Ongoing Acer Predator Triton 300 G53G LA-L191P woes

[chaoshusky]

Well, this is the second laptop of the exact same model and spec to die on me, the first one lasting a month. Fortunately that was returnable, but apparently the seller lied to me about this one and had even had its warranty registered, so very much out of luck there and can't afford to lose £900..

Anyway, it had been having issues for a while again with throttling, after fixing it briefly..almost like the heatsink is defective and slowly coming away! Worked great for a while, but started throttling on the GPU and throwing the odd blue screen, even when running in non turbo or at stock settings - absolutely no overclock. Can't say i liked using Turbo mode either, far too noisy!

It slowly got worse, to the point where i had a hotspot reading above 105C! Attempted to improve it by re-pasting and remounting the heatsink, making sure to do it in the proper order.. However at this point it was throttling so much it was barely scoring near an RTX 3060, not a 3070! GPU temp on its own was actually fine and the RAM was increasing but still nowhere near 100C, nevermind 105.. After a second attempt and remount, i did have improvement, almost reaching the typical Time Spy score and scoring about 8000 (usually 9.8K) but was still hitting a huge hot spot temperature exactly the same as before. No pads were missing and all thermal interfaces were present and making contact!

After switching it off and leaving it for a while, came back to the laptop to see if i could spot any bad mounting but all was well. Plugged in the AC adapter, the charging LED lit up blue. Pressed the power button and it would briefly try to power up, not even lighting the keyboard, then both LEDs would go out with the adapter still attached. To change that i had to unplug and re-plug the adapter. After a few attempts of powering it up, however...it died completely! No longer trying to power up, no charge light, nothing. First thought was actually the DC charge and input control but that seems to read fine. As the hot spot was GPU related i started measuring all rails..

FBRAM (GPU), FBIO and CPU rails all fine, measuring a couple dozen ohms.. About 24 on the RAM supply and under 40 on the CPU.. The GPU however? Way less than 0.1ohm, basically a dead short! That was directly at the Driver/VRM packages.. Nearby on the rows and columns of capacitors between the VRM and GPU, it was measuring 0.2.. Having managed to remove the former VRM packages with the usual hot air and flux method etc, this has actually changed to around 0.6ish at the capacitors and where the VRMs should be sat. So, i'm really hoping its not the GPU as that will be a pain to replace. And i definitely can't afford to just sell it for parts or pay over £400 for a replacement board. So any help would be massively appreciated! The Driver/VRM chips i removed do also measure somewhat differently to the replacement chips i have, which there are 5 of so one spare..

Again, any help or suggestions massively appreciated! I'll do some re-checking on the power input stages.. I'm just shocked they design the things so 19V is fed straight to the VRMs! I know its less efficient to run something like that on 12v and have the power supply produce more current, but that's around the level the battery normally is so it seems a little ridiculous that small ICs have to deal with dropping 19V to as low as a volt and under..

Kind regards,
Lea/Husky!

[Sephir0th]

Re: Ongoing Acer Predator Triton 300 G53G LA-L191P woes

I don't know how i can phrase this polite, but what do you expect when you cook your high end GPU for days while "playing" with the heatsink? A upper-end Gaming Laptop is a waste of money anyway, regardless how they are designed. The BGAs can't stand the heat over the time. That's it. Just saying.

Also 12V or 19V doesn't matter. both voltages can kill your buck converter loads with almost the same probability.

So.enough philosophy and basics for.the moment but to the Plot-Twist: Readings below 1 Ohms or even below 0.2 Ohms or lets say 0.1Ohms might be absolutely fine for the any upper end GTX 10XX and newer. For this reason this single reading is not enough to start playing with the Buck converter for the GPU_CORE since it might be fine. At least from the shared resistance measurements.

In fact the diagnosis is still missing. Because if the board doesn't do anything anymore then you have a issue way earlier in the sequence. With the Main power rail for example, or the 3V/5V. So you need to move there first to get the clue about what is going on with the board.

[chaoshusky]

Re: Ongoing Acer Predator Triton 300 G53G LA-L191P woes

I don't know how i can phrase this polite, but what do you expect when you cook your high end GPU for days while "playing" with the heatsink? A upper-end Gaming Laptop is a waste of money anyway, regardless how they are designed. The BGAs can't stand the heat over the time. That's it. Just saying.

Also 12V or 19V doesn't matter. both voltages can kill your buck converter loads with almost the same probability.

So.enough philosophy and basics for.the moment but to the Plot-Twist: Readings below 1 Ohms or even below 0.2 Ohms or lets say 0.1Ohms might be absolutely fine for the any upper end GTX 10XX and newer. For this reason this single reading is not enough to start playing with the Buck converter for the GPU_CORE since it might be fine. At least from the shared resistance measurements.

In fact the diagnosis is still missing. Because if the board doesn't do anything anymore then you have a issue way earlier in the sequence. With the Main power rail for example, or the 3V/5V. So you need to move there first to get the clue about what is going on with the board.

Some of this is true, though i wasn't "playing with the heatsink" like you seem to think and there seems to be design flaws with this laptop. It could well be a BGA problem too, but i think you're assuming quite a bit here yourself as i've been able to repair and build things for 30 years. I didn't come here to be insulted, i finally have the diagrams so i can dig more without having to reverse engineer the connections as i go. I guess my 'assumption' is based on the hotspot, the fact that it got worse, the fact that the heatsink is awful anyway and cramming over 200 watts into the thing is an awful idea! Additionally, the resistance changes further away from the VRM made me suspect it more. Having removed them, the resistance changed what i would call significantly and hasn't changed since with the board hot or cold. I have told people for years NOT to buy a "Gaming laptop" yet i got the bug myself, one MSI arrived DOA, the first Acer lasted a month (would power up and cycle the keyboard, pretty sure it screwed up on a UEFI update however i was able to send it back for my money back) and the 'used' one i was lied to about - it was far more used than they said, NOT refurbished and had already been registered for warranty..and was d*cked around with lol

If that reading is normal, fair enough..there are rather a lot of capacitors on the rails and low ESR or not that would likely add to it, but a reading that low is quite surprising! What also made me think it was the Driver VRMs is so far everything else appears normal, though it doesn't give charge any more so could well be one of the power controllers. However with the hotspot reaching over 105C i'd be inclined to replace them anyway as it was power limiting despite the brick and CPU running perfectly well, at that temperature any MosFET device would cook very quickly i'd imagine, especially when having to buck-convert 19V at several amps to 1v at more than 100A! I'll do some more measurements, as the adapter voltage does appear where its supposed to at the very start, so it certainly isn't the brick or port.

Snapping at me doesn't change the fact that these laptops are awfully designed though, and as people want thinner machines it was "just 19.9mm thick" which...does no favours to thermals. And of course their gimmicky metal fan for the GPU. Compared to the first laptop, this second one didn't have as much heat kicking out the sides, leading me to believe the heatpipes weren't doing their job! As well as signs it had been open, had one stick of RAM upgraded and marks on the heatsink screws, someone had already messed around with it. Some K5-Pro thermal putty also seems to be worse than many would lead you to believe, could well be it couldn't put enough pressure down due to it being rather viscous and wasn't touching the VRMs properly and there are a few creases in the heatpipes - not caused by me! Can't imagine they're working like they should so if she comes back to life, i'll replace the cooler as well.

No short circuits or open circuits, odd resistances were noted on any other power control ICs however, as everything is linked by sense and protect systems, pretty sure it went into 'protect mode' as a Dell Alienware did i fixed for a friend. That had simply flattened its battery too far, so manually charged the cells by direct access to revive the protect circuit inside the pack. I then applied the charger and it booted straight up after being completely dead for 3 weeks! The first error was CMOS reset, as apparently in that model, the RTC runs from the main battery, which had gone below save levels. My battery also isn't putting out power at all, so that kinda tells me the battery has been told to cut the output - thanks to them being able to communicate, no? I managed to also fix the thermal problems in the Dell and the fact that two people had previously messed around with it and ruined the Liquid Metal for the CPU. The fact that the resistance has also changed, whilst that could well be a BGA issue, aren't you jumping the gun there as well as the hotspot would be unrelated to the failure. If it were the BGA, applying pressure could have let it boot - i tried that and it didn't. That doesn't always work either though! lol

Any help finding a potential area would be appreciated, even if i have to remove various jumpers, inject a low voltage and test each part. My PSU is regulated and capable of high current, with variable voltage but the adjustment is rather coarse. A small pot to go from 0V to 16.4V, tempted to change it for a pot of multiple turns..

Anyway, thanks for the answer, but again i didn't come here to be insulted. Fixed a fair few BGA problems in my time, even curing the original "Red Ring" of the XBox 360s which still work now and can revive *most* PS3s with the "YLOD" error with a shim...that was a bonding issue of the die and carrier package. They're also quite mature now in terms of lead-free soldering and BGA doesn't seem to be as much of an issue, besides the brand new MSI laptop i got that would only POST and start up after it had been left switched on with a black screen, the power LED flashing an error code. In other words, as soon as it warmed up, it booted, was stable and passed all tests! Powered down and allowed to cool for just 10 minutes, it would die again. That was probably BGA. If mine was BGA it may well have similar behaviour but mine was acting like *something* was shorting a power rail. And i had no diagram or boardview!

Kind regards,
Lea

[Sephir0th]

Re: Ongoing Acer Predator Triton 300 G53G LA-L191P woes

Don't take it as offense, but as a motivation. I imagine it like me me trying to fix a freezer or a dish washer...

Yes, the history sounds worse and I understand why you're directly digging into the GPU circuit. That the thermal issue persisted only for hours instead of days makes the things a bit better.

However, i've seen many dead Laptop GPUs already and if it was killed by heat then the Laptop should be still able to turn on till some extent. The only reason what i can think of causing the behaviour is a so called High-Side short, so one of the DrMOS got internal shorted between Main Power Rail and GPU Core.

If so, or even if something else has failed, there is still a possibility to fix it without replacing the whole GPU. But you need to help us with clear pictures and clear measurements. No schematics avaliable normally won't be a issue in case of a no power fault.

So if the DrMOS or whatever is present there are still removed, the issue could be already fixed (LoL!), but we still need to take a look to the Main Power Rail, so 19V after(!) The second DC-IN Mosfet. Were you able to identify this rail?

Under normal circumstances Tests with power applied are not allowed (floating gates would kill the load), but since these are DrMOS (I've checked pictures of the board) you can still power it on and check for voltages if necessary. But as I read now youu have replaced them already. In this case, the coincidence might mislead us here as explained before regarding the untypical behaviour.

[chaoshusky]

Re: Ongoing Acer Predator Triton 300 G53G LA-L191P woes

Sorry for being a little b*tchy, I'm not always good with words which doesn't really help on a forum does it? Lol..

Thing is with pics, an embarrassing thing was..my hot air station was malfunctioning, not getting as hot as it should, so i've caused a little more discolouration to the PCB.. Had to fix that first of course! Nothing serious, but that's something i forgot to say! The VRM PCB location for the GPU was already "changing colour" if you know what i mean, from the heat? The Chokes don't seem to be damaged, but someone mentioned that too as dead chokes would be bad, or damaged ones. They are all closed circuit and about 0.3-0.4 Ohms which could be bad, obviously i need to get them on a meter that can measure mH! But i need to make some kind of preheater as i can't even buy something as simple as that at the moment, as the one i found was like £200!

Anyway, i can provide pictures or whatever you need.. It seems the RAM IO V+ is fine, the one that also connects to a few BGA points on the GPU to power the Memory Controller i'd imagine? That does measure fine, as do all the PEG/PCIe lines between the CPU and GPU..

And you're right in respect to having removed them, that could well be what's dead anyway lol and i can always (carefully) solder them back on.. As far as i can tell measurements around there seem normal..

I have been locating the +19V input and anything connected to it, have indeed found the DC-IN Fet/ICs so i need to get measuring those some more. I'd guess either a seemingly open circuit or most likely a dead short means failure? Be annoying if i have to buy more chips...haha!

I'd try to make a pre-heater out of some kind of incandescent or halogen bulb, as they put out some infra-red...but from what i can gather, photons and direct heat might actually be too much and cook something, as of course the proper pre-heater can be temperature controlled. With a bulb i'd have to put the flipping thing on a dimmer of some kind then use a thermocouple to find the right temperature range..

Anyway, sorry, i'm rambling.. But perhaps it isn't as dead as i feared? From the symptoms i hope you can see why i jumped to the conclusions i did considering the main problem was the GPU hotspot/throttling and it would still try to power on, then failed completely.. I can only hope it was a low-side Fet death as i'm fairly sure no chip that's designed to run on a volt would enjoy having 19V across it for mere nano or even pico seconds!

As for testing power, since the initial failure i haven't dared plug the PSU back in.. Fortunately i'm blonde but not completely stupid! Haha.. It seems according to the board and schematic there are various 'jumpers' i can use to disconnect certain rail supplies, i assume for testing at the factory before they're all bridged in the right places? If i have to remove all the darn solder jumpers and apply a single volt to the DC-IN jack, i bloody well will!

Kindest regards,
Lea/Husky

[Sephir0th]

Re: Ongoing Acer Predator Triton 300 G53G LA-L191P woes



But no, the Inductors/Chokes are almost never a issue in these circuits. lemme explain a bit more so we don't have a misunderstanding in the end. Look at the attached pictures.

In Picture 1) i have marked in red what i assume is the Main Power Rail after the DC-IN Fets. Share the reading please.

In Picture 2) you'll see a likely known-good overview about the Resistances to GND of the various Power Rails present on your board. I've reviewed the readings and they make sense to me. Compare please.

In Picture 3) i try to describe what might went wrong with the GPU. But still, it's just a guess. You'll notice the Inductors/Chokes circled in RED, and the DrMOS circled in Green. if one of these DrMOS got internal shourted between High-Side Drain (Main Power-Rail) and High-Side Source (GPU_CORE), the Main power rail will will be in fact shorted to GPU_CORE (Purple Arrows) and causes a awful low reading at the Current Sense resistor (Picture 1) as the logical result.

Some GPUs does not survive 19V coming in and some do...

This is just a example of a so called-High-Side short. It can happen with the VMEM supply at the bottom-right too, and also with the supplies related to the CPU. The exact measurement at the Current Sense resistor gives the clue.

But first things first. Meaure resistance of one of These Current sense resistors (Picture 1) to GND and share the result please.

[chaoshusky]

Re: Ongoing Acer Predator Triton 300 G53G LA-L191P woes

Sorry for not checking this for a while, i am desperate to fix this its just a matter of finding time to do it at the moment.. Doctor's appointment tomorrow!

I have done a few measurements myself, including making sure ADAPDET signal still present and other...probably pointless things! I'll check what you've posted shortly.. Get the joke? Shortly.. Heh, that was terrible. Anyway, a few curious things are the RTC battery is at 2.5v, plugged in it now drops to 1.2v.. Battery is disabled so no power, guessing the BATT_B/I signal and other lines have something to do with that! All the cells are charged fully though. Hmm.. I see your board has the RTX 3060, mine has the 3070 but i can't imagine that makes much difference, they seem to share the same layout and pinout.. I did also provide just 1 volt to the charger socket..nothing, no current draw so.. Also checked the PL101/2/3 and all are fine.

Right, I'll do the measurements asap and get back to you! Didn't realize it had almost been a week already..i need an assistant or something! lol

Thanks again everyone, i haven't given up yet! Bare with me and i'll get the board back on my table and check! Thanks Sephir0th, that's exactly what i needed to try and track down a fault! Cool name by the way..though i am starting to wonder if its actually to do with the PUB1 section. Trying not to go down rabbit holes though lol

Best regards,
Lea

[chaoshusky]

Re: Ongoing Acer Predator Triton 300 G53G LA-L191P woes

So.. On that side of PBR46 and PBR3 to ground, PBR46 measures like...38Kohms then slowly rises, the other side when i get a good connection suddenly goes to 247K then slowly drops..that's rather puzzling! A little resistor there measures correctly, but the super small ones nearby close together measure around 800K and 1.7M! I hope my meter isn't on the way out but I've also cleaned the probes.. Don't know if that will help you or not!

I'll now go and measure the rest and report back. Sometimes when i reconnect to those points (PRB46 or 3) it goes into the millions of ohms.. Confused! And to think i'm supposed to be an electronics engineer but this thing is so proprietary and...its dead so i can't even use my Oscilloscope for anything!

Kind regards,
Lea

[chaoshusky]

Re: Ongoing Acer Predator Triton 300 G53G LA-L191P woes

Well...this isn't looking good! I assumed 0v means a probe on ground, i tested each side of the inductors and here are the disturbing results..

Okay, from left to right, doing the 'top row' first along the board's longest edge we have 6.67M not 4M, 0.07ohm not 1.8k, 3.4k not 31k, 0.5ohm not 31k and 0.06 not 1k!
Next row is:
567ohm instead of 400, 0.4ohm on the GPU inductors not 1.1 (it used to read that across the inductors and across the GPU rail reads 0.6 now, i have installed the new FETs but can always remove them) and 28 instead of 105, which is a RAM rail and probably fine?
Final 'row' at the smaller edge of the board:
The first 4 inductors measure 50ohm not 28, the 5th measures 246 not 158 abd the last one near the DC-IN socket measures 18 instead of 11!
Obviously, a few of these are fine but..what the hell happened to my board?! I just wanted to play some games, had to sell my good PC to cover some bills so now I'm left with a machine that has a 6th Gen CPU and a GTX 1650.. Argh!

Any ideas? This is...strange...

Kind regards,
Lea

[Stevedb]

Re: Ongoing Acer Predator Triton 300 G53G LA-L191P woes

I would put all the information on hd picture so we can see which coil resistance belongs with which power supply on the board. And use best 200 ohms range on your multimeter and give also the resistance from your probes.

[chaoshusky]

Re: Ongoing Acer Predator Triton 300 G53G LA-L191P woes

Argh! I hope Seph can understand what i wrote, i tried my best! My phone camera tends to produce massive jpeg files lol that's why after each measurement i said "not xx" as on the pictures.. I'll do it later when i'm back from the doctors, i hope he hasn't disappeared..

Regards,
Lea

[Sephir0th]

Re: Ongoing Acer Predator Triton 300 G53G LA-L191P woes

I did not disappear, but i am always very busy...

So the good news the measurements are actually very helpful and the GPU doesn't seem to be the issue, but the PCH, which brings us to the bad news.

It looks like 3.3VALW, 1.05VALW and even 1.8VALW are dead shorted to GND and from my understanding there is no point to continue with the repair attempt from here. The indication for a shorted PCH can't be more clear. Sorry

[chaoshusky]

Re: Ongoing Acer Predator Triton 300 G53G LA-L191P woes

So.. A two year old laptop is now useless/only useful as a donor? I did always wonder why they never seem to put a heatsink on the PCH most of the time.. Seems its possible to swap them and clear the ME but you're saying its pointless? Just great.. How can Acer get away with that?! It cost me £900, the seller lied and it had its warranty registered, so i can't even return it to them.. I guess i should contact Acer and see if they will do anything? I can't really afford to replace the motherboard or get another laptop, selling the spares from it won't let me a whole lot back so..no idea what to do now. Not your fault obviously, the PCH being down does make sense, as i found other threads where the symptoms seemed the same including how it tried to power up then eventually failed there as well. Hmm.. Or i guess i could try and get another board that has a dead CPU or something? I did spot one on eBay where they claimed it was the CPU that was dead but..how would i know that's really what's wrong? I could get it and it also has a dead PCH..

Right to repair eh? Well that's a few months wasted trying to get it fixed.. No insurance either. I always said to people, don't buy a gaming laptop, they die..not even sure why i bothered at this point. Even the "brand new" MSI laptop i first ordered back at the end of 2021 arrived dead.. It would switch on, but wouldn't boot and flashed an error..however if i left it like that for a couple of minutes, forced power off then powered back on, it would boot and work fine! That would likely be a BGA issue? (When warmed it works) Which is...ridiculous. At this point i have nothing to lose though so i could try? I do have an external programmer i could clear the UEFI flash with or the ME, or of course re-flash it..but all those rails being shorted, i was hoping it would be replacing the actual buck regulators.. Hmm..

Thank you for all your help though, i really feel like crying hah.. Surprised the reading from the GPU is fine though, i mean i guess its good for parts? Maybe i can lift all the working chips and sell them or something stupid? I wouldn't want to do that and rip people off though.. Anyone have a board with something else blown i can steal the PCH from? Haha..

This right to repair thing..manufacturers need to start getting sued. I'll have to look into that, this doesn't seem fair at all. I'm in unexpected debt partly due to "helping" a RL "friend" that...backfired somewhat. I also have a car to sell i don't need to pay back most of the finance as i can't even drive, it was for them! Argh..

Regards,
Lea