The thermal runaway is a semiconductor issue. A self-amplifying problem which can be faced only by proper thermal protection and cooling system. Sometimes the one or the other fail... The CPU and GPU have a thermal shutdown protection implemented. So they do not die on a primary impact. On a primary impact, a MOSFET get killed. Mostly MOSFETs without an implemented over-temperature or over-current protection. Secondly, this will kill then the Dr-MOS and/or the CPU/GPU.
Spot on Sorin, there is something else, I've seen heatsink paste put on the wrong way too many times, too thick an application causes trouble not only with it being sub optimal performance, but it can make it dry out prematurely or even 'pump out' with many heating and cooling cycles, many people see this and mistake it for the paste being too thin, when in fact its because of expansion and contraction during the heating and cooling process, manufacturers do a 'good enough job to last the length of warrantee and a little bit more' to avoid a warrantee repair but end users with a bit of care can do better, the paste is there to fill in imperfections in the mating surfaces only, its not meant to have a thick blob in the middle like a sandwich where the mating surfaces dont actually touch, if you do it right - you wont see temps increase over the course of a few months or even years if you keep your coolers clean, I have not needed to replace mine in my laptop as yet, its still running the factory goop, but temps stay well in the safe zone and the keyboard doesnt get too warm, even though it gets used every day (MSI GE73 i7 8750h/GTX1060 32gb of ram 2x nvme +1 SATA SSD) and had it over a year before the pandemic and it sees some heavy gaming use in Star Citizen, I do use my own fan profileand throttlestop to undervolt and manage the heat the 8750h pushes out when gaming in the summer, in real terms it feels smoother as there is no ramp up or down, it is overdue for a change though but I dont like the current options out there.
Great explication Sorin! I've got a Fujitsu Siemens laptop 17 years ago. It's work now because i always used with a cool pad. The thermal fans never enough for me
i have watch hours of your videos , your shearing of knowledge is a spiritual thing , in my country of new zealand . i would just like to express my thanks ,
We call over-heating something when it heats more than it can acceptably and reliably work within its specification (but this is not entirely and always true, not exactly at least, example given below). It may or may not work slightly near or above the limit but it may get its lifespan reduced. Using the board as a heat spreader, a passive cooler is a decent idea but it works only for very low wattage like 2-4 watts of heat tops and even then it depends on the how much area and how fast it can dissipate the amount of heat generated by x component, especially without making nearby other nearby components either overheat in the process thus have their resistance increased and work less efficiently and/or begin to reduce their lifespan. The actual reason why components fail is because heat makes metals expand, which reduces their ability to hold electrons in a stable state in their outer shells by over-saturating them (with kinetic energy, which itself also generates small amounts magnetic energy as a field, hence increases EMI) leading to destabilizing their structure and so, they break apart causing the component to fail. That kinetic energy has an osculation frequency which, if nearby components (like other nearby mosfets working, oscillating almost in the exact frequency also amplifies this effect by a tiny amount depending on the power of the EMI signal created and distance). A component does not need actually need to overheat for this to happen/fail, it can also happen to temperatures below the limits for the component itself, simply because for example, a soldering pad or ball is not the correct size for the current it has to pass through it especially long enough to handle it so it either melts by a bit, or entirely or completely and as a natural consequence, disconnects the circuit which is connected to. This is why i always say, design things in a way that they can never overheat unless you are doing it on purpose to see if and/or when they will/can fail. Edit: I forgot to mention the amount of electromagnetic noise a short from a broken component can cause to nearby components especially if its a very strong one, like, imagine a mosfet blowing for example. This is why i never like it when PC cases have disk places behind motherboard trays or near they PSU, like, if your disk drives survive a short somewhere else, but they are close to such EMI emission, the disks regardless if they are flash based or magnetic, are still reliable to fail in someway or another and have some or all of your data lost.
I usually put thermal pads everywhere because i don't like the components to get hot. Hot means less life. Redesigning the motherboard with 3d printer and make new Copper heat sink it the best way.
WOW...Sorin has gotta nice clean tidy new desk 😂 Your going UP in the world my friend😉 My 'All Time Favourite Techie'. Btw- Hopefully everyone has seen Sorins 'Back Alley Fault finding techniques & Tricks' If you haven't, your missing out BIG TIME 😉
This is a great video to show why it is so important to not use computers in a very dirty environment, and if you do make sure you clean them often. I have many computers that are more than 10+ years running just fine as I take care of them. With the power draw, it is also important to point out the fact that people should keep their computers connected to power as often as possible and not only rely on battery as the battery will most likely not be able to give all the power for max performance.
A modern CPU will attempt to lower its voltage to frequency curve until it can no longer do so, then it will attempt shutdown, but if an external source keeps heating it or keeping it heated, especially more than it can handle (working or not) within a certain time limit (based on the material's thermal limits), it can still fail regardless, a saturated heatsink can cause that, unless the heat is getting away faster than the max amount of heat the heatsink can dissipate in time before being saturated and before the cooled component fails as well, which is when it needs active cooling. This video illustrates this perfectly.
Great video. Heat on motherboard is energy lost on processor power, so to compensate, more power is pushed to keep processing stable, but that can have bad consequences to moset. I never that of it that way
I have noticed that resistance actually drop with the rise of temperature, at least on some parts of the board. I also noticed this watching different repair videos. For example when shorted capacitor is removed and resistance is measured to check, it can be seen how resistance increasing as the board is cooling down.
From wikipidia; Silicon shows a peculiar profile, in that its electrical resistance increases with temperature up to about 160 °C, then starts decreasing, and drops further when the melting point is reached.
Interesting. I've been dabbling with electronics recently, to help feed a curious mind and attempt to repair a part rather than replace the whole. I find mechanical fault finding etc fairly easy, logical and straight forward. The electrical side is a little less intuitive and ofcourse if is not obvious visual evidence fault finding requires more knowledge than I as a layman possess. This video clarified some things I have been told by techs I've spoken to. Combined with comments from other viewers more knowledgeable than myself, I feel I have gained a decent base level of comprehension. Ofcourse as is most often the case, answering a question poses many more questions. Lol. It is a major task just to access the parts to service them. Equipment being serviced is downtime for production. The ervicing vs replacing becomes hard to justify fiscally. I currently have a pretty decent thermal imaging camera on insefinite loan from a friend. I'm wondering if something like this used at semi regular intervals to check for variations in temp would indicate when work was necessary before something failed. Obviously I'd be viewing the housing/outside of the equipment. Not the components themselves directly. Mostly stainless steel and sheet metal or plastic. I feel like even with accurate imaging the temp differential may not be enough to pre-empt failure?
Adding a thermal pad will not transfer heat very well if the back of the case is plastic. The thermal pad could trap heat if the pad is not sized properly and if there is no way to dissipate the heat that is building up. The thermal compound is to fill the small imperfections between IHS and the cold plate of the cooler. If one has to replace the thermal compound every 3 months then there are serious design issues in that laptop and cooling system.
I would been intreseting to see you heating like an input mosfet while the board is on, and seen that mosfet get shorted. Input mosfet, or cpu mosfet, just with some current limit, so you wouldnt kill it :D
Thermal pads on their own alone do nothing other than, some times become a saturated heatsink essentially "cooking" the component they are attempting to cool in the first place, you need to have something to take the heat off of them before the become saturated or somewhere in between the temperature limit the component can fail regardless because it can't handle the heat anyway. Passive cooling is only good when you can extract the heat passively fast enough something which it depends on the environmental temperature, which, the more it raises, especially getting near the component's temperature, the less the amount of heat will be dissipated/transferred away.
It's simple, semiconductors resistance decreases with the temperature rising. With less resistance devices tend to draw more current and more current means more heat and magic happens.
And people using laptops on a sofa or in bed which can block the air input/output. Better off on a desk, which somewhat depletes the point of a laptop 😀
Very informative video as usual Sorin. Very good advice to use thermal pads to transfer heat to the back of the keyboard or back cover. What was the current limit on the power supply vs the current of the original charger for this laptop? If we use a higher power charger on our laptop would this increase the risk of damaging the mosfets?
Higher current power supply is always better to use. Look how u pick psu for pc, it should be 2x more wattage than your pc use. It apply to every electronics device. My advice is "DONT BUY GAMING LAPTOP" get a PC.
What can also happen is the ABS plastic casing becomes brittle, drying out from the constant heat. The hinge stiffens up, probably from dried up grease, this then goes on to break the brittle hinge mounts, ending up with a broken hinge. The left hinge always breaks first, right next to where the heat is expelled from 99% of laptops. You have to ask the question, why is it designed that way. Bad design or again, planned obsolescence.
I have a question sir? my laptop samsung np535u3c has a light, its power has no display and its processor is not rotating, how can it be fixed sir, I hope you can help me
Guys, he is right but don't do that by yourself, I did that with my Acer Nitro5 AN-515-58-725A and it is not turning on after replacing the thermal paste. I accidentally unplugged the screen before Battery and the laptop shows shorts on the motherboard everywhere, it does not turn on, and I don't have a thermal camera and power supply to find the bad component on the motherboard. I am about to cry because I am a student and I really need my laptop I don't have it anymore since 2 months ago (dead laptop), and I am in a small city in the U.S. with very limited options including replacing the whole motherboard which is very expensive for me :(
Really I enjoyed the explanation for the problem Can I ask you a favour please? Do you have a video for reflowing the VGA in case it is not working properly ( especially in Dell laptops ) I tried to reflow it on 250 degree as many videos said but working for one week then played again So I need your experience if you have the same video just give me a link for it , or tell me what the suitable temperature to use when reflowing that chip Thanks in advance for all your efforts
Reflowing a dead chip helps only temporarily (the heat "fixes" broken things inside). To really fix, you need to REPLACE the component, not to reflow it. See Louis Rossmann video about it.
Sorry Sorin but i can', t agree with you this time... Replacing the thermal paste IS just a russian roulette for the gpus and cpus... The thermal dilatation, flexing the Board with the screwdriver and removing the pressure of heatsink can cause The balls to crack... Loosing audio... Loosing linees on gpus.... Loosing enterely operativity of CPU or PCH... Loosing the PCI express 16x... And a lot of things... Better just to clean fans and heatsinks from dust... The thermal paste can be replaced after 3 years... Just for chemical degradation.... And what about heatsinks/thermalpads on first 2 mosfets??? You saved the Board because the mosfet goes shorted... But first layer of the pcb was damaged... Pressing the mosfet and add thermal Pad can just increase the damage to internal layers making them fuses before The mosfet got killed... I saw lot of Asus x550 with layers merged under The first mosfet... Better to have a shorted mosfet than dig and scratch on layers to remove shorts.
So l have a dell Alienware 17 r4 when l try to power it on it shows orange and white light but no display, if l try more times it will display and go to the desktop what can be the issue
