820-00165 no greenlight in S0 state

[speggio91]

Hello everyone,

I've started to practise more on macbooks air board, I've got a bunch of defective ones, to see if I can gain some more knowledge.

I've been focused on a water damaged board of an 2017 Air.
Someone else already worked on it, the SMC has been reflowed plus some minor reflows around the board.

I've reached a dead end but I improved the state of the board.
Initially it was no power, no green light, no fan spin, stuck at S5 state.

After replacing some components (U7090, Q5500, Q7430, C7432, Q7630, U1950, R6115, R6116, Q3080) and reflowing many others (U8300, U4590, U6100, U6101, U8050, U7701, U5800, U7400, U7600, U1900)

The current situation is:
the board seems to be powering on (the CPU gets hot) but I get no chime, no greenlight on the charger (genuine one), the battery won't charge but I get a fan spin.
I can't test if I get an image because the LCD panel backlight has been broken.
If I do the SMC bypass procedure, the fan start spinning fast and it seems that in the LCD there is something showing. I don't have any SSD to test if the OS is booting.

I made several measurements to check if the S0 state is working:

PP3V3_SUS 3.366V

PP3V3_S4SW_SNS 3.365V

PP3V3_S0SW_SSD 3.365V

PP3V3_S0 3.322V

PM_SLP_S3_BUF_L 3.366V

PP5V_S0 5.135V

PP0V6_S0_DDRVTT 0.6V

PPVTT_S3_DDR_BUF 0.6V

PM_PCH_PWROK 3.4V

PM_PCH_SYS_PWROK 3.4V

CPU_VCCST_PWRGD 1.016V

PLT_RESET_L 3.366V

PP1V05_S0 1.060V

PP1V5_S0 1.514V

PP1V05_SUS 1.060V

PPVCC_S0_CPU 1.808V

PP1V05_S0SW_PCH_HSIO 1.052V

PP1V8_S3 1.8V

PP5V_S3 5.1V

PP1V2_S3 1.2V


PPBUS_G3H 8.600V

PP3V42_G3H 3.4V

SYS_PWROK_R 3.3V

PM_S0_PGOOD 3.408V

P5VS4RS3_PGOOD 3.36V

DDRREG_PGOOD 3.36V

P1V8S3_PGOOD 3.36V

P1V05S0_PGOOD 3.36V

ALL_SYS_PWRGD 3.36V


I think I've reached the end of my capabilities in troubleshooting this board, I'm happy of the progress but I ask you if there is something else that could be missing to check.

In my opinion the cpu is dead, because there was some hairy junk coming from the left side of it, I've tried to spray some IPA underneath with some air can, but I think it didn't do much.
I was thinking to reflow the CPU but I think I would destroy it without any hot plate or pre heater.

Also in the RAM chip zone there is more of that junk, I was considering a total reflow of all the chips as well.

[mon2]

Re: 820-00165 no greenlight in S0 state

Apply a flashlight from the backside of the display panel to mimic the missing backlight.

Does the caps key toggle the led on the keyboard?

[speggio91]

Re: 820-00165 no greenlight in S0 state

I didn't realize that also the panel itself is broken, I can't tell you what's the screen showing.

When I press the caps lock key, the green led show up.

Could it be an ISL problem? Maybe also the SMC could be..

[mon2]

Re: 820-00165 no greenlight in S0 state

If the green caps light toggles, the OS is booted. Carefully remove the LCD panel connector using the latch used to secure into place and inspect the contacts of the LCD panel cable. Wipe with IPA (alcohol) if required).

Post pics of the area of the display connector.

[speggio91]

Re: 820-00165 no greenlight in S0 state

Maybe I haven't explained very well the screen situation but the connector itself is in good conditions. The problem is that the macbook air i've been working on has been damaged by a previous technician that broke the screen assembly trying to pry open the adhesive bezel. The panel and the backlight are broken, I should buy a new screen to test the LCD screen output.

I've cleaned the connector as suggested, I've attached some pictures.
I've also uploaded few pics about the SMC area, there are some burnt component that i'm not sure if they are crucial for the SMC to work correctly.
I've measured those components and the resistors value are correct. The L5001 is 0 ohm value.

[reformatt]

Re: 820-00165 no greenlight in S0 state

The test point connected to R5381 is likely open circuit, would explain why your battery is not charging (since this is a SMBUS line to the battery). Corroded test points very common issue on these boards.

If you've had that much corrosion right next to the SMC, then likely there is corrosion under it. But given it's booting you might be ok. I usually reball them even if a reflow is successful.

A damaged display is still usually detected to at least enable backlight. You should get 25V at C7797 once the board boots. I have a couple of damaged LCD's that I use for this purpose on the bench. Buying a replacement display makes these uneconomic for resale unfortunately. I can only get about AUD$400 for them, and the displays are about $190 aftermarket. Couple with new battery and maybe a SSD and adaptor, you don't make much.

[reformatt]

Re: 820-00165 no greenlight in S0 state

For displays, keep an eye out on eBay for faulty A1466 from 2013~2017. The 2013/14 models not much in demand as they have 820-3437 boards and only run Big Sur natively, but the rest of the chassis incl screen can be used on 2015/2017 models. Units with screens with no corner dents or keyboard markings on the LCD are the most valuable.

[speggio91]

Re: 820-00165 no greenlight in S0 state

Reformatt, thanks for spotting that corroded pad, I really didn't see it in the first place.
Later I will try to fix it and see if the battery charging circuit will work again.

Also thanks for the tips about the older macbook market model, I will take a look.
I also seen some aftermarket screen from aliexpress for 50€, that would be my last stand in case I won't find anything cheaper on ebay or similar marketplaces.
Later I'll test the backlight voltage as well.

About the SMC, I was thinking also to reball it just to be sure the reflow didn't work 100% but I was a bit hesitant because I'm not familiar with the process. I also don't have the BGA stencil. But I think I'll give it a try, I don't have anything to lose actually. Do you think reballing the chip with a solder blob will be enough or would make the balls too much unevenly the same size?
Alternatively I have the solder paste. In case I will get the stencil, does the paste need to be applied right away or need to be dried first? The reflowing temperature to melt the paste is roughly 250 C? What about the airflow? I know the settings differs a lot from station and location, I just wanted to have a general indication to start trying the reball process.

About the fact that the charger light won't light up, is it related to the SYS ONEWIRE even if the board boots or it is related to the SMC itself?

Also about the fact that the board boots only under SMC bypass, it can only sensor related or it might be that SMC that is not working correctly?

Sorry for the many questions

[reformatt]

Re: 820-00165 no greenlight in S0 state

About the fact that the charger light won't light up, is it related to the SYS ONEWIRE even if the board boots or it is related to the SMC itself?

Also about the fact that the board boots only under SMC bypass, it can only sensor related or it might be that SMC that is not working correctly?

The charger light is supposed to be out in SMC bypass. It goes from green/orange to out as a visual indication you are in bypass. PPBUS_G3H will also drop from 8.6V (with no battery) to 8.3V, and fan will also spin at max speed since sensors are ignored.

Check for a missing PP3V3_S5_AVREF_SMC, this is the most common reason for only working in bypass. Often due to a shorted cap C5125, or corrosion around U5110.

As far as reballing goes, this is a whole other topic. Deal with the other issues first.

[speggio91]

Re: 820-00165 no greenlight in S0 state

-I've fixed that corroded pad but there's no change to the green light/orange light in the charger behaviour, both are not yet present.
I've measured the battery voltage on the pins 1,2,3 of J6950 and there seems to be some activity there, the starting voltage with no charger connected is 1.5V. After connecting it it raise up to 4.5/5V and it seems to be rising.

-on C7797 I measured 40V when the display connector is connected. That voltage is present even if I'm not in the SMC bypass mode. Does that mean the board is powered on normally or that doesn't mean anything? Yesterday I replaced U7701 from a donor board because it seemed to be quite corroded.. maybe that did something to the backlight circuit even if I didn't test it until now.

-in the SMC area i've redone the pads of R5177, R5178, R5179, R5180, they were quite corroded and the lines were interrupted.
I've replaced U3710 and R5625 which weren't in a good condition at all.

Then I've desoldered Q5158 and it was cracked, the pin 2 was detached from the rest. In the process of replacing it with a donor one, it slipped away and I didn't find it again. I guess that component is important for the CPU to communicate the SMC its temperature. I will try to get more of those transistors.

-I've tested the line PP3V3_S5_AVREF_SMC and it measure 3.33V without entering in the SMC bypass mode.

Do you think the board is actually booting even if there's no light coming from the charger?
I guess the next step will be to take a decent display and see what's going on there..

[reformatt]

Re: 820-00165 no greenlight in S0 state

40V for backlight? I've only ever seen that once and I think I resolved it with a new LP8550. It should be 25V.

Do you have a USB stick with a LED that blinks when accessed? When an Apple board successfully POST's, it will probe the USB ports for something to boot off. The LED in a USB stick should blink 3 times to indicate the port is being accessed. I use a Chipmunk USB tester which is specifically designed for just this thing.

You can also use a magsafe 2 cable from a power supply and monitor input current. It's important to use a genuine magsafe cable, as 820-00165 boards are particularly fussy about them and will hold PM_BATLOW_L low if they are dodgy, and they won't reach PM_SLP_S4 as a result. 820-00165 boards will initially go 550 to 600ma, then settle back to around 330ma or so once it's POST'ed.

From your descriptions, I think you at least need to reflow the SMC or reball it. If the SMC is not in bypass, it will not charge the battery without onewire comms to the charger. You can also have a bad SMC as liquid damage can at times damage the SMBUS data connections to it.

[speggio91]

Good news!

I've purchased an early 2015 air in immaculate conditions that had no power issue. I've fixed the board and I swapped it with the one of this topic.

The machine is actually turning on, the screen image and backlight are working and the battery is charging.
The charger led is also working, turned out to be the flat cable 821-1722 that had a pin that desoldered after it got shorted when it got water damaged.

After booting, the OS is kinda slow due the missing transistor Q5158 that I will replace as soon as I will get some more of them.

The only issue I've noticed is in the backlight circuit. When I try to adjust it, the screen won't dim but will abruptly lower the brightness at the minimum value when the UI indicator in the OS is at half way.
Looking around in the schematics and in the wiki I think the part to check is the LED-Retun / Feedback / Current sink one, right?
I already tried to reset PRAM and SMC with no changes.

Reformatt, I think you previously mentioned what is managing the backlight intensity but the last few messages of last week has been lost after this forum has been updated.

I'll try tomorrow to take a look at the schematics and wiki to better understand how the circuit work.

Thank you again for the hints!

[reformatt]

EDP_BKLT_PWM is a square wave, and the duty cycle is varied to adjust brightness. Backlight voltage on a working system is around 25V or so. Current is adjusted via the LED_RETURN_X paths so voltages you measure aren't an overall indication. I'm sure if you put in a new LP8550 it'll sort your issue. I've not had much luck re-using these off other boards as the print on top tends to disappear after reballing and you can't determine what is pin 1. They are only about $3 each so I just buy them in packs of 10 off SUHMS on Aliexpress.

[piernov]

I've already received bad LP8550 from unofficial distributor in the past. For once, this part is available at Mouser so I'd highly recommend buying there instead.
Note that when they are new they need to be configured with the proper slope setting.

[speggio91]

I've finally received some LP8550 alongside with some MMBT3904LP-7B DFN-3 (Q5158). After replacing both components the backlight is functioning properly, it only need to have its slope reprogrammed as suggested by piernov.
When the system boots, the fan is still spinning at full speed and the system is a little bit slow. I've downloaded macs fan control, but all the sensor reads good temperatures and the cpu doesn't seem to be thermal throttling.
I've tried to do the PRAM and SMC reset and they both gave no improvements.
If I startup the Macbook in SMC bypass mode, the system become more reactive and snappier.

Where should I focus the searches to find the sensor failure? Should I inspect all the sensor circuits at pages 43-44 of the schematics?

[mon2]

Review your trackpad cable. The sensors are often in the trackpad assembly. Do check your smbus lines to see if they are stuck which will cause the sensor feedback to fail (i2c based). Measure the voltage to ground of each smbus scl and also the sda lines.

[speggio91]

Thank you for the hints Mon2.

The trackpad i'm using belong to a known good case, I'ìve tested it with another functioning board with no issues on the sensors.

I've took some measurements for all the SMBUS scl and sda lines:

SMBUS_SMC_0_S0_SCL = 2.9V (voltage to ground) and 0.58V (diode mode). With the display connector detached the diode mode value is: 0.77V and the voltage to ground is 3.3V

SMBUS_SMC_0_S0_SDA = 2.9V (voltage to ground) and 0.59V (diode mode). With the display connector detached the diode mode value is: 0.77V and the voltage to ground is 3.3V

SMBUS_SMC_1_S0_SCL = 3.3V (voltage to ground) and 0.59V (diode mode).

SMBUS_SMC_1_S0_SDA =3.3V (voltage to ground) and 0.54V (diode mode).

SMBUS_SMC_2_S3_SCL = 3.3V (voltage to ground) 0.77V (diode mode).

SMBUS_SMC_2_S3_SDA = 3.3V (voltage to ground) 0.77V (diode mode).

SMBUS_SMC_3_SCL = 3.3V (voltage to ground) and 0.66V (diode mode).

SMBUS_SMC_3_SDA = 3.3V (voltage to ground) and 0.66V (diode mode).

SMBUS_SMC_5_G3_SCL= 3.3V (voltage to ground) and 0.55V (diode mode).

SMBUS_SMC_5_G3_SDA = 3.3V (voltage to ground) and 0.55V (diode mode).


Anything else to check?

[mon2]

Each of these readings look good. SMBUS (I2C) lines are not stuck. You will need to run ASD (I think that is the name) tool to perform an exhaustive diagnostic of the logic board. That tool will then report which sensor is at fault.

Also review this thread:

[Solved]: A1466 (820-00165-A) sensor issue | Rossmanngroup Forums

[speggio91]

I've downloaded and ran ASD EFI 3S156 from an USB drive, I did a full test cycle and an error has been found. I've attached a picture about it. No sensor issues was found though.

I might try to remove and replace U5810, as seen in the Rossmann forum topic, even if I'm a bit skeptical that it would make any difference.

[mon2]

Not at my computer so you will have to hunt for the lcd backlight controller. Measure the voltage to ground of the smbus scl and sda lines that mate with this controller. Post each measurement.

if not convenient, remove all power. Meter in diode mode. With no power, safer and easier to measure test points without shorting the board.

red meter probe, yes red to ground.
black meter probe to backlight controller smbus scl line. What is the diode mode reading?

repeat for sda line. Post each measurement.