820-00244-a short on ppbus_g3h

[gadgetfix]

hey guys. this one came from another store and there has been a previous repair attempt. I only noticed because L7210 was missing I have no other history and suspect no prior liquid damage.

ppdcin_g3h has 20v present but both ppvbat_g3h and ppbus_g3h on F7040 both showing 0.6v. I removed F7040 and measured both to ground and short is on ppbus_g3h. I soldered a wire to ppbus_g3h and put around 8v through - it was drawing close to 1.8A but no heat? Should i go higher?

Thanks

 

[2informaticos]

Where did you hear to inject in such crazy manner?
Always search first if no high-side MOSFET is short on any secondary power supply.
Because if not, you can directly inject high voltage into a low voltage rail!
Also NEVER start with hihg voltage, even if you are in the main power rail.
Post the exact value for that short.
Then search all the (big) coils for similar value.
This way you can detect a shorted high-side MOSFET, if exists.

 

[gadgetfix]

Hey informaticos i started with lower voltage with no success so i went up until it got to 8. i will test tomorrow mornign and report back

 

[gadgetfix]

.7ohm pm ppbus_g3h

 

[2informaticos]

Now check if other big coil has the same value.
If any found check then between it and PPBUS_G3H.
If you get 0.000-0.001 means shorted MOSFET. If you decide to inject voltage into PPBUS_G3H, do not exceed 120% of that coil.
If you get 0.014-0.015 means no short. In such case you can inject even 8V on PPBUS_3H.
Hoiwever, always start with low voltage...

 

[gadgetfix]

ahhhhh i see what you are saying. thanks informaticos

 

[2informaticos]

In case of shorted MOSFET, remove it and then you don't have a low limit for voltage injection anymore...

 

[gadgetfix]

ok. short is not on ppbus_g3h its on PPBUS_S5_HS_COMPUTING_ISNS

 

[JohnB8812]

Those are literally the exact same rail...try this remove the CPU heatsink and use freeze spray if possible. Inject voltage on pin 2 of F7040 at 1.8 volts and see if by chance the CPU die evaporates first. If so, dead CPU. If not, then something else must absolutely get hot.

 

[gadgetfix]

John must i use freeze spray?

 

[gadgetfix]

ok tried thermal cam and freeze spray. done the entire board both sides no go.

i have previously removed some components that i have yet to put back on in pursuit of finding the short.

PPBUS_S5_HS_COMPUTING_ISNS is 3.4ohm short to ground.

 

[JohnB8812]

You don't have to use freeze spray, but it helps. Something I'm seeing more and more on these new ones is a dead short to ground through the PCH die of the CPU. Inspect that. Also inspect if any cap gets obviously hot

 

[gadgetfix]

hmmm no caps getting hot. anything else you can recommend?

 

[JohnB8812]

Like I said, I wouldn't be surprised if it's the die on the CPU.

 

[gadgetfix]

anyway to test?>

 

[JohnB8812]

Freeze spray is best with canned air upside down.

 

[2informaticos]

"PPBUS_S5_HS_COMPUTING_ISNS is 3.4ohm short to ground."
Did you remove R5450 to confirm that?
Check now the exact value at L7210/30/90, L7410/30, L7900/60, L8000/50.
This way you can discard a shorted MOSFET.
If is not the case, you can inject voltage without fear.
That rail has many filter caps, and also goes to U7800.

 

[gadgetfix]

"PPBUS_S5_HS_COMPUTING_ISNS is 3.4ohm short to ground."
Did you remove R5450 to confirm that?

yes

I have also removed u7800 but i will check mosfets.

 

[gadgetfix]

L7210 1.1ohm
l7230 1.4ohm
l7290 25.4ohm

l7410 10.8ohm
l7430 10.8ohm

l7900 446ohm
l7960 112.5ohm

l8000 185ohm
l8050 152ohm

 

[2informaticos]

Now post resistance value between L7210/30 and PPBUS_S5_HS_COMPUTING_ISNS.
Lower than 3 ohm (aprox 2 ohm) means one of corresponding MOSFETs is shorted.
In such case, 95% dead CPU.
If you get aprox 4.5-5 ohm you are lucky, CPU can still be good.
Then should be one cap which must heat injecting voltage...