this weekend. So I guess this one will go undiagnosed and remain a mystery

I also tried heating up the SOC but no luck. I don't really want going to reflow it since they tend to bend if unevenly heated/cool. Anyway, thank you guys for your input

[mcplslg123]

Re: Weird but interesting issue with DELL 3542 (intel mb 13269-1)

Yeh, sometimes we've to let it go without being sure what was wrong. Its part of business ...sometimes.

[looofy]

Re: Weird but interesting issue with DELL 3542 (intel mb 13269-1)

did you check 3D3V_S0 impedance with ohm with ground?
if not you can check pin 8 ,9 for u3601, it is around 400 ohm.
if its too low then the pch is . i hope its not

[Spider1211]

Re: Weird but interesting issue with DELL 3542 (intel mb 13269-1)

@looofy; As mentioned, all voltage rails are present once the laptop boots. There definitely was no short on 3D3_S0 when I tested and if I remember correctly, resistance was around 400ish ohm.

Its also obviously not a dead PCH since the laptop boots. I was tempted to do a reflow but got caught up with other repairs and sadly had to return the board.

[looofy]

Re: Weird but interesting issue with DELL 3542 (intel mb 13269-1)

@looofy; As mentioned, all voltage rails are present once the laptop boots. There definitely was no short on 3D3_S0 when I tested and if I remember correctly, resistance was around 400ish ohm.

Its also obviously not a dead PCH since the laptop boots. I was tempted to do a reflow but got caught up with other repairs and sadly had to return the board.

i faced much of such type of this mb i just remove D2501 , i guess u will know the rest

[c.daniel]

Re: Weird but interesting issue with DELL 3542 (intel mb 13269-1)

@Spider1211: hi; i have exactly the same laptop, and i have weird voltage 3.7 and 5.2 instead of 3.3 and 5; after many hours, i found it the guilty ic; it was U3504, a protection chip for usb .... i never think those chip gonna messing up the voltage....it has 0,7ohm between pin 1 and 5 .... it blow my mind; take me aprox. 7 hours to find it ....

[Spider1211]

Re: Weird but interesting issue with DELL 3542 (intel mb 13269-1)

@c.daniel, thats interesting. I wish I still had the board to test. How did you eventually find the fault? Can you please share your diagnostic approach.

[c.daniel]

Re: Weird but interesting issue with DELL 3542 (intel mb 13269-1)

Man, yes, i can share; i'm at begining with repair, so my method was absolutely dodgee; here it is:
- i attach multimeter, on voltages, on rail 3,3volts, wich it has 3,7volts and let it do the reading
- i begining to heating up the mainboard, with hot air station, low flow, very hot air; and, i walk around the mainboard, to see where those 3,7 volts beginning to change value
- after i find the area, from 3,7 going up to 3,73 volts, i cool down with vacuum cleaner, and check again with 60watts soldering iron, on chip capsule, to find exactly what chip making the "3.3" to change
- all the method took me kind of 3-4 minutes .... but, after i spend almost 7 hours trying....
- but, me is verry happy, and that's is all about

[c.daniel]

Re: Weird but interesting issue with DELL 3542 (intel mb 13269-1)

Offtopic: i check on google where is Morcellement; man, you live exactly in the middle of water )) nice place.

[Spider1211]

Re: Weird but interesting issue with DELL 3542 (intel mb 13269-1)

Sometimes dodgy is the way to go

Yes, Mauritius is an island and a tiny dot on the map but has beautiful mountain ranges for hiking and amazing seaside

[mcplslg123]

Re: Weird but interesting issue with DELL 3542 (intel mb 13269-1)

@spider1211, got very similar problem on same mb(uma). Changing the caps on 3v,5v regulator fixed it. I related them with tantalum caps. Used to turn in 1 out of 10-15 attempt.

[Spider1211]

Re: Weird but interesting issue with DELL 3542 (intel mb 13269-1)

@mcplslg123, ahhhh those tantalum strike again! . I'll try to contact the dude from the shop and see if they can have the device back if it still hasn't been fixed elsewhere. It still bugs me to not have found the root issue