Microsoft Surface Pro 5 1796 - No Power

[j64x86]

Hey There,

Trying to troubleshoot a SP 5 1796 (M1007506-015), most of the caps around the CPU area are shorted.
When tracing from the DC in the first shorts I could find is near the USB & DP connectors.
Tested with diode mode, red to ground.

Two bad caps, marked in red on the attached picture, I currently don't have a bench PSU to attempt power injection, any best guesses as to which of the two could be the culprit?

[mon2]

Re: Microsoft Surface Pro 5 1796 - No Power

Link or post a schematic for the logic board.

Cpu power rails are low voltage rails. Respectively, they will also be low resistance rails.

With no power, measure the resistance of these rails and post their exact resistance readings.

With power, check the voltage to ground of each large coil. Each coil is an inductor and is linked to a local switching power supply.

[j64x86]

Re: Microsoft Surface Pro 5 1796 - No Power

Unfortunately I don't have schematics, really wish I did, couldn't find them without some sort of paywall.

I measured the voltage with the 15v DC connected with and without the battery, it appears that the battery does receive some voltage.

[mon2]

Re: Microsoft Surface Pro 5 1796 - No Power

See attached.

[mon2]

Re: Microsoft Surface Pro 5 1796 - No Power

Start with measuring the voltage to ground of:

a) charger IC U6302 -> pin #28 REGN

b) +VSYS

[j64x86]

Re: Microsoft Surface Pro 5 1796 - No Power

Thank you so much for the power rail info.

on IC U6302
pin #28 REGN = 5.98 V
pin #22 VSYS = 8.8 V

[mon2]

Re: Microsoft Surface Pro 5 1796 - No Power

Personally have not worked on one of these but the rails look to be ok.

See attached for the power up sequence. In the end, step # 24 is the critical one and this voltage should not be 0 volts. That is, this is an active LOW signal (denoted with the # suffix) so if this rail is 0 volts = logic low, the target is in RESET mode. Respectively, if all is well, this output should be a logic '1' = ~3 volts.

Can you confirm it? It if is indeed ~0 volts, then each step from #1 onwards must be reviewed as shown in this chart to locate the fault.

What is the voltage to ground of the attached and gate @ U2202 ?

[j64x86]

Re: Microsoft Surface Pro 5 1796 - No Power

It's my first time attempting board level repair on something as advanced as this.

Good news, step 24 is reading 0 Volts, so I have the opportunity to troubleshoot Steps 1-23

[mon2]

Re: Microsoft Surface Pro 5 1796 - No Power

Yes, proceed with the chart in order and you should locate your fault. Continue to update your review. At this time, the conditions have not been met to release this reset line to be a logic '1'. Have to do some toilet repair for my mother. Yuch. Will check back later.

[j64x86]

Re: Microsoft Surface Pro 5 1796 - No Power

Good luck with the plumbing!

Had a go at testing voltage on the various component. Couldn't perform step 6 as the group of components is not present on my board, the other point is under the processor. Step 7&8 are both 0 volts, Step 7 has a # so assuming it should have positive voltage.

Now I'm not sure if the answer to the riddle is in Step 6?

Step 1: 3.299V
Step 2 3.3V
Step 3. 3.291 V
Step 4: 8.8v
Step 5: Skip
Step 6: PCH_DPWROK / Declare deep power good to PCH (can't locate on board)
Step 7: 0 Volts!
Step 8: 0 volts


Step 6


It's getting late this side, happy festivities!

[j64x86]

Re: Microsoft Surface Pro 5 1796 - No Power

Interesting observation, when pressing the power button the smaller portion of the CPU instantly heats up for a moment....

Not sure what this means but it's a new data point.

[reformatt]

Re: Microsoft Surface Pro 5 1796 - No Power

That's the PCH ;-(. Check the PCH power rail (usually +1.05V) resistance to ground. Usually thats at least 100 ohms or more.

[j64x86]

Re: Microsoft Surface Pro 5 1796 - No Power

That's the PCH ;-(. Check the PCH power rail (usually +1.05V) resistance to ground. Usually thats at least 100 ohms or more.

I tested this +1VSB to ground and it's 0 ohms...





The caps connected to this fuse U4703 are also shorting to ground, not sure if this is the behavior of the fuse or not (pn AP2553FDC).. U4703 seems to be related to the display port so might be OK to ignore,





I know the the 3.3v rail does have power as I get 3.3V from the SAM, not sure where it's shorting


In other news... I noticed a 2.2uH inductor gets hot. (related to the charge circuit)

[j64x86]

Re: Microsoft Surface Pro 5 1796 - No Power

Re-testing a bunch of things..

If I measure VSYS on R6319 I get 0.737 V is this a little low?

[mon2]

Re: Microsoft Surface Pro 5 1796 - No Power

If pch is heating up and it's power rail is showing the low resistance to ground (ie. 0 ohms for this case), then pch is shorted. CPU is dead.

[j64x86]

Re: Microsoft Surface Pro 5 1796 - No Power

If pch is heating up and it's power rail is showing the low resistance to ground (ie. 0 ohms for this case), then pch is shorted. CPU is dead.

That did cross my mind... oh well it was fun investigating.
When I have more time I will see if I can figure out what shorted pch (besides the CPU)

Thanks for all the help, this forum has solid members!

[ReeceyBurger123]

Re: Microsoft Surface Pro 5 1796 - No Power

Inject voltage see what gets warm. I’ve had a bad MLCC in the mix of the cpu caps on the pro 5 ��

[reformatt]

Re: Microsoft Surface Pro 5 1796 - No Power

For future reference, 1st troubleshooting step is to measure each coils resistance to ground. This would have shown you the short on the 1V line. Voltage injection/thermal camera is then used to pinpoint the short which in this case is your PCH. Alternative is to use some alcohol on the PCH and inject 1V and see if it evaporates quickly. Logic suggests that your PCH doesn't get warm if there was something else on the rail that was shorted.

[j64x86]

Re: Microsoft Surface Pro 5 1796 - No Power

Inject voltage see what gets warm. I've had a bad MLCC in the mix of the cpu caps on the pro 5 ��

I took a chance and injected onto VSYS rail and this FET (CSD87334Q3D) heated up, after removal short to ground on VSYS is resolved.




Now VSYS powers the VCORE rail... not sure if the PCH is toast as a result of the failed FET or if it somehow was protected.

VCORE is open, probably because VSYS is awol

[j64x86]

Re: Microsoft Surface Pro 5 1796 - No Power

For future reference, 1st troubleshooting step is to measure each coils resistance to ground. This would have shown you the short on the 1V line. Voltage injection/thermal camera is then used to pinpoint the short which in this case is your PCH. Alternative is to use some alcohol on the PCH and inject 1V and see if it evaporates quickly. Logic suggests that your PCH doesn't get warm if there was something else on the rail that was shorted.

Great general advice, I'm still learning and I've learnt soo much about S0-5 states, power rails, logic etc.

I reckon a decent microscope, bench supply and thermal camera will go a long way in diagnosing shorts