A lesson in gathering all the data first - LFC
ฝัง
- เผยแพร่เมื่อ 29 พ.ค. 2024
- This laptop has a dead PCH, but I got so hung up on trying to find a main-rail short that I didn't bother to properly check the secondary rails until some 2hrs into the work. A casual glance at the 1v and 3v rails being shorted to each other should have told me what I was getting into within five minutes.
Check out Adamant IT 2 for the Pod Cast and More: / adamantit2
I stream on games Thursdays and Sundays, 7pm UK: / nethesem
Join the Discord right here: / discord
Support the channel directly at: / adamantit
Follow us on Twitter at: / adamantit
Check out our Instagram at: / adamant.it
Chapters:
00:00 - Intro & Disassembly
05:00 - The GPU is low resistance
15:35 - Power Injection
17:30 - Secondary Rails
21:22 - Everything is shorted together
24:55 - The PCH is Dead - วิทยาศาสตร์และเทคโนโลยี
Side note, this is why we always inject at 1v or less, because you never know when the injection voltage might go to unexpected places. If you're below the voltage of everything on the board, the risk of causing further damage is minimal.
100%. 0.5-1 volt injection is always safe to my knowledge.no chip is gonna blow with voltage that low.Dont forget to set limits to current as well,start from very low and slow and stready increase if needed.Also,as a general rule, pch cpu,gpu surfaces,need to heat up uniformaly when taking power.If during injection the surface heats up partialy,its gone to silicon heaven.
These acer helios and predators almost always have a dead gpu or pch. Very rare i have found one not to. I have replaced mosfets to get light on charger but nothing. Terrible laptops when they go.
As you demonstrated, Compal boards allow power rails to be isolated for testing by removing the solder blobs provided for this purpose. This aids repairability and should be on all boards.
An easier way to check if a high side mosfet is shorted or partially shorted is just to check the resistance between drain to source, since in those vrms the phases are all in parallel if one mosfet is shorted then all of them will show as shorted. If the resistance is high then you pretty much know that all the high side mosftes are not shorted, still doesn't mean they're not faulty but you know the supply didn't shot through.
All of the mosfets are in parallel but their gates are not, so like Graham said, measuring their gates is the best bet for spotting the bad one. Mosfets generally (not always) when they get shorted they short all the terminals together so the bad mosfet will have a different reading from drain/source to gate compared to the other ones in the same vrm.
Similarly you can check resistance from the main power rail to the coil which is the same as measuring drain to source.
Done watching, thank you very much for the informative troubleshooting & repair video. I have learned significantly more troubleshooting & repair lessons in this tutorial video and to your other repair videos as well compared to my ENTIRE 4 YEARS OF COLLEGE EDUCATION. I hope you will soon have a mini-series for Schematic & Boardview-free Voltage/Power Rail Tracing[12V/18-20V Main Voltage Rail, 5V, 3.3V, CPU/GPU Core Voltage Rail, DRAM Voltage Rail, IGPU Voltage Rail, System Agent/Northbridge Voltage Rail, PCH Voltage Rail, BIOS Voltage Rail, Battery Power Rail], Proper method of testing/checking of potentially faulty MOSFETs & ICs/Controller Chips, CPU/GPU/PCH Reballing and BIOS Bin File Editing.
Another youtuber i watch had a similar problem with an 1080 he was sent to fix. The first thing the previous technician did was measure resistance to ground on the vcore coils, saw a low resistance and started removing everything that is connected to ground on that rail, he removed all the low side mosfets saw that the short didn't disappear and called a dead chip when the real problem was a shorted capacitor on the 5v rail.
This board you were trying to fix is a complete nightmare on the other hand, shorts everywhere.
This was good content though and a lot can be learned from this for sure.
Great video.
Cheers!
Hey, thank you so much for sharing your knowledge, so methodically it's wonderful ! Great teaching skills btw, can't get enough ! Cheers
Never seen the soldered jumpers before - learning something new all the time so thanks! :)
There are several reasons why modern GPUs and CPUs have a low resistance, but the easiest explanation is based on Ohm's law. Modern processors run on much lower voltages then most other IC's. In the early days of transistor IC's 12~14 volts was not unusual. While currently most processors run on 1.0~1.5 volts. The number of transistors in the CPU has also increased exponentially by staggering amounts (currently we're in the 1 billion to 10 billion range for CPUs and twice that for GPUs). With the increase in computational power, interconnectivity has also increased and more outputs are required to turn high to signal to other IC's that the processor is turning on. Meaning more current paths are open while the processor is booting up. GPUs usually have a lower resistance than CPUs because they have higher parallel processing capability (which requires more transistors to be in parallel).
Thank you for the information.
That is a good valuable Lesson. Thank you for teaching us. Have a good health for helping us further, then.🎉❤
Thanks for the video brother. Been slowly picking things up to start board repair.
It is interesting to see how to properly diagnose a dead mainboard that i work on on a daily basis. Instead of being told to just replace the whole board. These videos make me want to properly try and fix laptops instead of having to through as many systems as possible.
Masive lesson learnt... reminds me of the night and day I removed a bag full of components then stopped by the very last mosfet.
I learned so much from this.. Thank you sir..
Always fun to go on a diagnostic caper.
"This is Acer Preadator Hideous" ;-)
15:50 I was triggered by the use of a red wire for the GND connection !
A very interesting one again thnx.
Godbless you sir,awsome effort😊
I am really wondering what caused PCH to short in the first place. Was it a power surge?
It helps to place the, "laptop's name", in the title. People actually search for laptop
repairs on TH-cam by name, believe it or not. **Its a difficult concept for ants I guess.**
I’ve recently started watching these videos as I’m getting more into board repair and really enjoyed the amount of knowledge that’s passed on to viewers. Any ideas on groups to look into as I have a graphics card on my bench I’m slowly working on repairing?
hey! long time, no see! Thanks for the video!
Excellent 👍
I always struggle to get solder to bridge across pads (intentionally). The previous repairer seems to be a pro in that regard!
You need a multimeter that "short" beeps on reading a normal diode and "long" beeps when there's a short.
We all know ALL multimeters in existance beeps on (when you touch both +- leads) continuity mode (resistance) BUT there's better multimeters that beeps on diode mode as well.
Besides the overpriced YELLOW brand, I've found a few cheap ones that does this, Uni-T UT60 & Uni-T UT89. Costs only $30 or less. You must have them!
I fixed one of these with a shorted GPU mosfet and luckilly it returned to life. Hard to spot the short but was possible
I would think the reasoning could be. Find short - inject - find regulators getting hot - check pinouts on datasheet to inject again at the output side of reg that will then allow you to see heat on PCH. Dead PCH done. I was wondering why you took so long to inject. Maybe prefer hunting with probes may be. Anyway. Thanks for sharing. Always learning.
Because mr Adamant has learned the hard learned lesson that there is such a thing as injecting too early. And most experienced technicians have learned this by eagerly injecting at the first sight of a short, and bricked silicon on a bord that was actually fixable. Thus sending the GPU/CPU/PCH to an early grave.
@@sindreskjelbostad6436 injecting at 1v will avoid those issues. How can 1v kill a CPU or PCH?
@Swenser it probably won't, but it altso depends on much amp you send. And at such low voltage, a short may not show, especially with a sub par thermal camera, tempting you to increase the voltage. And while 1v prob won't do any harm, 1,5 probably will. At any rate, if voltage injection is always your go to, you never don't really learn much about the electronics you are repairing. Wich is what Adamant and a few other technicians are trying to learn beginners.
Hi , what software are you using for board viewing and reading schematic cause the one you're using here is so good and I need it
Boardviewer
I know they say if it's not a live don't bother typing till you've watched it all. But my brain says @7:07 there are 4 smc ceramic caps very close together and just to the left of them (Up by his left thumb), there are 2 black rectangular items, the one to the right is that a hole or is that Pin 1 ID? Just curiosity as the other has no hole or dimple?
what about reflowing , reballing the same shorter PCH. could that clear the short
Not gonna argue with your result without testing the laptop myself, BUT:
- I find it odd for a piece of silicon to "connect" multiple rails so perfectly to ground at the same time. Usually (IMHO) with the PCH, only one rail is shorted and that's game over.
- When I declare I PCH dead, I want to inject power to the short and see the PCH getting hot and nothing else. If I don't see that, I'm not done.
yes, you have to see a single dot spot on the PCH if it is a dead one
I'm told by folks I trust (I checked in with them while I was working on this one, just before I did the conclusion saying I'm sure the PCH is dead) who says that it's quite common for a dead PCH to short the 1v and 3v rails together. I agree that injection to see some heat on the PCH to confirm it makes sense, but with the entire board all showing 0ohms and likely at least one bad regulator somewhere, it felt like injection wasn't showing anything useful, like trying to find a hole in a sieve.
@@Adamant_ITThanks, the part about 1 and 3V beeing both shorted is good to know.
As for the conclusion:
I don't disagree. I just presented my way of definitively declaring an IC dead and that would have been in this case: Remove any other part, that gets hot during voltage injection, until either the short goes away or the PCH is the last thing, that gets hot.
What about replacing the PCH? What would the cost of that be?
i know you have said but what brand microscope do you use
you can replace the pch from a donor board tho if you are good enough with reballing and stuff
I asked Graham if he took in donar parts but he doesn't coz ot having loads of stuff stored amd in some cases never used but I know one guy who does but then again he has more staff plus storage space.
Yea PCH replacement is possible... but you need the right PCH, with the right firmware, then reball and replace - and after all of that, then you have to fix anything and everything else that might be broken on the board, and it's likely that one or more secondary regulators are busted as well.
If anything, I'd say this board is more useful as a CPU/GPU donor to another board with simpler faults.
Either way, to the next tech to look at it, _good luck_
@@Adamant_IT But you'll tell the customer that the laptop is dead, right?
No, I told them the PCH was dead. They're seeking a PCH replacement, but as per my previous comment, that's a job I'll let someone else take a swing at.
@@Adamant_IT Ok.
How did you get a badcaps login? I've signed up multiple times, but never successfull..
I've had mine for a while, I'm not sure if they're harder to get these days, I just signed up like any other forum - but that was like, five years ago now.
Plz how you get the schematic is it an application or what ??
Schematics are usually PDF files, and occasionally you can also get a board view (which shows where things are located on the board) which requires an application like FlexBoard View to view.
You can find them for googling for the board part number, for example a common MacBook Air board would be "820-00165 schematic"
@@Adamant_IT thank you i enjoy your videos keep going 💕
Why PCH dies so often?
Because it has so many connections to so many places, including input/output ports, that might receive overvoltages (ESD) a lot more often then the parts inside your laptop.
Excuse my ignorance, but what is a PCH? Googling it says it's a Platform Controller Hub for Intel platforms, but I'm still no nearer
Well, there is a Wikipedia article titled "Platform Controller Hub". So reading that should answer every question you could have.
If are more familiar with the term "South Bridge", you can also think of it like that.
Really short answer would be:
It connected the CPU to "everything" else, like Ethernet, SATA, USB, Wifi, Keyboard/Trackpad and so on.
@@NebukadV I saw that on the wiki article - what I don't get is how (or why) this chip went 'rogue' in the first place if no other fault was found; something must have caused a fault in the chip.
@@Tapsnapper This has nothing to do with your initial comment, but whatever:
As the PCH has electrical interfaces to the "outside world", like USB to name the most likely corporate, it can be killed from any overvoltage or ESD that might be happening on those interfaces (from dodgy mains wiring inbetween different devices and faulty devices).
Or one of the voltage regulators feeding it might just have sent 12V to it. This kills a PCH exactly like it kills a CPU. Those voltage regulators all looked sus throughout the whole video.
Can you replace the board?...I just checked the price for a board...$700 new...$300 used so I guess not...not worth it.
Why didn't you try the Chipmunk?
The chipmunk can give useful information, but only if the board is able to switch on. It can't tell you anything if the board doesn't power up. If the PCH had failed without shorting everything together, and we had a No POST, that's when the chipmunk would be a useful indicator.
as to why cpu/gpu have low resistance . ( feel free to rectify .. im just a humble enthousiast who likes looking stuff up ) imagine pushing 100A at 2.5 v into an 2 ohm load xd.. your power company will love u .. and your pc will dual function as a stove .. xd. ( loads of transistors and nodes in series and parallel is how i understand with my ignorant brain.. just put 10 bilion mosfets in parallel .. you only need to look at them to have rds-on to switch open ...)
Why do gaming laptops look so ugly?
because it is a cheap Acer😊
First! yay!
Acer is hot garbage, cheap parts low quality
Yuck the charge cable connects in the middle at the back. I'm offended. I'm gonna throw up....the thing, the whole thing, it's hideous.
its a Gaming laptop so you don't really ever unplug that cable so its better at the back out of the way.
@@KBFix ok
this was not a pch we fix this all the time this tech is inexperienced
What would your theory be? Do you think it's just a few bad regs pulling everything down?
If you "fix this all the time" then you know what the problem could be, right? Teach us! Tell us what could be wrong with the board! Don't just make a comment like that. Thank you in advance. We'll be waiting for your help and explanations.
From a business standpoint, how do you suck the cost of the time in events like this? Do you charge the customer still and if you do is it just a nominal fee?