820-00850 5v no power, PPBUS_G3H shorted, no PP3v3_G3H

Narek

Member
I'm working on 820-00850 board, it has no signs of liquid or physical damage. Customer states computer just turned off by itself. Some info about board:

- Stuck at 5v 0,45A
- LDO's on CD3215 are not shorted
- PPBUS_G3H is 0V, shorted to groud on system side (diode mode 0,000V)
- PPBUS_HS_CPU 0V, shorted to ground (diode mode 0,000V)
- PPBUS_G3H and PPBUS_HS_CPU are shorted to each other (diode mode 0,000V, 0,000ohm resistance)
- PP3V3_G3H is 0V, not shorted (diode mode 0,384V)

Is there any sense do dig into this board? I guess CPU got 12V from PPBUS and got fried, right?
 

2informaticos

Administrator
Staff member
Very likely CPU is fried.
You can lift CPU coils one by one and check if the short from PPBUS_G3H disappear.
Then check for short (ohm mode) on CPU rails.
 

Narek

Member
I have removed all CPU coils and short on PPBUS_G3H is not gone.

These are readings on CPU rails (CPU side):

PPV095_S0_CPUVCCIO_REG_R - 43,0 ohm
PPVCCSA_CPU_R - 9,9 ohm
PPVCC_CPU_PH1- 2,9 ohm
PPVCC_CPU_PH2 - 2,9 ohm
PPVCCGT_CPU_PH1 - 2,5 ohm
PPVCCGT_CPU_PH2 - 2,5 ohm
PPVCCGT_CPU_PH3 - 2,5 ohm
PPVCCEDRAM_S0_REG_R - 24 ohm
PPVCC_S0_CPU - 3,0 ohm
 

2informaticos

Administrator
Staff member
Then may be just a shorted cap.
Check first no short on other big coils, from the other secondary power supplies.
 

Narek

Member
Other coils are not shorted. Caps on PPBUS_G3H and PPBUS_HS_CPU looks fine, no damages. The only way to find damaged one is to inject voltage and use thermal cam. Is it safe for CPU now to inject voltage to PPBUS_G3H with all CPU coils removed? Or I should separate PPBUS_G3H and PPBUS_HS_CPU somehow (if it's possible..)?
 

2informaticos

Administrator
Staff member
You've said not same value to ground on CPU coils, as in PPBUS_G3H.
If you are sure about that, the CPU power supplies are good.
 

Narek

Member
I'm sure about these readings. I have injected safe 2V 2 Amps and foud out, that C6455 on PPBUS_G3H_SPKRAMP_LEFT line is glowing on thermal cam like crazy. After removing it PPBUS_G3H and PPBUS_HS_CPU are measuring 0,384V in diode mode, so I think case is solved.
 
Last edited:
Top