820-3208 - no fan spin

jpadie

New member
Hi

i have an 820-3208 in an S0 state. but no fan spin.

The missing S0 rails are
PP0V75_S0_DDRVTT
PPVCCSA_S0_CPU
PPVCORE_S0_CPU
PPVCORE_S0_AXG


incidentally PP1V5_S3_CPU_VCCD is at 1.35v and not 1.5v

all_sys_pwrgd is not present (50mv)

I have changed Q7950 twice just in case

inputs on Q7950 are

VMON_Q2_BASE 1.06v
VMON_Q3_BASE 1.3v
VMON_Q4_BASE 0.7v
S0PGD 0.65v

the voltages coming in to each of the resistors (and dividers) feeding Q7950 are fine. the resistors themselves are fine. and the all_sys_pwrgd resistors also seem fine

any thoughts on where to go next with this board?

thanks

justin
 

2informaticos

Administrator
Staff member
If you have doubts with Q7950, remove it.
The board should work without it.
If ALL_SYS_PWRGD stays low, VCORE will not appear.
So check wich power supply involved in generating ALL_SYS_PWRGD fails.
Check first if R7967 gets 3V at pin1.
Next, remove one by one R7901 and R7963-66.
When ALL_SYS_PWRGD goes high, you just found the culprit...
 

jpadie

New member
thank you

removing R7901, R7903-6 sends pwrgd high. measuring on the incoming side of each removed resistor pad gives the following:

p5vs3_pgood 0.5v
p3v3s5_pgood 0.5v
pvccsa_pgood 0v
cpuvccios0_pggod 0.6v
p1v8s0_pggod 0.6v
CPUIMVPAXG_PGOOD 0v

so i guess a problem on one of the 0v lines.... any known issues to look for on those?
 

2informaticos

Administrator
Staff member
"R7903-6"
???

You must check ALL_SYS_PWRGD after each resistor removed.
You cannot get 3V on the other side of the resistors, as the pull-up R7967 remains on ALL_SYS_PWRGD side...
 

jpadie

New member
for example leave R7967 in place; remove the other resistors 7963-8 and test for voltage on pad 1.

unfortunately now a retrograde step in any event. i had identified that CPUVCCIOS0_EN was not present. went to make a cup of tea and have come back to find that the board is no longer in an s0 state. so I have lost most of the rails. I still have the hot rails but PPVRTC_G3H is only 2.7v.

no orange light on the charger.

damn.
 

jpadie

New member
i have all the always on rails.just RTC is at 2.7v rather than 3v. i will replace the clock chip\
 

jpadie

New member
that would be annoying. but I can't see why the PCH would spontaneously have fried.

I have swapped the clock chip. Output on RTC is still 2.7v . Also pp3v3_s5 is at 2.7v as is the P3V3S5_DRVH line coming from U7201

no enable on U7941. dead SMC?
 

2informaticos

Administrator
Staff member
"I can't see why the PCH would spontaneously have fried
---------------------------------
i had identified that CPUVCCIOS0_EN was not present"
You already have something PCH related...

Check resistance to ground on VCCRTC and 3V3_S5 lines; be aware, both go to PCH.
Compare with a good board, if possible.
Also check again all RTC signals.
 

jpadie

New member
I don't think it can be PCH as that enable signal is generated from pm_slp_s3_r which is also required for other enable lines that I know are functional.

i guess either r or c7981 must have been damaged.

But of course now the problem is rather wider with no full S5 rails and only 2.7v on pp3v3. I will check the resistances tomorrow morning before work.
 

2informaticos

Administrator
Staff member
I don't want to dissapoint you, but PM_SLP_S3_L is controlled by PCH; it also has the pull-up resistor inside.
All the sleep signals are controlled by PCH.
 

jpadie

New member
Sure. But if the signal wasn't being generated then all the other rails that rely on it would not be present. And previously the board was in an S0 state meaning that pm_slp_s3_l was not in an active low state .
 

2informaticos

Administrator
Staff member
You're right, it "WAS"...
So you can't discard PCH now.
C1802/03 are prone to fail on this board.
But them cannot drop so much VRTC rail.

If U2700 receives correct 3V42, and all caps from VRTC are good (disconnect them for test), then no other reason than bad PCH remains...
 
Last edited:

jpadie

New member
resistances on the requested lines are
PPVRTC- 4MOhm
PP3v3_S5 - 3.8kOhm

the always-on rails are present (save that PPVRTC is at 2.7v)

SMC_RESET_L,SMS_PWRBTN_L are both high.
SMC_PM_G2_EN is at at a few mV.

dead SMC?
 

2informaticos

Administrator
Staff member
"If U2700 receives correct 3V42, and all caps from VRTC are good (disconnect them for test), then no other reason than bad PCH remains..."
Did you check???

PCH can have high resistance to ground, but it can enter in some kind of avalanche effect, droping VRTC...
No matters SMC is good or not.
Bad RTC block will not allow PCH starting power on sequence!
 

jpadie

New member
i used an RTC from a donor board. but since VRTC is the same as pp3v3_s5 I suspect the issue is not the clock chip but pp3v3_s5.

i will double check the caps.

the more interesting question, forensically, is that the board had fan spin when the resistor to all_sys_pwrgd was remove to prevent it pulling low. The heat sink was in place and, as said, fan spinning happily and normally. It then spontaneously stopped within the next twenty minutes. No apparent overheating (although I was not present). No evident corrosion.
 

2informaticos

Administrator
Staff member
PCH should work with 2.7V on VRTC.
The suspect thing is if that voltage arrives to PCH as is, or is pulled down by PCH...

Do you also have 2.7V at 3V3_S5 coil?
If yes, check corresponding power supply.
Pay atention to feedback resistors, R7260/61.
 

jpadie

New member
yes. with L7260 removed there is 2.6+v at both pads (which I find a bit odd and makes me suspicious of U7201)

Interestingly R7260 and61 measure 11k and 7.8k respectievely . on a donor board they measure closer to 20 and 9k

replaced the clock chip again. ppvrtc is at 2.3v now. PP3v3_s5 remains the same. I'm not getting a 32kHz signal on the clock chip though. starts at 16kHz and then spirals down to 8.
 

jpadie

New member
also to note that there is no 3v on the middle pin of the magsafe.

U2700 receives pp3v42_g3h. i cannot see any cap that is bad around U2700. there is no short on that rail.

But ... i cannot see the logic that says that the PCH must start up just because U2700 is working. doesn't the PCH need all_sys_pwrgd AND (U2750) CPUIMVP_PGOOD AND (U2760) SMC_DELAYED_PWRG to start up?
 
Last edited:
Top