edm 010
Boot sequence 510 for 10 seconds and then sleeps at 0.009
During boot sequence: DA9081 hot and SSD controller very hot
Short:
Capacitors pins 3/39/40 1.2V_FC_VDDQFIO
Capacitors pins 21/20/25/24 1.2V_FC_VDDQMIO
Some capacitors behind the SSD controller
Uart:
80891001
80894002
80892003
Is it safe to inject 1V - 1A to check for any shorted capacitors?
@Calvin my friend 
You can give it a try with injecting voltage. My money is on the SSD controller.
Two bad caps on two different rails would be unusual.
100 ohms predicted - 63.5 detected
650 ohms predicted - 0.07 detected
0.9k predicted - 1.2k detected
3-4 megs predicted - 389k detected
1.3-1.7k predicted - 0.01 detected
I agree with you, they’re two different lines
To be safe, I injected 0.8V - 1A and the SSD controller heats up immediately.
I’ll try to remove it and see if the short goes away
Replaced the SSD controller. Initially no boot sequence, then I did a light reflow and everything was fine.
I tried to turn it on and blod for 2 seconds.
80810001 - 217F uart
No different errors.
DA9081 the voltages are correct and I have no shorts. 0.8V that goes to SSD controller shows 290 ohm resistance to ground
Any suggestions?
I would reball the SSD controller again and see if it solve the issue.
I’ve replaced the SSD controller twice, but unfortunately the issue remains the same. After inspecting it carefully under the microscope, I noticed that the BGA balls around the edges solder properly, while the balls in the center remain slightly lifted because the PCB is slightly warped.
What do you recommend?
Bend PCBs are a pain. I still have a good mainboard but it is also slightly warped and if heated up for resoldering the APU, it warps so that the edges of the APU are touching the board and squeeze the BGA balls partial together.
Maybe lower but longer heat helps preventing warping?
It’s the first time this has happened to me, but yes, I think it was caused by too much heat from the top at a high temperature without any preheating.
Can you recommend a preheater? I need to replace another SSD controller, but honestly I’m a bit scared now 
To solve this, I’m considering using 0.50 mm solder balls to improve the contact. What do you think? Also, would you use flux under the chip or not? Would you recommend heating it from the bottom instead of the top?
Good question. At southbridge and SSD controller I usually use 450°C heat and 30% airflow without a nozzle to de- and resolder without extra bottom heat (It is a Weller Hotair station and I don’t know what is the airflow in compare to other hotair stations).
I move the hotair all the time and probe with tweezers if the chip starts to move and if, pick it up.
At resoldering, I apply a thin layer of flux and place a new or reballed chip on the pads and start heating also while moving the nozzle. When the chip sit down it moves a bit and than I check again with tweezers by very slightly tapping the side and see if the chip moves and fall back in position.
My stencil I use for reballing Southbridge and SSD controller has 0.4 mm. But I guess 0.5 mm shouldn’t be an issue. You can place one solderball on a pad and check if it looks ok from the diameter.
I used stencil with 0.5 mm solder balls. Boot stuck at 0.55 V. After inspecting the SSD controller under the microscope, I noticed that some of the balls along the edges hadn’t made proper contact.
I performed a reflow, but nothing changed. Then I measured the 0.8 V rail from the DA9081 and found an abnormally high resistance. After another reflow, the console reached a partial boot (0.221 A), and the resistance returned to normal. Unfortunately, under the microscope I found two solder bridges on one corner of the SSD controller. Damn!
At this point, I believe the PCB is very slightly warped. I think using 0.50 mm balls helps compensate for it. Now I just need to figure out the correct reflow profile. I’m using an ATTEN ST-862D at 425°C and 50% airflow, but I suspect the airflow is too high. Since I’m not using a nozzle, it may be pushing the chip down too aggressively.
Maybe in this case it would be helpfull to warm up the board to 150°C before resoldering the chip, to prevent the board from warping and keep the needed time and heat for resoldering short and lower.
I bought a T-8280 and it was time! As soon as it arrives, I’ll finish this repair.
