820-3437B dead CPU?

[Bouba148]

Hi 2Informaticos and All....
board came zith serious WD; cleaned and repair all what I could what was visual before plugging the charger; then:

Ppbus_g3h : 8.19V
PP3V42_h3h: 3.423V
PPVRTC_G3H: 3.423V
PP5V_S5: 0V

Reheat the SMC then I saw sparks almost flames around the C3085 (thunderbolt) area

Removed U3210 and C3084 C3085 C3096 qnd C3097 because of short to gnd
PP15V_TBT at 8.61V !!!!!!!! so I've also removed Q3080 as Pin 1 and 4 = 8.61V

Ppbus_g3h : 8.61V
PP3V42_h3h: 3.420V
PPVRTC_G3H: 3.330V
PP5V_S5: 5.129V
PP3V3_S5: 3.331V
PP5V_S4RS3: 5.130V
PP3V3_S4: 3.331V
PP3V3_SUS: 3.331V
PP3V3_S3: 3.331V
PP5V_S0: 5.129V
PP3V3_S0: 3.328V

PP1V8_S3: 1.835V
PP1V2_S3: 1.228V
PP1V05_SUS: 1.085V
PP1V5_S0: 1.535V
PP0V6_S0_DDRVTT: 0.630V
PP1V05_S0: 1.092V
PP1V05_S0SW_PCH_HSIO: 1.090V
PPVCC_S0_CPU: 1.8V

Dead CPU?

 

[Bouba148]

I also noticed that at U6100 pin 6 , SPI_MLB_CLK is at 0V
Then I went thru R6121 and R6111: SPI_CLK_R is also at 0V which is provided by the CPU......

 

[JohnB8812]

I don?t think so. If you had flames around your TBT stuff perhaps something is shorted/causing issues in that area and causing no post. All TBT stuff removed?

 

[2informaticos]

SPI I/O pins must be low when PCH accesses U6100, so that looks correct.
But you must check all SPI Bus Termination resistors and maybe remove J6100; sometime gets corroded and blocks the communication.
You can also try new BIOS, as sometime liquid damage cause BIOS corruption.
Maybe need to put good flux and heat CPU; no more than 180-200 degrees, just to burn eventually crap.
Ultrasonic cleaning in the end could help too.

 

[Bouba148]

@JohnB8812, "All TBT stuff removed?"

As I've removed Q3080, now I have nothing going to the THBT level, so I guess I should be OK no?
You agree that it's NOT normal that Q3080 was passing (closed ) with pin 1 and 4 at the same voltage (8.61V)

[USER][/USER] :
I've forgot to say that I succeed reading the SPI Rom with the Medusa, but only after several tries as it had difficulties to connect.
I've cleaned the ME region and PRAM area

When you write: "SPI I/O pins must be low when PCH accesses U6100"
How long does that sequence take, only few seconds?
Correct me if I'm wrong but I use to remember that all SPI pins had a level of 3.3V except the 4 of course.

Also I still have NO GREEN light on the charger......which is secondary

Best Regards

 

[2informaticos]

Write new BIOS; you could have corrupted pure BIOS area not only ME region.

SPI pins should go high on the first moment, but I/O pins will drop immediatelly and should stay low.

Test the MLB without LIO flex connected.
Once the board works, then you have time to fix Magsafe light issue.