820-00239 fans spin, most rails ok but no other sign of life

[Outofwhack]

Hi All,

I am new to board repair but have an education in electronics from college two decades ago. Having lost my job in aviation I'm hoping to make a go of Mac board repair in the UK.

Ive had some luck with repairing a couple of MacBooks using Louis’s fantastic videos but the A1706 appears quite a chore to re-assemble just to spot fans spin with ribbon cables hidden by ribbon cables with the fans being left in the case and all.

After the liquid spill its owner sent it to sleep and delivered to me within a day [he says].
Wont boot, no display but fans spin, no external display.
Burnt LVDS ribbon cable adapter to screen (burnt at screen control board too), tried some test screens but waiting on new ribbon cable adaptor before further testing.

So the fans spin on this one already and this has thrown me. I am lost and need help!

I have attached photos below of all corrosion I have seen on it in the hope for opinions on which must be removed, reflowed or replaced. I gradually purchasing spares of components replaced in Louis's videos but I have only gained one A1465 donor board so far.

From what Ive learned from watching Louis’s videos it appears to reach the S0 state…..
But there are a few rails missing (are these supposed to be missing).
Anyhow Ive embedded a few questions along the way (in brackets). Fanspin doesn't seem to be the reward for this board - what is?

(I will just add my multimeter UNI-T UT139C seems to take a long time to settle down when measuring ppbus_g3h and the like. If I hold the probes on it might well drop half a volt after 40 seconds - not sure whats going on there - my other multimeters behave the same)

ppbus_g3h 13.1v
pp3v3_sus 3.3v
pp3v3_s0 3.3v
ppvcc_s0 0.73v (shouldn’t this be 1.5v)
ppvccsa_s0 0.85v (why does pdf say range 0.55 to 1.5v)
ppvccgt_s0 0v
pp1v2_s0 1.2v
pp0v6_s0 0.6v
pp1v8_s0 1.8v
pp1v0_s0sw 1v
pp5v_s0 5.2v

ppvccedram_s0_cpu 0v (40ohm to earth, I expect this is ok????)
ppvccedram_reg_f 0v
pm_opc_zvm_l 0v
cpu_zvm_l 0v (signal comes from U0500 - what does this chip do?)

pp5v_s0_vcore1 5v
pp5v_s0_vcore1 5v
ppvcc_s0_cpu 0.5v (should this be 1.5v, page 68 of pdf indicates 0.55 - 1.5v Why?)
It seems to wander between 0.43 and 0.53 according to my multimeter.

pp5v_s0_fan_conn 5.2v Does this indicate fan wold be spinning if it was back in case?
smc_lid 3.4v Does this indicate lid(screen) open or closed?
lcd_bklt_en_l 13v Is this an ok level?
bklt_en_r 0v
edp_bklt_en 0v

all_sys_pwrgd 3.2v I left mentioning this till last because Im sure it was 0v when I first measured it. But now its there. Hmmmm

Any help appreciated.

Regards, Andy in the UK

 

[2informaticos]

First of all, welcome to the forum!

You should start eliminating any corrosion sign.
Board looks really ugly around U7800, U7960 and U8000.
Remove all QFN (or similar) chips and re-tin the pads with leaded solder.
U7800 is more delicate (BGA package); reflow it with quality flux first.
Change corroded caps and resistors.
Check all traces around U6100/01.

A lot of questions, not all relevant in this repair.
"lcd_bklt_en_l 13v Is this an ok level?
bklt_en_r 0v
edp_bklt_en 0v"
None good, but no reason to concern about them, as we don't even know if board boots yet.
You've found a damage on video connector.
Expexting image is the last test should do in such case.
Did you check for USB activity?

Some PCH and CPU voltages are adjusted in real time; that's the reason of schematic range, instead of fixed value.

"ppvcc_s0 0.73v (shouldn’t this be 1.5v)"
I'm not sure what this mean; PPVCC_S0_CPU maybe?
That could be good, but you've posted another value below.

As a rule for new thread, start posting 820-xxxxx board code.

 

[Outofwhack]

Many thanks for your fast response 2informaticos,

Ive updated the title of the thread with 820-00239
I will look up how to check USB activity in these thread but its late here 10pm.
I will remove/re-tin/reflow as suggested - thanks.

I am keen to understand if there is another way to know if a board boots than looking for fan spin as its so tedious to reconnect the board inside the case just to connect the fans.

Can I rely on every MacBook board with USB-C or older ones with hdmi/thunderbolt automatically outputting a video signal to an external monitor as a way of verifying boot when main screen not connected?

I did mean PPVCC_S0_CPU - you were correct. I made measurements over multiple days and it does seem to have changed. I am embarrassed how long I have had this board - I tried and tried over 10 days before deciding to join the forum.

Forgot to say, my USB volt/ammeter shows 20volts supplied but 0.000 amps drawn.

Thanks again

 

[2informaticos]

0.7V at CPU VCORE can be sign that board boots; should drop from initial over 1V.
But you can't expect internal video, with such damage on the connector.
Checking external video is a good idea, but a problem with iGPU may stop that too.

Checking USB activity is a MUST DO test; read how to do that properly:

HOW TO collection

I hope this "How to" collection will help our members. I will possibly modify this post in time, if updates available. Trigger SMC_ONOFF_L (PMU_ONOFF_L): It reffers to simulate power button function, when the keyboard is not connected. We need to do this when we work with the board...

boards.rossmanngroup.com

 

[Outofwhack]

No USB activity indicated by SNOW edition USB drive.

Aside: Customers decision to proceed with in-depth board repair could be dependant on if they also need a new screen after the burned screen connection.
Question: can backlight be tested with just a variable voltage on the [burned] connector pin ?

 

[2informaticos]

When LCD connector is burnt, very likely CPU eDP output is dead; because of leakage from backlight output.

Your board has many areas affected by corrosion.
BGA chips may also have corrosion underneath.
You should think a bit more, before invest more time.
May be too expensive and client wouldn't accept the price finally.

 

[Outofwhack]

Thanks. Yes I think for a beginner Ive got a pretty hopeless board.

I have talked customer around into purchasing a refurbed board (at his expense) but before going down this route he wants confidence in the rest of his computer working first. Its around $500 as it was 3.3GHz and 16GB/512GB
If I can keep him happy chances are I get to keep the broken board as a much needed donor.

So is there any way to get some confidence in the screen without using a working logic board.
e.g. should the backlight come on with just the right voltage on the supply line and an earth? or does the screen rely on other signals to light the backlight?

I wonder how does the screen supply business test returned screens? Im wondering if there is any cost effective solution to test them other than keeping a stock of working logic boards. Ive collected a bunch of broken screens from customer repairs but from examining pinouts theres big danger in assuming just because a connector fits together they have the same pinout.

 

[2informaticos]

Checking backlight presence on spare LCD, will not assure it has a good image.

You can use a lab PSU to apply voltage on the backlight power rail, but I do not recommend to inject directly in the board; unless you have a dead board to use for this purpose.
You need to find proper points on the small board attached on the LCD assembly.

You can start with 35V and go up slowly, until backlight comes on.
Older screens need 24-40V, depends by model; newer screens (retina) require 44-50V.
Do not exceed nominal voltage, or LEDs may get damaged.
Just in case, set current limit to 0.1A.

 

[piernov]

As far as I know there must be communication between LED driver on main logic board and backlight controller on LCD T-Con board for backlight to work on Retina displays.
Messing with 40V around an LCD connector is a good way to damage it too so I wouldn't recommend it.

 

[2informaticos]

LED driver requires such communication.
I explained him how to test bypasssing the driver.
If starts the test with low voltage and current has a correct limit, cannot burn the LEDs.
However, the test should be done with scrap board, or (better) injecting directly into pads of LCD T-Con board.