Dell Latitude 7390 - all power rails up except VCC_GT, no image.

[kitor]

Hello,

I have Dell 7390, LA-F311P.

AFAIK there's no schematics available, but there's one for LA-F312P in this thread. It differs in no Thunderbolt controller (iirc that's for 5290/5390), however rest of the schematics/BV seems to match 1:1.

Behavior:

Connected barrel plug, no RAM:
- power LED turns on + orange light on front
- turns off after a second or two

Connected barrel plug, with RAM:
- power led turns on + while light on front
- after a few seconds power led turns off, white light on front stays lit

Connected Type C PSU, with or without RAM:
- power led and white light on font blinks for half a second, nothing more

Connected battery, with RAM:
- power led turns on + while light on front
- after a few seconds both turns off

Extra checks (barrel plug or battery, with ram)
Hold D when connecting power activates LCD test

Hold CTRL+ESC when connecting power and keeping it held (bios recovery combo)
- power led turns on + white light on font
- after a few seconds power led turns off, but immediately goes lit again
- this happens in infinite loop

What I checked:

BIOS dump seems sane (or at least UEFItool doesn't complain) - see attachment. I tried two other dumps that I found on this forum without any luck.

While laptop is powered from battery/barrel plug AND power LED is on, all voltage rails show up, except +VCC_GT. By all I mean all - I checked even stuff like +3V3_CAM... See attachment - rails in red are always on, in blue are on while power LED is lit.

According to schematics, ISL95857AHRTZ shall start +VCC_SA, +VCC_GT and +VCC_CORE at the same time (via IMP_VR_ON).

PU612 seems to get all the signals from ISL.
- Output resistance to ground is around 11ohm.
- PWM_GT is 0.5M to GND, FCCM_GT is 1.5M, which matches other two rails.
- There's +5V_ALW at pint 3 and 15 (VCC and PVCC)
- There's around 4,8V on pin BOOT, which matches other two rails while they are down. While SA and CORE are up, this pin measures around 5.2V.

As I had damaged 5490 board that uses virtually the same power circuit, I swapped ISL and DryMOS, no change in behavior.

Any suggestions where to look next? BIOS? EC firmware? Something around ISL? Bad CPU?

[kitor]

Re: Dell Latitude 7390 - all power rails up except VCC_GT, no image.

This weird behavior on USB C PSU led me to testing that part. Turned out UT7 was quite hot and giving around 3V instead of 5, UT8 was outputting around 1V instead of 3V3. Output of UT3 was around 12ohm to ground which was definitely bad.

I injected voltage to UT3 output and found out that Thunderbolt controller (UT1) is pulling around 0.3A and getting noticeably hot. After disconnecting power to UT1, machine boots from USB C power supply.

Still no boot from barrel plug or battery (same symptoms as before).
When booted with USB C PSU battery is charging. After disconnecting C power supply it runs on battery, but very slow.
CPU slows to 400MHz and it won't recover even when C PSU is connected again.

So there's definitely something more to fix, but that's a progress.

[kitor]

Re: Dell Latitude 7390 - all power rails up except VCC_GT, no image.

Ok, I found the last piece of the puzzle.

After much measuring I found that +5V_PD_VDD is very low with battery/barrel plug power. I also noticed that 5V_TBT_VBUS is around 1V, which was interesting as in that power configuration, thunderbolt was inactive and this shouldn't happen.

Quick look into schematics determined DT2 diode as source of the problem. After replacement, laptop works fine on all three power types.

Just the TBT3 controller is left to be replaced later.

[mon2]

Re: Dell Latitude 7390 - all power rails up except VCC_GT, no image.

Excellent!! A very nice read. Thanks for the updates.

[kitor]

Re: Dell Latitude 7390 - all power rails up except VCC_GT, no image.

Funny thing - even though TBT3 controller (JHL6340) is broken, it kinda works.

For future readers - there's a schematics available for 7280 (LA-E121P) which is based on older version of the same project - and TBT3 part is almost identical.

Issue is with 3.3V_TBT_SX (or VCC3P3_SX) which is to my understanding suspend power. After reconnecting PJP6, TBT3 IC gets quite hot instantly (even at suspend) but TBT3 bus works - I connected to a dock with no issues.

Interesting design is that 3.3V_TBT_SX comes from +3.3V_VDD_PIC which itself is generated by UT7 / AP2112K-3.3, rated up to 650mA (more than enough to supply everything on that line).
But the 5V supply line for UT7 isn't rated for that, with around 1W (~300mA at UT7 output) voltage drops to 4V and DT1 (sourcing 5V from +5V_ALW) is getting warm.

I compared just in case to other board, and indeed JHL6340 itself just needs a replacement.

[Zilicon]

Re: Dell Latitude 7390 - all power rails up except VCC_GT, no image.

Do you have an idea on one that shows BSOD immediately +VCC_GT gets power?

the machine runs fine when there's no voltage at the PL612, but BSOD with the error "WHEA uncorrectable error" will be displayed immediately after the PL612 gets 0.7v.

BSOD happens frequently but the system works fine in safe mode.

I got the same BSOD when I tried to boot from an external device