Apple 820-3662 VCore shorted to ground.

[barry101]

Hi, Would appreciate some help.
I have an Apple Late 2013 15" A1398 MacBook Pro with Logic Board 820-3662.
Get charger light but no fan spin or chime. No liquid damage.
Found PPBUS and VCORE both shorted to ground. Other rails were ok.
Removed VCORE MOSFETS U7310, U7320 and U7330. Found U7330 was in fact shorted.
With the 3 MOSFETS removed, get PPBUS and fan spin. VCORE is still hard shorted to ground (0 ohms).
Attempted voltage insertion on VCORE to locate short using bench supply set at 1.0 VDC and 4.95 Amps (max current for my supply).
Short reduced supply output voltage to 0.75 VDC and measured only 0.25 VDC at board. Only about 1 Watt of power
None of the VCORE capacitors heated up more than any others. None look suspicious.
Considered raising voltage setting on supply to increase voltage at board but concerned about frying the CPU.

Questions:
1-Am I wasting my time with this board as the CPU is most likely fried from the U7330 short?
2-Are the other 2 non-shorted MOSFETs still considered useable for use on other boards if needed?

Thanks.

[m1ch43lzm]

1. It's very likely you have a dead CPU if the shorted MOSFET sent 12V to it, if it's taking almost 5A on VCORE, the voltage drop is likely due to the cable you're using to inject voltage from the bench PSU, best to use a thicker gauge and the voltage loss across the wire will be lower (reflected on the PSU voltage meter, it will drop further, you have almost 0 ohms on the other end)
Remove heatsink and check with thermal camera/a drop of isopropyl alcohol on the CPU

2. Better get new ones as replacements, IIRC those MOSFETS on the 3662/3787 can cause some issues with random crashes/black screens
Source: https://logi.wiki/index.php/A1398_Common_issues

Late 2013-Mid 2014 (820-3662, 820-3787)


Screen off with computer on, or crashing: Sounds like the FDMF6808 chips are failing and causing black screening. You can try fixing this by replacing all 3 of them with NEW FDMF6708 or FDMF3030 chips from a reputable source like Mouser or Digikey. Possible software fix linked here

If you have dedicated graphics on your board then the above fix has a pretty low success rate.

If you have internal graphics, aka no separate graphics chip on your board then the above fix success rate is very good.

(source)


Additional notes : Symptoms are very unlikely to be a screen or screen cable issue. If you’re not getting image when this happens, then it’s not related to backlight either. Very very small chance it’s a GPU/CPU issue unless the fail is major. The FDMF chips issue arises from the board flexing. The chips tend to fail and cause the device to not wake from sleep/randomly shut down due to CPU VCore dropping out. Machine will stay on but basically never wake from sleep at times. If you’ve never messed with microsoldering before, I would recommend sending it to someone who has experience fixing this issue. If you have done board repair, then this repair is a little tedious so you want to make sure you get GOOD chips as the ones claimed to be new in China often end up being sanded down garbage to look new.

[barry101]

Thanks for getting back to me. Greatly appreciate it!
Thanks for your recommendation that the CPU is probably dead due to that short.
Regarding the 5 amp draw, I set the current limit on the supply to 5 Amps so I believe it's operating in constant current mode. I am using a pretty heavy gauge wire so I am surprised it’s loosing so much voltage. The wire is the size of the MOSFET output pad. Probably can’t use a wire much larger and get it attached.
I have the heatsink off and was using an infrared thermometer (unfortunately I don’t have a thermal camera) to check temperature on all the caps and the CPU. Highest reading was about 91. The CPU was a one or two degrees warmer than any of the caps which were mostly around the same temperature as the board. None were hot to the touch. Left the power on for about 15 minutes with no appreciable temperature rise. Didn’t think the alcohol would help under the circumstances but I can try it.
Will take your advice and not re-use the MOSFETs.
The board will probably be heading to the donor pile. It’s s shame, I have a few of these boards with the same PPBUS/VCORE short to ground.
Thanks again for your help!

[m1ch43lzm]

Usually the high side MOSFET on those FDMF gets shorted, therefore sending PPBUS_G3H straight to the CPU...
iR thermometer is useless for this, as it takes an average around a big spot, the trick with alcohol is to add a drop or 2, then connect the wire to the power supply for voltage injection
If the alcohol evaporates quicker (compared to room temp) in one spot when you apply power, it's dead
Usually one corner of the CPU gets hot when it gets shorted like that

[barry101]

Didn't realize that. I'll give it a try. Thanks!

[barry101]

Alcohol approach revealed definite hot spot in front left quadrant of CPU around pins in the 250 region. All the capacitors looked fine. Board is going into donor pile. Thanks for the tip!