(347) 552-2258 [email protected]

Macbook Air 820-00164 Need U8080

M

MexicoTV

New member
Hello,

New to the forum, so glad to be part of it! I have much experience with electronics but not as much with MBs though I have experience there too. This is my first time working with a shop that will hire me if I can prove I can fix boards for them.

I have gotten a macbook air with no power but green and orange light. No power but does charge. The board makes it to S0 state. PPV3V3_S0 is present but PPV5V_S0 is not. I found that the enable signal to U8080 was not present. I then found that resistor R8187 which is more a jumper was open. This jumper takes the P5VS0_EN signal from U8180 and sends it over to U8080. U8180 does have all the right voltages including the outgoing engage signal for U8080. I jumpered R8187 and found that engage signal still does not make it and it even goes missing at U8180 now too. If I measure the incoming voltages to U8180, the VCC voltage is still there but PM_SLP_S3_L begins to "blink" as if is being overwhelmed and protecting itself. U8080 does also have it's VDD. It's as if U8080 engage pin is shorted to groun and when I reconnect U8180 it messes it up too. Also after this test it seemed like U8080 was hot. U8080 doesn't measure low to ground on output, VDD, or engage pin. I believe the IC might just be dead. Is my diagnosis correct you think?

I see the IC is for sale from china but would take a long time to get here.

Does anyone have this IC on hand who can sell it to me? Even if it's from a dead board. SLG5AP1443V is the part number. Cannot find datasheet.

Thanks!
 
Last edited:
2informaticos

2informaticos

Administrator
Staff member
You have a problem on PP5V_S0 power rail.
This is the reason U8080 gets hot and PM_SLP_S3_L also pulses.
Just for test, U8080 can be shorted I/O (3-5).
If you don't have short/low value to ground on PP5V_S0 is not easy to find the problem; post the value however.
I recommend to remove U8080 first and try bypassing it (as above mentioned); just to discard it.
 
M

MexicoTV

New member
Looks like youre right, i get the same pulsing after hooking those two lines up. This time im seeing a resistance between PP5V_S0 and GND. 63K Ohm when black is on ground, 32Kohm when red is on ground. This is still high for a 5V line. It would be great if someone with a working board could give me their reading. I did notice U2800 getting kinda warm when this thing was plugged in. Not very warm but noticeably warm, kinda the only warm thing on the board. Besides this I have almost no other clues, everything else is squeaky clean. No oxidation.

I guess I can start by looking at all the sub rails that depend on PP5V_S0
 
Last edited:
M

MexicoTV

New member
So how much of a bad idea would it be if I just forced 5V into this missing rail and watch for smoke or heat? I plan on hooking a usb charger up to it and limiting the current with an incandescent bulb.
 
2informaticos

2informaticos

Administrator
Staff member
As you don't have a short, nor even low value, probably nothing heats.
Unless you are lucky and the culprit enters in some kind of avalanche process.

Do you get a pulse on all big coils?
Did you try to start in SMC bypass mode?
 
M

MexicoTV

New member
What I'm wondering is if maybe one of the buck or boost circuits that feeds from PP5V_S0 is the one with a short on it's output. I snooped around a but so far nothing has happened. My theory was that if a dependent buck converter was drawing excessive current, or perhaps any other circuit where IT'S output has a short, it would cause PP5V_S0 to draw excessive current and it be commanded to stay off. If the short was on the output of a dependent circuit this would also mask the short when measuring PP5V_S0. Does that sound plausible?

I am not aware as to how to perform SMC bypass. Can you please explain? A google search brings up mostly SMC reset results, which i did.

EDIT: As for big coils, which ones do you mean? The ones that give the CPU power?
 
Last edited:
M

MexicoTV

New member
Okay i measured it. L3095 and L7701 are the only two coils that don't tick. All other coils peak at around .52 or so then go off. L7520 peaks at around 1.8v then disables. L7560 is makiing 3.3V. L7130 is making 8.5V. Neither of the coils that do not pulse show low resistance to ground. LL7701 is backlight boost and L3095 is thunderbolt.

U2800 is thunder host which is the part I notice is warm when I leave this thing plugged in. Is this normal? Possible lead? The thing is neither of the respective chips for L3095 or L7701 are receiving their enable signals.

Q7706 gets voltage in but none out. This is what sends voltage to L7701. U7701 enable signal too.
Q7706 gets 8.1 voltage at gate. This turns it off. Q7707 has to pull down Q7706 gate but this doesn't occurr.
Q7707 needs signals EDP_BKLT_EN and BKLT_PLT_RST_L which do not arrive. Both have origins in the PCH though the second one goes through a logic gate first, which is getting one of it's signals. It's an AND gate. PCH is not telling this to turn on.

Q3080 gets voltage in but none out. This is what sends voltage to U3090 vin and L3095.
Q3080 postive 8.5 at gate which turns it off.
Q3080 has to be turned on by Q3005.
Q3005 has no positive voltage on TBT_A_HV_EN. Originating from U2800

This is what I have found so far.

EDIT: I put a 390 ohm between PP5V_S4RS3 and PP5V_S0 pins instead of a bridge (where U8080 was) and even at a theoretical limit of 20mA the 5V PS shuts down. This time no pulsing.
 
Last edited:
M

MexicoTV

New member
My latest finding

Figuring that maybe the short isnt on the 5v line and that perhaps the activation of that 5V line is turning on an IC or transistor that shorts out a different unrelated line I cut the incoming 16.5v and put an ammeter there. The board idles at 20mA and when it peaks it hits close to 100mA then goes back down and resettles at 20mA.

I also measured current at PP5V_S0 again and only about 10mA tries to go in there.

EDIT:

So I made a mistake, the computer is NOT making it into SO state. PP3V3_0 is also doing the same thing. I confused it with PP3V3_S5.
 
Last edited:
M

MexicoTV

New member
I got fan spin!!!! I removed the enable signal to U8030. It's going to be a bitch figuring out why that helped though. No ground on this rail. This creates PP3V3_S0

No liquid spill. Not sure what the history is. Looks spotless.
 
Last edited:
2informaticos

2informaticos

Administrator
Staff member
Do you got all voltages now?
Looks like a problem on 3V3_S0.
Check connections between R/U5440.
Does it work in SMC bypass mode (with U8030 enabled)?
 
M

MexicoTV

New member
Im getting all coil voltages except L7701, L3095, L2920, L7310, and L7320.

I don't know how to do SMC bypass. Please explain?

These last two produce PPVCC_S0_CPU 1 and 2. These two i mentioned before as reading 13 ohms to ground. I took out one of the current sense resistors and on CPU side it measures about 8 ohms. Coil side reads normal. CPUs tend to read low but is that too much?

SO to summarize, for some reason PP3V3_S0 being on, makes PP5V_S4RS3 trip out and shut down, at the same time the enable signal shuts down for U8080 and U8030 and PP3V3_S0, PP5V_S0, and PP5V_S4RS3 shut down while PP3V3_S5 keeps going fine.

If I disable PP3V3_S0 then PP5V_S4S3 and PP5V_S0 both stay on as does PP3V3_S5 and a whole lot of rails come to life. Likewise if I keep PP5V_S4S3 disconnected from PP5V_S0 it and PP3V3_S0 both exist happily side by side.

Im gonna guess that whatever causes the issue is on the 5V line, not sure on what step though. 3V S0 is tripping something on the 5v line to over consume. We know that it isnt 5V S0.. but perhaps S4, S3 etc.
 
Last edited:
M

MexicoTV

New member
My updates for today.

I figured out what i think is the SMC bypass routine. Put everything back together and connected the magsafe charger while holding down the power button for a while. Nothing happened.

Tried SMC reset too, with and without charger connected. With the charger i can see the amber light go back to green for a moment.

Went around analyzing PP5V_S4RS3. This voltage dips and ticks if PP3V3_S0 is around at the same time. It is created by Q7520 and U7501 both of which feed from PPBUS_GH3. This rails flickers a bit when the fault occurs but only goes down by about 0.2 or 0.3 volts. So the fault is definitely on PP5V_S4RS3 or its associated components. This same IC creates PP3V_S5 through Q7560, this rail is working fine.

I went looking for things that could cause a new load on our PP5V_S4RS3 when PP3V3_S0 is activated and found:

U7600 untested other than for shots.
U8080 which we know works fine as long as PP3V3_S0 isn't on so no short or issue here as far as i can tell. All of PP5V_S0 is clear as it works while PP3V3_S0 is not available.
L4011, not on this board.
U4600 USB port power. I looked for shorts, none and also to make sure there was no pin throwing a fault. This pin is high 3,3V which means its okay as its active low. If im not mistaken.
U6410 speaker IC. Removed a jumper to isolate it, did not help.

There are some more things but this is what i looked into.

The only other main place I can see is the DC in/Audio board I actually did find a bit of corrosion here, it's light but there. Hiding under the ribbon cable that connects this board logic to the main. I found that the codec IC has several resistors that read ground on both sides. These resistors were removed and the ground was only on one side. However all these resistors are all only 0 ohm resistors so perhaps this is how they are. Removing them did not fix the issue but perhaps there is a short inside the IC? Why there would be these external connections to ground for the IC?

If anyone had me that could confirm that would be great, it's my only real clue right now. The board is APN 923-0430 aka 820-3453-A. I see a thread here with some also describing problems with PP5V_S4RS3. Would hate to buy this board only to find it's not the issue. https://boards.rossmanngroup.com/for...r-on-ppbus-g3h

Audio Codec is a CIRRUS 4208-CRZ. Found chip info but not much help.

NVM about the Codec, those were inductors to ground.
 
Last edited:
2informaticos

2informaticos

Administrator
Staff member
Disconnect LIO flex, keep only 5 pins DC cable connected.
Board should start automatically when charger is connected.

You don't have a problem directly on 3V3/5V_S0 lines.
I suppose some chip is enabled when 3V3_S0 appears and then generates overcurrent on 5V_S0, or 3V3_S0 rail...
 
M

MexicoTV

New member
Yes I did.

I have managed to inch closer to the fault. When I ground the enable pin to U8005 1.05V PCH HSIO Switch all the rails turn on, including some of the ones that didnt turn on before like L7310 and L7320 which are making 1.05V volts.

The offending line seems to be PP1V05_S0SW_PCH_HSIO. Concerning me is the fact that some of theses connections go to the PCH. Many of them in fact. Any clues? Onward....

There is no shorts on it's output or input. Need to see if this is enable on high or enable on low.

EDIT: Okay so turning it OFF brings everything else back online.
 
Last edited:
You must log in or register to reply here.
Top