(347) 552-2258 help@rossmanngroup.com

820-3332 - Momentary fan spin

V

victor

Member
Hello wonderful people! Congrats on the new forum - looks great.

I have a problematic laptop here which I need help with.

The history...
  1. It DEFINITELY has no liquid damage
  2. The person I got it from claims that it was working perfectly, they changed the battery (for a non-genuine one) and then the problem appeared
  3. Upon opening the laptop I was immediately aware of a gold-sparkly substance which was literally spread all over the board. It turned out to be a thermal paste that they had used on the CPU heatsink. I have no idea how they managed to get it everywhere.
  4. After studying the board relentlessly under scope and confirming absolutely no signs of corrosion - I elected to ultrasonic it in my crest to remove the nasty paste. I did test a small clump of the paste between the probes of my multimeter - it doesn't seem to have continuity or even register a resistance, so I think it is unlikely it shorted anything out.
The problem...
  • You can plug it in, or use it with the battery and the same thing happens either way.
    • When power is applied, the fans spin for a moment and then stop. Nothing else occurs.
    • Occasionally the fans will spin for a while before stopping, sometimes they barely move.
    • Once in a blue moon, the fans spin up to full speed and remain that way - nothing else happens.
      • I have only been able to recreate this twice.
What I've tested...
  • I have narrowed the problem down, i think...
    • Working down the rails, the first S4 rail not to be present is P5V3V3_S4_EN
      • This seems to be created by the presence of SMC_S4_WAKESRC_EN and PM_SLP_S5_L, NEITHER of these are present and I cannot figure out how they are made.
Really looking forward to any help you can offer!

Thanks.
 
2informaticos

2informaticos

Administrator
Staff member
When fan spins, 5V_S0 appears; obviously 5V_S4 is present at same time.
If all voltages disappear, faster or not, could be a SMC sensor issue.
So try first to start in SMC bypass mode.

Test the board alone, no cables connected, apart of charger.
Post exact PPBUS_G3H, 3V42_G3H, and 3V3_S5/SUS voltage.
Check if SMC_RESET_L is stable.
Also post PM_BATLOW_L, PM_RSMRST_L, PM_PWRBTN_L, PCH_DSW_PWRGD, PCH_DSWVRMEN and the 4 RTC signals.
Trigger SMC_ONOFF_L and check for drop pulse at PM_PWRBTN_L.
Use R5015/16 pads, instead of power button.
 
V

victor

Member
Thanks a lot for your reply.
Board is completely disconnected from everything apart from charger and fans.

No result from SMC Bypass mode.

PPBUS_G3H - 12.25v
3V42_G3H - 3.42v
PP3V3_S5 - 3.33v
PP3V3_SUS - 3.33v
SMC_RESET_L - 3.38v
PM_BATLOW_L - 3.32v
PM_RSMRST_L- 3.30v
PM_PWRBTN_L - 3.41v
PM_DSW_PWRGD - 3.41v
PCH_DSWVRMEN - 3.21v
PPVRTC_G3H - 3.33v
RTC_RESET_L - 3.32v
SYSCLK_CLK32K_RTC = 0.433v
PPVOUT_G3_PCH_DCPRTC - 1.552v

PM_PWRBTN_L drops to 0.35v when SMC_ONOFF_L is shorted to ground and then shoots back up to 3.41v.

I note that the problem seems to have gotten worse - the fans now don't spin at all, but i hear a little click like they are trying to.
Hope I covered everything you asked for there!
 
2informaticos

2informaticos

Administrator
Staff member
"Hope I covered everything you asked for there "
Not exactly...
"RTC_RESET_L - 3.32v "
If you look at RTC block inside PCH, there are 3 signals more.
Probably bot related now, after other values reported.
But keep in mind that for the future.

"PM_PWRBTN_L drops to 0.35v when SMC_ONOFF_L is shorted to ground and then shoots back up to 3.41 v"
Check for pulse at all PM_SLP_Sx_L signals at same time.
Skip PM_SLP_SUS_L, is already steady high.
 
V

victor

Member
Okay! Will remember that for next time - thank you.

PM_SLP_S5_L - pulses to 3.33v and down to 0
PM_SLP_S4_L - pulses to ? around 2v and then down to 0
PM_SLP_S3_L - pulses to ? around 3v and then down to 0
PM_SLP_S3_R_L - pulses to ? around 3.2v and then down to 0
 
2informaticos

2informaticos

Administrator
Staff member
Then all S3 even S0 voltages should pulse at same time; maybe not VCORE.
So check all big coils and pages 68/69 too.
 
V

victor

Member
I'm not exactly sure what you're looking for when you say big coils. If you're talking about the large inductors on the board...

L2406 - pulses to 0.8v
L7530 - nothing
L7520 - nothing
L7510 - nothing
L7550 - nothing
L7560 - nothing
L8960 - nothing
L8930 - nothing
L9830 - pulse to 0.5v
L7330 - pulse to 0.5v
L7030 - 12.25v
L7220 - pulse to 4v

As for pages 68/69
P1V8S0_EN - Pulses
P1V8S0_PGOOD - pulses to 3v
PP1V8_S0 - pulses to 1.8v
PP3V3_S5 - solid 3.32v
P1V5S0_PGOOD - pulses to 3.2v
PP5V_S0_P1V5_LDO - pulses to 4.4v
PP1V8_S0_P1V5_LDO - pulses to 1.6v
PP3V3_SUS - solid 3.32v

Sorry if i have misunderstood your directions
 
2informaticos

2informaticos

Administrator
Staff member
Resolder U8900; use leaded solder.
Post exact ohm value to ground at L8930/60.

Did you check all 9 output voltages from page 69?
If no pulse on any, check it for short to ground.
Also check 1V05_SUS (page 68).
 
V

victor

Member
"Resolder U8900; use leaded solder. "
I should have mentioned - this was the first thing I did. I've done it and checked it twice.

L8930 Resistance to ground - 26.2ohms
L8960 Resistance to ground - 26.2ohms

1V05_SUS - 1.062v

PP3V3_S4_FET - pulses to 1.7v
PP3V3_S3_FET - pulses to 3v
PP5V_S3_FET - pulses to 4.6v
PP1V5_S3RS0_FET - pulse to 1.1v
PP3V3_S0GPU_MISC_FET - no pulse, steady at 0.06v (checked for short - no short!)
PP3V3_SUS_FET - 3.32v steady
PP5V_SUS_FET - 5.00v steady
PP3V3_S0_FET - pulses to 3.32v
PP3V3_S0GPU_FET - nothing (checked for short - no short!)
PP5V_S0_FET - pulses to 5v
 
V

victor

Member
Oh dear... the plot thickens.

After the above, I decided to resolder U8900 again, this time removing it completely from the board and once again replacing all the solder and soldering it back on. The joints looked absolutely perfect and look perfect again now. However, after doing this - the behaviour of the laptop temporarily changed. The fans were turning on and remaining on, spinning up faster and faster and then eventually switching off after about 45 seconds, still nothing on screen though. The fans still did this even with the trackpad connected. I was able to start it in SMC bypass mode, but nothing on screen and same switch off after around 45 seconds, however i know that it was in bypass mode because the fans came on full rather than gradually climbing. So I opened up the boardview and schematic, got the multimeter back out and went to turn it on again to begin testing.... now it's back to doing the momentary fan spin! sometimes spins for a second or two, sometimes barely spins at all. This is making me think that there is an intermittent connection somewhere or a faulty chip which is being influenced by heat and movement. What do you think?
 
V

victor

Member
is there anything I can do to absolutely confirm that the GPU is faulty? Also anything i can do to disable it and just use intel gpu?
 
2informaticos

2informaticos

Administrator
Staff member
Few hardware mods available to disable GPU on this board; never tested however.
I recently posted on other thread, use search function...
 
V

victor

Member
2informaticos said:
Few hardware mods available to disable GPU on this board; never tested however.
I recently posted on other thread, use search function...
Click to expand...

Thank you so much for all your help.

Are there any signals i should check to absolutely confirm the GPU is at fault?
 
V

victor

Member
2informaticos said:
This is not an exact science.
But your description points to GPU...
Click to expand...

Okay, So i have more news on this.

I searched for the dGPU hardware disable mod as you suggested. I found it on the facebook post and I did it, all of it - it was not fun lol.

The board still has the exact same symptoms though! So that suggests that the problem is NOT the GPU. Can you help?
 
2informaticos

2informaticos

Administrator
Staff member
Disabling GPU doesn't remove it from the board.
GPU can still block some data lines of CPU.

If is not GPU, is something which reacted to heat, usually a BGA chip...
 
You must log in or register to reply here.
Top