820-00138 won't boot when SSD is connected

[ajdoc]

Cant find any corossion.
Board won't boot when PCIE SSD is connected (SSD controlled on SSD is hot).
After press power button fans are spinning (trackpad clicks, caps lock LED not working) and screen is blank.

As soon SSD is removed all works fine if boot from USB port.

Tested with known good SSD and same issue.
Checked both SSDs on another machine and SSDs are working fine.

Compared diode mode with working board of J3700 and all readings are correct.

Any idea?

 

[2informaticos]

Did you compare voltage and signals too?

Post U3740 voltages, with SSD connected.

 

[ajdoc]

Signals as well.

U3740/1 - 3.41V
U3740/2 - 0.99V
U3740/3 - 3.32V
U3740/4 - 3.41V
U3740/6 - 3.32V
U3740/7 - 0.0V
U3740/8 - 0.0V

 

[2informaticos]

R3740 is good?
Check diode mode at pin 24 of J3700, w/ and w/o SSD inserted; compare with good board.
Do the same for pins 7, 8 and 47-50.

 

[ajdoc]

R3740 is good

J3700/24 - the same values on known working board:
w/o SSD - 0.456V
w/ SSD - 0.386V

J3700/7 - the same values on known working board:
w/o SSD - 0.290V
w/ SSD - 0.290V

J3700/8 - the same values on known working board:
w/o SSD - 0.440V
w/ SSD - 0.387V

J3700/47 - the same values on known working board:
w/o SSD - 0.411V
w/ SSD - 0.411V

J3700/48:
w/o SSD - 0.650V | known good board 0.450V
w/ SSD - 0.650V | known good board 0.450V

J3700/49 - the same values on known working board:
w/o SSD - 1.750V
w/ SSD - 1.750V

J3700/50 - the same values on known working board:
w/o SSD - 0.433V
w/ SSD - 0V

 

[2informaticos]

Are these readings voltages???
PLEASE, stop writing "V" after diode mode readings.
Avoid any misunderstood.

 

[ajdoc]

Sorry, thats diode mode readings

 

[2informaticos]

Then, what this means?
"w/ SSD - 0V"
It is short to ground (0.000), or OL (open)?

 

[ajdoc]

Short to ground (0.5 ohm to ground) when SSD connected.

 

[2informaticos]

Possibly communication problem between PCH and SSD.
It may block PCH somehow.

Just in case, change U3710/11.
Compare their voltage and diode mode readings with good board.
Also compare readings on both sides of all C3710-17.

 

[ajdoc]

Found something wrong.
One of C3710-17 dont match.

With SSD inserted on working board got diode mode reading 0.380 and on clients board got OL.

My 820-00138 boardview is in fact 820-3662 model and that area looks different .
Will show on photo below where got wrong reading.

It looks like it is C3717 on board view.

 

[2informaticos]

You should confirm checking continuity with J3700.

820-00163 boardview is also different there?

 

[ajdoc]

There was connection between C3717/2 and J3700/22 so it was strange why there is no any value when SSD is inserted.

Checked inside of J3700 and pin 22 was bented/damaged.

J3700 replaced and all works fine again.
Many thank for all help.

 

[2informaticos]

Thanks for solution feedback, as this was a really weird issue.
It may help other members in the future.