A great informative video. I had similar issues as you when heating the board to remove a failed MOSFET. The issue is that the board is essentially full of copper traces and these will immediately wick away any heat from the area you've focussed your heat gun at. This is more evident when trying to prepare the bare PCB surface for re-seating the new component, and you'll notice that despite the use of flux and a very hot iron, the solder will not flow as it should. Again because the board is wicking the heat away from the solder and the iron. I decided to purchase a single electric hot plate (around £13-£25 from Amazon) to pre-heat the entire card, as this way you can get the components up to temperature and removed/re-soldered faster and easier. If the board is already hot (not too hot), it can't steal the heat from the area where you're working, you just have to be a little more careful with your working area and any plastic tools around it.
Good job for your first try, to remove them easyer I found reducing your airflow to almost nothing at 500 degrees , the MOSFET comes off very easy give it a try .
to desolder a mosfet properly and quickly you need to use the small nozzle and set the temperature to the highest (450 - 500°C) with a bit of air (3rd level max). 350°C isn't enough as it will take a lot of time to melt the ground solder and it will cause harm to the components near the mosfet! :)
im curious (as a newbie) If the Mosfet was damaged, i presume this is current overload or something, which is heat? So, at what temp does the mosfet start to be damaged? How hot can they get before failing?
@@gepisar presumably, it was slightly off spec to begin with, and the issue simply wasn't strong enough to be detected in testing, maybe it got static damaged after it was factory examined but before it was put on the board, but maybe it was overheated in operation as well, which is why thermal pads shrinking with time and getting unstuck from a surface is a potential issue. Critical die temperature is around 150°C in operation, but ideally you want a lot less for long life, since momentary temperatures during switching count and not just average over a time. Support components are worth a look as well, cracked ceramic capacitors in vicinity and other components designed to tune the behaviour of the circuit, but it's rarely the issue. High temperature causes electromigration and gradually degrades the semiconductor, reducing its properties. On-state resistance will get higher (worse) and switching speed will get slower. The MOSFET emits most of the heat during the transition while it's being switched, and relatively less when it's fully on and zero when fully off. Problem is that if the semiconductor is already not ideal, either from abuse or from manufacturing, it will emit more heat than usual. Then it will have to be run harder. Soon you get a runaway condition where it's running hot because it's damaged, and it's getting damaged because it's hot.
Nice video! I agree, mosfets are particularly difficult to remove with the non-leaded solder used these days. Most times, I have to break the package to get them off. One tip I have for you is to use a soldering iron and a wick to remove the non-leaded solder from the factory and re-tin the pads with some regular leaded solder. It will make reinstalling the new mosfet easier and use less temperature.
I agree. Where you can, add some lead before you start. And as iPadRehab always says "add more flux" :) ...I followed Louis Rossman's advice and acquired some Amtech NC-559-V2-TF "sticky flux", and it's fantastic.
Duuuude, I have a 970 SSC that has coffee damage, starting testing like your procedure today, with the way cards are these days I'd like to save it so as not to buy a new one!! Your info has me on the right path I think, thanks!
Finally! a technician which explains everything straight to the point with simple words, unlike others which they try to sound like scientists. You earned my sub keep up the good work!
If you want to listen to a technician you should go watch Northridge Fix. She said she wasn't good.... And it shows actually. Using a heat gun around capacitors (those cylindrical components) is asking for a DISASTER. Those components blow up when there's too much heat!!! They should be protected with some sort of shield so there isn't heat transfer... And has she said... She had to apply a lot of heat to the mosfet. Which is VERY close to the capacitors!!! It's LUCKY she didn't got injured! SAFETY people!!! And bro... She's limited on equipment... No voltage injection and heat camera?!? Yeah... She ain't going far with this
Can't wait for the next issue! I have been waiting for someone to go step by step and let us see what you are doing. I pulled all the mosfet from several boards I purchased on ebay and I am trying to fix them! I have checked all the mosfets pulled and they are good. Had I seen this video before I could have left the mosfets on the board and waited for you to do the next step. Thanks and keep doing videos for those of us who don't know how to go through the steps for testing all the pieces, which I hope you will do. Thanks again and stay safe and well.
How unfortunate ☹️ I really hate when components are as stubborn as that MOSFET were. I've purchased a preheater to use underneath the PCB for when working on such multilayer boards as I've heard the heat can cause the boards to bend and warp from being applied from one side only. I've seen videos of preheaters in use making the removal much easier as the area you are working on is closer to the temp needed and so you don't require such a long dwell time with the hot air tool on the top of the component. I do hope you don't give up on the board and manage to find the culprit and repair it. Good luck!
Thank you very much, I'll do my best! You have a good point there I never thought about the possibility for the board to bent. I've never seen one or had the issue but definately something to consider. A few days ago I made a new video in which I try using a ceramic IR heating element for soldering and well (spoiler: it does work) but since then I was playing with these old, broken graphics cards from last video and used the heating element to preheat the pcb to around 180°C which worked so much better than I ever thought so there will also be a video in which I show the possibility for it to be used for preheating pcb's.
I had to repair my GTX 1080 Ti Waterforce, it had blown mosfet, it infact had melted it self into the groundplane but i managed to save the card by using dremel and drilling most of it out to isolate it. In the end it worked but is running with 11 mosfets instead of stock 12 smosfets. Sold it later to friend and to date its still working perfect on my friends pc, i repaired it about 3 years ago and sold it about 2 years ago.
Hi, Anna, where do you get Evga graphics data sheets from and what do they look like and cover? Looked on Evga web site and nothing there for my Evga e-Geforce 9600 GT (512MB) graphics card. Also, my brand new LAP multimeter does not beep when I set it to continuity and touch the probes together. I get readings on 5 of the six pins in the power port of the card, but no beep on the odd one out. Does this mean there is a short? Love the way you progress very methodically and clearly through the fault-finding process and the colourful cuticles. Have subscribed. Thanks.
The reason for hard desoldering of shorted high current transistors is that some times the drain welds it self to the PCB, if that is the case I would use a dremel with a stedy hand under a microscope in order not to damage the pcb or other components with excesive heat. Also, before this method try by adding flux as many times as needed and use a 10 mm nozzle with ~400 C. I use 480 C every day on automotive ECU's (FET's, SMD micro-controllers) BUT NOT on BGA like these GPU processors. 40% leaded solder doesn't help at all with hot air desoldering,it helps only with direct contact soldering station for a higher contact surface of heat transfer and a proper shaped soldering tip. Also you should have measured the gate trace resistance before put in a new FET (it is allways a good practice in order not to work for nothing), measure and open the FET with a multimeter to check if it's body diode is shorted and if the FET is opening (do not forget that when you open the FET by multimeter, drain-source will act as shorted because of the internal capacitances, you must close the FET by shorting drain to source before conclusions). Chears!
Keep in mind that the passive core resistance can be between 0,5-10ohm depending on model and the other vrm rail like for memory in the 10-300ohm again depending on model. So you using the beep function that does that from 100ohm can mislead. The vrm resistance will indicate the gpu health. And it can be shorted from vcc to core og gnd to vcc on the core. And the switch controller can be faulty as well. It does take time and even broken cards that you break more to know and lean what and how the card works. The 12v voltage rail from pcie slot, should be around the 3-4kohm and the pll or plex ind the 150-500ohm depending on model. I have a r9 290X that have 14ohm on the 12v pcie rail and it should be the 3.5kohm because a shorted pll core in the gpu it should be 30-40 ohm.
Where is the 2nd part of this video? I like the way you explained everything. I want to learn more about graphic card repairing. Can you please help me or teach me how to properly trace different faults.
I just answered someone who had the same question. So basically I had to wait for the mosfet drivers to arrive just to discover the card had another problem I first had to figure out about and then I also got sick at the beginning of this year. So everything got delayed but I'm going to make another two parts. One with a lot of talking so you can follow along the process I had to go through to figure everything out and the hopefully last video to fix it once and for all.
Good evening, greetings from Argentina, excellent video, in my case (I have the same GPU) the mosfets that mark a short, it's Q13, Q21 and Q22... the whole phase would be, so I think the problem comes from another place... or not?
These ones are feisty, they sit directly on gargantuan copper pours in inner layers, which wick away all the heat! This isn't the case with the driver chip and its friends, they're barely connected thermally to the copper mass. It may be wise to preheat the board to run ahead of the soak and protect the components you don't want to hurt with some alu tape. The current handling capability isn't the only difference. For each branch of the VRM, you have one high-side FET and two low-side or free-wheeling FETs in parallel. The high side FET passes a high voltage and low current, its main operational requirement is fast switching and low gate capacity, they are also on for only a very short burst of the cycle. The low side FETs pass a low voltage and extremely high current, they need low on-state resistance, they are on for most of the cycle. For this reason, low- and high-side FETs ideally shouldn't be switched around, the specified component should be used. The layout of this board is a little tricky to see which one belongs to which branch and side.
afteryou remove the part , put some flux on and use solder wick to clean the pads , then use a solder iron and some tin to retin the pads. This way there is no contamination or remainder of the old part (especially if it broke during removal)
I strongly suggest you check the vcore and vmem inductor resistance everytime before plugging this thing into your pc - you don't want 12v of PD flowing into the GPU core as it'll kill it instantly.
Very good video, I want to change the mosfet of a 1070 sc evga and I would like to know which one you recommend that I buy because I do not know which model they are.. Thank you
Awesome Video thank you. Was able to spot a broken Mosfet on my PCB. But do you know where to buy the Mosfets separately and how I can find out what type of mosfets are on the PCB? I am trying to fix a R9 390 8GB
neat thanks. will wait for the next part. i dont really have any idea about gfx card repairing, but seems they have a typical failure rate of dying at every 4-5 years.
Pre heat the board a bit before going on the chip so it doesnt soak the heat as much. I would also protect the electrolitic caps because they explode under high heat. Oh and do use flux! :D
Do you know the specs of the ceramic capacitor on the other side of the board? C848. It's on the opposite side of the Q13 mosfet in this video. The upper left mosfet on the board. I could also use the specs for that mosfet. Mine are 4C06N - RCC4R. I found 4c06n mosfets but I don't know the amp rating I need. Will be looking for your other video next.
my 970 had a scare the other day, just died and the fan would not start, it's a zotac blower, already stripped, cleaned and it's working now though it needs a fan replacement, thankfully, to avoid using an uncooled card in the beginning of summer, my brother is more into console and had a spare 1080! cannot say enough just how reliable a zotac 970 is, years of use and it's only the fan that goes! roll on 2 weeks when my rtx3060ti arrives!:)
I'm trying to be brave and tackle a faulty mosfet on my Evga 980Ti. As far a finding the new part, is Alibaba the best place? Any other channels? I live in Toronto area.
This helped me fix my GPU twice, The one next to your bad one was bad for me, Worked for a while and then the same one in your video on my card went bad too. Replaced that and its working again. I wonder if it would be worth it to replace them all.
Thanks for watching! I'm glad the video helped you fix your GPU. In your case as another mosfet died I would say it is definately worth replacing all of them. It seems like they got hit hard when the first one went out.
Hello there Do you know how to check a PWM controller with a multimeter? Because I suspect it causing a short on the Mosfets, I removed a Mosfet and measured it alone there's not short. The PWM model is NCP81174 for a ZOTAC GTX 970.
Everyone is an expert here of course but a couple of tips. Cover the caps with a heat sink before heating as they tend to explode with too much thermal abuse. Pre-heat the board prior to removal and use good quality flux and wick to remove old solder and replace with leaded solder before replacing the component.
Hey thanks for your comment! Preheating is something I do ever since I had problems in this video, as for flux I now have different types for different applications. At the time I made this video I still had a supply of leaded solder but it ran out and everything I can get my hands on is awful quality. So I mostly use lead free solder nowadays. I was aware of the risk with overheating capacitors but never considered using a heatsink on them. I don't know if it is that necessary but still thankful you mentioned it. Even if I'm not going to do it others might read your comment and decide it is the right joice for them. (edit: I had caps blowing up on me in my early days of the hobby but has been years since it happened last)
Enjoyed your video alot, you pretty smart :) and can explain stuff so good to, sick of people not being more forward on what this and that does, but you i say good job :) keep up the great work, i know i learned something new from your knowledge on electronics
One of the mosfets on the far right side I believe is burnt as it has a white spot that isn’t from thermal pads on my 970. Would my card be repairable?
Generally, if a part in an electronic blows, there's something else wrong. Unless you fix said other issue, replacing the part isn't going to fix it. First guess says run lower power through it with a thermal camera and check other parts of the board for shorts and excessive hear. It sounds like you were doing a bit of voltage prodding as well, which can also be a good indicator.
Hi Anna, my gtx 970 "kinda died" recently. It provides image and sometimes I can still play, but it eventually gives me a black or glitched screen and I have to reset the PC. Does this problem look familiar to you?
Where would you go after finding a shorted gate on a mosfet if you removed it and the mosfet was good? How would you verify that the mosfet driver is go or not? What is before the driver? Please do more of these if you can. Thanks again.
I have the exact same board, when it went bad the computer turned off and would not turn on. I put in another lessor graphics card, and the computer ran normally so I know that the GTX 970 is bad. There is no visible damage. I would like to repair this board as replacing it would be expensive. I am a retired technician, I used to do this kind of work for a living. Having a schematic would help, but those are likely proprietary.
Hi Anna, I would like to ask to you, How if we found source mosfet not connected. I found it in my gpu that already no display for few month, But fan keep running loud 100% directly. Do i just to jumper it? Or any else to searching? Thank you for help anna, Ikrna
for mosfets with these type of heat stations use 4 to 5 on air flow, and heat to 400c thats enoug plus use a nice flux dont forget tho.. you can cover the caps with aluminum so they dont get tosted, dont pull them hard till the solder its liquid, dont be afraid its ok in the way to learn to make some mistakes and most of the times you need to replace all mosfet and driver of the line other way they will fail again in short time or allmost instantly
Seu vídeo foi muito bom, mas acredito que nesta data de hoje você tenha melhorado suas técnicas de solda. Você judiou muito da placa, mesmo assim, parabéns!
I love the Nails! Also I am currently using a Radeon 3000 IGP that microsoft decided to no longer supprt, thanks, Microsoft, you wankers. Anyway I have a 5770 and 4780(I think) somewhere in the house and thought I'd have a look. So thank you for puting this together, free knowledge is exceptionally underrated these days
Hi, great tutorial, but i don´t undestand. You said that the Lower rated mosfet was broke, wich is the 46amp (4C10N), but you finally installed the higher rated mosfet of 69amp (4C06N), maybe that's the problem?
This was not the problem here but I get where you come from. These two mosfets are the exact same, except build for higher/ lower current than the other one is. Switching behavior is important and they are, as I stated, the exact same in every regard but current rating. However you might have been right if they were from two different *generations* (lacking a better word here) and if I didn't check the switching behavior. You may want to watch my newest video to understand why it didn't work the first time around and what other problems I ran into :)
Yeah, love the nails also! You explain this very well, you should think about a career in teaching, plus you are honest and do not edit out the messy stuff.
You are very thorough and amazing with your handi work I have a card that is giving me black screen turns on with fan on full blast lights but no screen any ideas I have checked the mosfets they are all good I can't find anything
Hi, nice video. I learned a lot from watching it: thanks very much. I'm like you, have a different foundation (mine is materials engineering not physics as yours is), and am using these sorts of things to learn about electronics. Couple comments: - At start, you might say what's wrong with the device: I'm assuming it is just dead, but you need to watch more to see that. Some GPU's have erractic/intermittent issues. - Also at start, you did say it likely wasn't worth a great deal of effort to fix, but I think is good to say what it's initial cost was, and current 2nd hand value (just to show how that judgement was made) - Maybe comment on this GPU's nature: the basic ones start by being powered by the PCIe connector, but after 75Watts, it needs to draw from the auxilliary inputs. And this one is fairly high powered at xx Watts. - In terms of troubleshooting, I wonder if you fixed the two shorts you found (by simply disconnecting that part), does the GPU work at a low level? Drawing low power? Can test that by a functional GTX 970 (there are many of them), and seeing if it will work without the aux power lead. Maybe not. In hindsight, for the failed Mosfet I wonder if you might just cut it out (say with a Dremmel tool): which wouldn't heat the board up at all. Then add a new one over top of it. Doesn't need to look pretty, or be flush. I didn't see the Mosfet type, but looking at the data sheet, you can see how much heat you are allowed to use to install it. I suspect it's close to impossible to remove it without a VERY good desoldering station. Especially without cooking the neighbouring ones. And maybe is good idea to shield the others whilst trying to remove the failed one. I'm bad at doing SMT's......barbeque time...:-} I see this guy managed it: th-cam.com/video/CtK1O-UoLMM/w-d-xo.html And he replaced the driver too: couldn't get same so tried what he felt was similar. And it worked. Anyways, nice video: thanks.
Hey, thanks for your comment and welcome to my channel! :) Now the video is for everyone interested to watch, people that now about computers I don't have to tell them what a GTX 970 is worth and for others I thought that it is just unnecessary information. Same goes for the information that a GPU like that wont even startup without pci-e power connected. It literally tells you in startup, black screen, white letters: pci-e connector missing I wont start. That's at least the gist of it. As for the damage to the gpu, well without diagnosing it I can't really tell you. 49 seconds into the video I said that I suspected a mosfet to be the problem but I checked for everyone to see and follow if they try to diagnose a GPU problem like that. At the end of the video I said that the problem is not fixed and I have to do more to fix it. Since then I did a lot more on the GPU and found numerous problems with it. This thing is not a simple fix at least not as simple as I thought. I need to make the next part and there I'll talk about everything I did and give some information on how GPU power delivery works for a better understanding. When you work on stuff like that you don't need a good hot air rework station, just the one I have is good enough. Now obviously better ones make your life easier but that was not the problem here. The mosfet simply got partially welded to the pad underneath which can happen if there is not enough solder underneath (meaning layer thickness) and the component gets hot for an extended amount of time. Others were moving long before this one did. And with shielding you mean use kapton tape to protect other components? That is something I had a comment about before which I did not answer to. If you mean something else ignore the following and tell me! :) "It is supposed to protect other components from heat" is what you hear all the time. Kapton tape is a heat resistant, non electrically conductive, self-adhesive material that has many uses. But it won't "protect" any component from heat. Lets just say we have kapton around one of these mosfets, the hotair does not directly hit the surrounding components but the air heats up the kapton tape. It is not an perfect insulator, no perfect insulator exists and it is so thin that it will still transfer a significant portion of the heat energy towards components underneath. So no protection there. This makes it even worse considering that the mosfet or component you are working on is electrically connected to other components and for that also thermally connected. When you try to cool SMD fets like these one trick we engineers do is not only to attach a heatsink to the component which is less efficient as if they were a tht fet but also connect a heatsink to the inductor as it is electrically connected to the mosfet and has very good thermal transfer. What you see looking on top of the mosfet is a relatively thick layer of plastic like material so more heat energy can be transferred through the pcb. Thats one reason why we also cool inductors. To make matters even worse is that in order to desolder a compoenent you have to saturate the pcb/ traces electrically connected with the component you are working on in order to reach necessary temperature for soldering. It literally works as a heatsink. So blocking the heat is exactly what you shouln't do. Just go on youtube and watch professionals (I'm not one of them) doing board level repair, none uses kapton tape this way, it makes no sense. Again it has its uses and I got plenty of it but that's not something it should be used for. However you are absolutely right that components do have diagrams showing the temperature curve and if you are concerned about damaging anything taking a look at it doesn't hurt but this is more for reflow ovens and not hot air soldering as you can set the hot air temperature to whatever and the distance to your work piece decides actual soldering temperature. Has a lot to do with experience, just do a bit of smd work and after a short time you get the hang of it. It is really easy. My hot air station is set between 300°C and 350°C although my solder melts at 186°C if I remember correctly but heat is only part of the "problem", time is the other. So soldering with 300°C a mosfet can be quickly soldered, maybe 5 to 15 seconds. Depends on current pcb temperature. Since this video I started to preheat the pcb so it doesn't take as long to saturate the pcb with heat energy, also I use lower melting point solder than the one the components were originally soldered in place with. Again thank you very much for your comment, I really appreciate the feedback. :)
Ola, eu sou do Brasil e parabéns pelas explicações ... Só faltou mostrar os valores do multímetro nas medições da entrada de energia. Eu tenho uma duvida na minha placa ... o pino neutro central de cada entrada 2x6 está apresentando resistência, nao fecha em zero e nao bipa. Todos os videos que vi até agora a condutividade fecha no neutro central, a minha nao.
Any suggestions for desoldering the mosfet? I’ve been using a heat gun, a soldering iron and flux but I can’t get the bottom of the mosfet to come off (it broke in two pieces)
to much heat for to much time. Try preheat area and than blast mosfet with higher temp for shorter time, but firstly try higher airflow, ramp at max your airflow, flux. Train properly
i have amd rx 480 8gb gddr5 when i connect 8 pin of smps to the graphic card the 2 fans rotates but when i connect graphic card to the slot on the motherboard fan does not rotates and graphic card is not detected by the computer. Is my graphic card faulty?
Hey, welcome to my channel! :) You have to make sure that these mosfets are the exact ones you need or have similar properties to the ones your GPU requires. They need to match in terms of switching current, Rds(on) value and switching properties. As for the voltage rating, any mosfet that can at least handle the 12V input plus some spare should do the job. I got mine from ebay (Link: www.ebay.de/itm/20-x-NTMFS4C06NT1G-4C06N-NTMFS4C06N-NTMFS4C06-QFN8-Power-MOSFET-Single-N-Channel/183967962550?ssPageName=STRK%3AMEBIDX%3AIT&_trksid=p2060353.m2749.l2649 )
@@PhysicistAnna Hi, my display card also died recently and I am trying to fix it. It is so sad that you have failed. I suspect that the new mosfet may not be an real one. I check the ebay link above, the price of the mosfet is too cheap. It may be fake. I suggest you check the new mosfet by multi-meter. Make sure the pins is behave as the real mosfet. If the pin don't work as normal, it must be fake.
Hi quick question are you able to tell what size capacitors those are on the actual GPU chip around the dye...on the green part between the silver part and the actual GPU dye...and secondly will the resistance be different in all? One broke off my 2080 super and I need to replace it but not sure of the value. I'm being told it's a 01005 0402 capacitor and a 1uf should suffice? They come in various voltages can you please advise. The lowest I could find was a 2.5v as you mentioned chip uses very low voltage...would this be it uk.farnell.com/murata/grm022d80e104me15l/cap-0-1uf-2-5v-mlcc-01005/dp/2997876
I was cheering for you because I have a 960 in the same condition and I was hoping to be able to fix it, although I'm afraid I wouldn't be able to replace SMD components so close together without doing damage T_T...
There were two different mosfets used. I however used the higher rated 4c06n for the repair. The other lower rated one was the 4c10n if I'm not mistaken.
Replacing the upper two mosfet drivers got me into windows. It seems as if the lower two are not working as I ran into some more problems especially under heavy load. Will replace them in video part 2 and hopefully that solves the issue once and for all. If you are interested in how I got there, here is part 1.5 in which I tell you my journey so far in terms of diagnosing the issue. th-cam.com/video/jXlV53GZS18/w-d-xo.html
Great tutorial Ana, you explained better in 17 m then others in 50 minutes .
Thanks and welcome to my channel :)
exactly
yeah write, i learned how not to do it, tnx for that
Yeah well explained 👍🔥
I would agree. I watch a lot of these. I must break a lot of sh*t
The initial steps to find the fault with just a multimeter are VERY useful and clearly explained. This video is a gem. Thank you!
A great informative video.
I had similar issues as you when heating the board to remove a failed MOSFET. The issue is that the board is essentially full of copper traces and these will immediately wick away any heat from the area you've focussed your heat gun at. This is more evident when trying to prepare the bare PCB surface for re-seating the new component, and you'll notice that despite the use of flux and a very hot iron, the solder will not flow as it should. Again because the board is wicking the heat away from the solder and the iron.
I decided to purchase a single electric hot plate (around £13-£25 from Amazon) to pre-heat the entire card, as this way you can get the components up to temperature and removed/re-soldered faster and easier. If the board is already hot (not too hot), it can't steal the heat from the area where you're working, you just have to be a little more careful with your working area and any plastic tools around it.
Good job for your first try, to remove them easyer I found reducing your airflow to almost nothing at 500 degrees , the MOSFET comes off very easy give it a try .
to desolder a mosfet properly and quickly you need to use the small nozzle and set the temperature to the highest (450 - 500°C) with a bit of air (3rd level max).
350°C isn't enough as it will take a lot of time to melt the ground solder and it will cause harm to the components near the mosfet! :)
im curious (as a newbie) If the Mosfet was damaged, i presume this is current overload or something, which is heat? So, at what temp does the mosfet start to be damaged? How hot can they get before failing?
@@gepisar If you look at the mosfet data sheet, it will usually tell you a package max temperature, just assume above that bad things start happening.
Для выпаивания мосфетов лучше использовать нижний подогрев . Иначе легко повредить текстолит , перегреть электролитический кондесатор .
@@gepisar presumably, it was slightly off spec to begin with, and the issue simply wasn't strong enough to be detected in testing, maybe it got static damaged after it was factory examined but before it was put on the board, but maybe it was overheated in operation as well, which is why thermal pads shrinking with time and getting unstuck from a surface is a potential issue. Critical die temperature is around 150°C in operation, but ideally you want a lot less for long life, since momentary temperatures during switching count and not just average over a time. Support components are worth a look as well, cracked ceramic capacitors in vicinity and other components designed to tune the behaviour of the circuit, but it's rarely the issue.
High temperature causes electromigration and gradually degrades the semiconductor, reducing its properties. On-state resistance will get higher (worse) and switching speed will get slower. The MOSFET emits most of the heat during the transition while it's being switched, and relatively less when it's fully on and zero when fully off.
Problem is that if the semiconductor is already not ideal, either from abuse or from manufacturing, it will emit more heat than usual. Then it will have to be run harder. Soon you get a runaway condition where it's running hot because it's damaged, and it's getting damaged because it's hot.
@@MrAlxalx So heat it up from the other side of the card?
Nice video! I agree, mosfets are particularly difficult to remove with the non-leaded solder used these days. Most times, I have to break the package to get them off. One tip I have for you is to use a soldering iron and a wick to remove the non-leaded solder from the factory and re-tin the pads with some regular leaded solder. It will make reinstalling the new mosfet easier and use less temperature.
I agree. Where you can, add some lead before you start. And as iPadRehab always says "add more flux" :) ...I followed Louis Rossman's advice and acquired some Amtech NC-559-V2-TF "sticky flux", and it's fantastic.
GPU repair+ ASMR?? thank you youtube algorithm gods for bringing me here
Duuuude, I have a 970 SSC that has coffee damage, starting testing like your procedure today, with the way cards are these days I'd like to save it so as not to buy a new one!! Your info has me on the right path I think, thanks!
Finally! a technician which explains everything straight to the point with simple words, unlike others which they try to sound like scientists. You earned my sub keep up the good work!
If you want to listen to a technician you should go watch Northridge Fix. She said she wasn't good.... And it shows actually. Using a heat gun around capacitors (those cylindrical components) is asking for a DISASTER. Those components blow up when there's too much heat!!! They should be protected with some sort of shield so there isn't heat transfer... And has she said... She had to apply a lot of heat to the mosfet. Which is VERY close to the capacitors!!! It's LUCKY she didn't got injured! SAFETY people!!!
And bro... She's limited on equipment... No voltage injection and heat camera?!? Yeah... She ain't going far with this
I really enjoyed your explanations, camera angles and steadiness, as well as your technical knowledge and accuracy.
Looking forward to part 2!
Great vid, the more we learn to repair these things, the less E-Waste the world has to put up with.
Damn she's like the female version of EliTech XD
Great explanation BTW!
Can't wait for the next issue! I have been waiting for someone to go step by step and let us see what you are doing. I pulled all the mosfet from several boards I purchased on ebay and I am trying to fix them! I have checked all the mosfets pulled and they are good. Had I seen this video before I could have left the mosfets on the board and waited for you to do the next step. Thanks and keep doing videos for those of us who don't know how to go through the steps for testing all the pieces, which I hope you will do. Thanks again and stay safe and well.
excellent . The sound and pronunciation of the words are excellent.
How unfortunate ☹️ I really hate when components are as stubborn as that MOSFET were. I've purchased a preheater to use underneath the PCB for when working on such multilayer boards as I've heard the heat can cause the boards to bend and warp from being applied from one side only. I've seen videos of preheaters in use making the removal much easier as the area you are working on is closer to the temp needed and so you don't require such a long dwell time with the hot air tool on the top of the component.
I do hope you don't give up on the board and manage to find the culprit and repair it. Good luck!
Thank you very much, I'll do my best!
You have a good point there I never thought about the possibility for the board to bent. I've never seen one or had the issue but definately something to consider. A few days ago I made a new video in which I try using a ceramic IR heating element for soldering and well (spoiler: it does work) but since then I was playing with these old, broken graphics cards from last video and used the heating element to preheat the pcb to around 180°C which worked so much better than I ever thought so there will also be a video in which I show the possibility for it to be used for preheating pcb's.
I had to repair my GTX 1080 Ti Waterforce, it had blown mosfet, it infact had melted it self into the groundplane but i managed to save the card by using dremel and drilling most of it out to isolate it. In the end it worked but is running with 11 mosfets instead of stock 12 smosfets. Sold it later to friend and to date its still working perfect on my friends pc, i repaired it about 3 years ago and sold it about 2 years ago.
Hi, Anna, where do you get Evga graphics data sheets from and what do they look like and cover? Looked on Evga web site and nothing there for my Evga e-Geforce 9600 GT (512MB) graphics card. Also, my brand new LAP multimeter does not beep when I set it to continuity and touch the probes together. I get readings on 5 of the six pins in the power port of the card, but no beep on the odd one out. Does this mean there is a short? Love the way you progress very methodically and clearly through the fault-finding process and the colourful cuticles. Have subscribed. Thanks.
My inno3d 1050ti doesn't look like this can you please explain mosfet or power stage which should be replaced?
The reason for hard desoldering of shorted high current transistors is that some times the drain welds it self to the PCB, if that is the case I would use a dremel with a stedy hand under a microscope in order not to damage the pcb or other components with excesive heat. Also, before this method try by adding flux as many times as needed and use a 10 mm nozzle with ~400 C. I use 480 C every day on automotive ECU's (FET's, SMD micro-controllers) BUT NOT on BGA like these GPU processors. 40% leaded solder doesn't help at all with hot air desoldering,it helps only with direct contact soldering station for a higher contact surface of heat transfer and a proper shaped soldering tip.
Also you should have measured the gate trace resistance before put in a new FET (it is allways a good practice in order not to work for nothing), measure and open the FET with a multimeter to check if it's body diode is shorted and if the FET is opening (do not forget that when you open the FET by multimeter, drain-source will act as shorted because of the internal capacitances, you must close the FET by shorting drain to source before conclusions).
Chears!
Girl you're amazing...saved my gpu...make.more videos on amd rx570 and 580 i have 3 bad ones
Keep in mind that the passive core resistance can be between 0,5-10ohm depending on model and the other vrm rail like for memory in the 10-300ohm again depending on model. So you using the beep function that does that from 100ohm can mislead. The vrm resistance will indicate the gpu health. And it can be shorted from vcc to core og gnd to vcc on the core. And the switch controller can be faulty as well. It does take time and even broken cards that you break more to know and lean what and how the card works. The 12v voltage rail from pcie slot, should be around the 3-4kohm and the pll or plex ind the 150-500ohm depending on model. I have a r9 290X that have 14ohm on the 12v pcie rail and it should be the 3.5kohm because a shorted pll core in the gpu it should be 30-40 ohm.
Where is the 2nd part of this video? I like the way you explained everything. I want to learn more about graphic card repairing. Can you please help me or teach me how to properly trace different faults.
I just answered someone who had the same question.
So basically I had to wait for the mosfet drivers to arrive just to discover the card had another problem I first had to figure out about and then I also got sick at the beginning of this year. So everything got delayed but I'm going to make another two parts. One with a lot of talking so you can follow along the process I had to go through to figure everything out and the hopefully last video to fix it once and for all.
@@PhysicistAnna I also like the way you explained the whole situation 😂 so excited to watch next parts of this video. Thanks
Good evening, greetings from Argentina, excellent video, in my case (I have the same GPU) the mosfets that mark a short, it's Q13, Q21 and Q22... the whole phase would be, so I think the problem comes from another place... or not?
These ones are feisty, they sit directly on gargantuan copper pours in inner layers, which wick away all the heat! This isn't the case with the driver chip and its friends, they're barely connected thermally to the copper mass. It may be wise to preheat the board to run ahead of the soak and protect the components you don't want to hurt with some alu tape.
The current handling capability isn't the only difference. For each branch of the VRM, you have one high-side FET and two low-side or free-wheeling FETs in parallel. The high side FET passes a high voltage and low current, its main operational requirement is fast switching and low gate capacity, they are also on for only a very short burst of the cycle. The low side FETs pass a low voltage and extremely high current, they need low on-state resistance, they are on for most of the cycle. For this reason, low- and high-side FETs ideally shouldn't be switched around, the specified component should be used. The layout of this board is a little tricky to see which one belongs to which branch and side.
afteryou remove the part , put some flux on and use solder wick to clean the pads , then use a solder iron and some tin to retin the pads.
This way there is no contamination or remainder of the old part (especially if it broke during removal)
I strongly suggest you check the vcore and vmem inductor resistance everytime before plugging this thing into your pc - you don't want 12v of PD flowing into the GPU core as it'll kill it instantly.
Very good video, I want to change the mosfet of a 1070 sc evga and I would like to know which one you recommend that I buy because I do not know which model they are.. Thank you
where did you order the mosfets ?
Awesome Video thank you. Was able to spot a broken Mosfet on my PCB. But do you know where to buy the Mosfets separately and how I can find out what type of mosfets are on the PCB? I am trying to fix a R9 390 8GB
Before using heatgun, you must cover other components with thermal insulator tape.
Then, the desoldering will be safe.
neat thanks. will wait for the next part. i dont really have any idea about gfx card repairing, but seems they have a typical failure rate of dying at every 4-5 years.
Does this guide of yours also apply to board-powered only cards like some 1050 Tis?
Pre heat the board a bit before going on the chip so it doesnt soak the heat as much. I would also protect the electrolitic caps because they explode under high heat. Oh and do use flux! :D
Amazing work
Do you know the specs of the ceramic capacitor on the other side of the board? C848. It's on the opposite side of the Q13 mosfet in this video. The upper left mosfet on the board. I could also use the specs for that mosfet. Mine are 4C06N - RCC4R. I found 4c06n mosfets but I don't know the amp rating I need. Will be looking for your other video next.
my 970 had a scare the other day, just died and the fan would not start, it's a zotac blower, already stripped, cleaned and it's working now though it needs a fan replacement, thankfully, to avoid using an uncooled card in the beginning of summer, my brother is more into console and had a spare 1080! cannot say enough just how reliable a zotac 970 is, years of use and it's only the fan that goes! roll on 2 weeks when my rtx3060ti arrives!:)
I'm trying to be brave and tackle a faulty mosfet on my Evga 980Ti. As far a finding the new part, is Alibaba the best place? Any other channels? I live in Toronto area.
Brilliant👏🏻👏🏻👏🏻👏🏻👏🏻Thank you for an incredible explanation of repairing video card👍🏻👍🏻👍🏻👍🏻
You are extremely intelligent, your video is really well made. Very informative, no bullshit.
This helped me fix my GPU twice, The one next to your bad one was bad for me, Worked for a while and then the same one in your video on my card went bad too. Replaced that and its working again. I wonder if it would be worth it to replace them all.
Thanks for watching! I'm glad the video helped you fix your GPU.
In your case as another mosfet died I would say it is definately worth replacing all of them. It seems like they got hit hard when the first one went out.
Hello there
Do you know how to check a PWM controller with a multimeter? Because I suspect it causing a short on the Mosfets, I removed a Mosfet and measured it alone there's not short. The PWM model is NCP81174 for a ZOTAC GTX 970.
Everyone is an expert here of course but a couple of tips. Cover the caps with a heat sink before heating as they tend to explode with too much thermal abuse. Pre-heat the board prior to removal and use good quality flux and wick to remove old solder and replace with leaded solder before replacing the component.
Hey thanks for your comment!
Preheating is something I do ever since I had problems in this video, as for flux I now have different types for different applications. At the time I made this video I still had a supply of leaded solder but it ran out and everything I can get my hands on is awful quality. So I mostly use lead free solder nowadays.
I was aware of the risk with overheating capacitors but never considered using a heatsink on them. I don't know if it is that necessary but still thankful you mentioned it. Even if I'm not going to do it others might read your comment and decide it is the right joice for them.
(edit: I had caps blowing up on me in my early days of the hobby but has been years since it happened last)
Enjoyed your video alot, you pretty smart :) and can explain stuff so good to, sick of people not being more forward on what this and that does, but you i say good job :) keep up the great work, i know i learned something new from your knowledge on electronics
Thank you very much
One of the mosfets on the far right side I believe is burnt as it has a white spot that isn’t from thermal pads on my 970. Would my card be repairable?
Generally, if a part in an electronic blows, there's something else wrong. Unless you fix said other issue, replacing the part isn't going to fix it. First guess says run lower power through it with a thermal camera and check other parts of the board for shorts and excessive hear. It sounds like you were doing a bit of voltage prodding as well, which can also be a good indicator.
Hi Anna, my gtx 970 "kinda died" recently. It provides image and sometimes I can still play, but it eventually gives me a black or glitched screen and I have to reset the PC.
Does this problem look familiar to you?
Where would you go after finding a shorted gate on a mosfet if you removed it and the mosfet was good? How would you verify that the mosfet driver is go or not? What is before the driver? Please do more of these if you can. Thanks again.
kapton polyimide tape helps keep the surrounding micro components from moving. Nice vid
I have the exact same board, when it went bad the computer turned off and would not turn on. I put in another lessor graphics card, and the computer ran normally so I know that the GTX 970 is bad. There is no visible damage. I would like to repair this board as replacing it would be expensive. I am a retired technician, I used to do this kind of work for a living. Having a schematic would help, but those are likely proprietary.
I just fix the same card with the same problem was kind of tricky
Hi Anna,
I would like to ask to you,
How if we found source mosfet not connected.
I found it in my gpu that already no display for few month,
But fan keep running loud 100% directly.
Do i just to jumper it? Or any else to searching?
Thank you for help anna,
Ikrna
Hi Anna what type of Multimeter are you using?
this video helped me out alot! i have the exact same problem and have already ordered my replacement mosfets. thanks
for mosfets with these type of heat stations use 4 to 5 on air flow, and heat to 400c thats enoug plus use a nice flux dont forget tho.. you can cover the caps with aluminum so they dont get tosted, dont pull them hard till the solder its liquid, dont be afraid its ok in the way to learn to make some mistakes and most of the times you need to replace all mosfet and driver of the line other way they will fail again in short time or allmost instantly
Seu vídeo foi muito bom, mas acredito que nesta data de hoje você tenha melhorado suas técnicas de solda. Você judiou muito da placa, mesmo assim, parabéns!
Which of these Mosfets can I use
MTB09N03H8
PH7030AL
PH2530AL
Instead of this
MDU1513
I love the Nails! Also I am currently using a Radeon 3000 IGP that microsoft decided to no longer supprt, thanks, Microsoft, you wankers. Anyway I have a 5770 and 4780(I think) somewhere in the house and thought I'd have a look. So thank you for puting this together, free knowledge is exceptionally underrated these days
Hi, great tutorial, but i don´t undestand. You said that the Lower rated mosfet was broke, wich is the 46amp (4C10N), but you finally installed the higher rated mosfet of 69amp (4C06N), maybe that's the problem?
This was not the problem here but I get where you come from. These two mosfets are the exact same, except build for higher/ lower current than the other one is. Switching behavior is important and they are, as I stated, the exact same in every regard but current rating.
However you might have been right if they were from two different *generations* (lacking a better word here) and if I didn't check the switching behavior. You may want to watch my newest video to understand why it didn't work the first time around and what other problems I ran into :)
The mosfets are different because one of them is a high side mosfet and 2 are low side mosfet. The one that broke I think is the highside mosfet.
Where did you bought new mosfet from
i dunno what to say , just u are AWESOME , i hope i can learn something about fixing things like that
Me-
"damn, that mosfet really is not wanting to move, I wonder if HOLY moly her nails are awesome!"
Yeah, love the nails also! You explain this very well, you should think about a career in teaching, plus you are honest and do not edit out the messy stuff.
You are very thorough and amazing with your handi work I have a card that is giving me black screen turns on with fan on full blast lights but no screen any ideas I have checked the mosfets they are all good I can't find anything
Dead temp sensors maybe ? 🤷🏻
@@mattmin45 I was wondering how and where the temp sens is
se te agradece compartir la experiencia
You are grateful for sharing the experience
Hi, nice video.
I learned a lot from watching it: thanks very much.
I'm like you, have a different foundation (mine is materials engineering not physics as yours is), and am using these sorts of things to learn about electronics.
Couple comments:
- At start, you might say what's wrong with the device: I'm assuming it is just dead, but you need to watch more to see that.
Some GPU's have erractic/intermittent issues.
- Also at start, you did say it likely wasn't worth a great deal of effort to fix, but I think is good to say what it's initial cost was, and current 2nd hand value (just to show how that judgement was made)
- Maybe comment on this GPU's nature: the basic ones start by being powered by the PCIe connector, but after 75Watts, it needs to draw from the auxilliary inputs.
And this one is fairly high powered at xx Watts.
- In terms of troubleshooting, I wonder if you fixed the two shorts you found (by simply disconnecting that part), does the GPU work at a low level? Drawing low power?
Can test that by a functional GTX 970 (there are many of them), and seeing if it will work without the aux power lead. Maybe not.
In hindsight, for the failed Mosfet I wonder if you might just cut it out (say with a Dremmel tool): which wouldn't heat the board up at all.
Then add a new one over top of it.
Doesn't need to look pretty, or be flush.
I didn't see the Mosfet type, but looking at the data sheet, you can see how much heat you are allowed to use to install it.
I suspect it's close to impossible to remove it without a VERY good desoldering station.
Especially without cooking the neighbouring ones.
And maybe is good idea to shield the others whilst trying to remove the failed one.
I'm bad at doing SMT's......barbeque time...:-}
I see this guy managed it: th-cam.com/video/CtK1O-UoLMM/w-d-xo.html
And he replaced the driver too: couldn't get same so tried what he felt was similar.
And it worked.
Anyways, nice video: thanks.
Hey, thanks for your comment and welcome to my channel! :)
Now the video is for everyone interested to watch, people that now about computers I don't have to tell them what a GTX 970 is worth and for others I thought that it is just unnecessary information. Same goes for the information that a GPU like that wont even startup without pci-e power connected. It literally tells you in startup, black screen, white letters: pci-e connector missing I wont start. That's at least the gist of it.
As for the damage to the gpu, well without diagnosing it I can't really tell you. 49 seconds into the video I said that I suspected a mosfet to be the problem but I checked for everyone to see and follow if they try to diagnose a GPU problem like that.
At the end of the video I said that the problem is not fixed and I have to do more to fix it. Since then I did a lot more on the GPU and found numerous problems with it. This thing is not a simple fix at least not as simple as I thought. I need to make the next part and there I'll talk about everything I did and give some information on how GPU power delivery works for a better understanding.
When you work on stuff like that you don't need a good hot air rework station, just the one I have is good enough. Now obviously better ones make your life easier but that was not the problem here. The mosfet simply got partially welded to the pad underneath which can happen if there is not enough solder underneath (meaning layer thickness) and the component gets hot for an extended amount of time. Others were moving long before this one did.
And with shielding you mean use kapton tape to protect other components? That is something I had a comment about before which I did not answer to. If you mean something else ignore the following and tell me! :)
"It is supposed to protect other components from heat" is what you hear all the time. Kapton tape is a heat resistant, non electrically conductive, self-adhesive material that has many uses. But it won't "protect" any component from heat. Lets just say we have kapton around one of these mosfets, the hotair does not directly hit the surrounding components but the air heats up the kapton tape. It is not an perfect insulator, no perfect insulator exists and it is so thin that it will still transfer a significant portion of the heat energy towards components underneath. So no protection there.
This makes it even worse considering that the mosfet or component you are working on is electrically connected to other components and for that also thermally connected. When you try to cool SMD fets like these one trick we engineers do is not only to attach a heatsink to the component which is less efficient as if they were a tht fet but also connect a heatsink to the inductor as it is electrically connected to the mosfet and has very good thermal transfer. What you see looking on top of the mosfet is a relatively thick layer of plastic like material so more heat energy can be transferred through the pcb. Thats one reason why we also cool inductors.
To make matters even worse is that in order to desolder a compoenent you have to saturate the pcb/ traces electrically connected with the component you are working on in order to reach necessary temperature for soldering. It literally works as a heatsink. So blocking the heat is exactly what you shouln't do.
Just go on youtube and watch professionals (I'm not one of them) doing board level repair, none uses kapton tape this way, it makes no sense.
Again it has its uses and I got plenty of it but that's not something it should be used for.
However you are absolutely right that components do have diagrams showing the temperature curve and if you are concerned about damaging anything taking a look at it doesn't hurt but this is more for reflow ovens and not hot air soldering as you can set the hot air temperature to whatever and the distance to your work piece decides actual soldering temperature. Has a lot to do with experience, just do a bit of smd work and after a short time you get the hang of it. It is really easy.
My hot air station is set between 300°C and 350°C although my solder melts at 186°C if I remember correctly but heat is only part of the "problem", time is the other. So soldering with 300°C a mosfet can be quickly soldered, maybe 5 to 15 seconds. Depends on current pcb temperature. Since this video I started to preheat the pcb so it doesn't take as long to saturate the pcb with heat energy, also I use lower melting point solder than the one the components were originally soldered in place with.
Again thank you very much for your comment, I really appreciate the feedback. :)
@@PhysicistAnna true, the idea of using the tape is to simply keep parts in place , like that resistor and cap that moved on you :P
Put some leaded solder paste around the mosfet! It will melt and mix with the lead free solder making removal easy!
That's a great idea, thanks! :D Will definitely try the next time I have this problem :)
Great tutorial. Hoping for more gpu repair like this.
Great video, keep up the great work, waiting for the next part.
Hey, thanks and welcome to my channel!😊
@@PhysicistAnna do you repair GPU for fan??🤔
Ola, eu sou do Brasil e parabéns pelas explicações ... Só faltou mostrar os valores do multímetro nas medições da entrada de energia.
Eu tenho uma duvida na minha placa ... o pino neutro central de cada entrada 2x6 está apresentando resistência, nao fecha em zero e nao bipa. Todos os videos que vi até agora a condutividade fecha no neutro central, a minha nao.
cara e que tal escrever em ingles ou usar o google? Ela e alema, fala ingles e tu vem em portugues?
Any suggestions for desoldering the mosfet? I’ve been using a heat gun, a soldering iron and flux but I can’t get the bottom of the mosfet to come off (it broke in two pieces)
to much heat for to much time. Try preheat area and than blast mosfet with higher temp for shorter time, but firstly try higher airflow, ramp at max your airflow, flux. Train properly
i have amd rx 480 8gb gddr5 when i connect 8 pin of smps to the graphic card the 2 fans rotates but when i connect graphic card to the slot on the motherboard fan does not rotates and graphic card is not detected by the computer. Is my graphic card faulty?
Explained down to the multy meter where to set it. Thank you , I always wanted to know that. Continuity,
where did you get your MOSFETs from? I'm trying to get into GPU fixing and the one I have is this exact same problem?
Hey, welcome to my channel! :)
You have to make sure that these mosfets are the exact ones you need or have similar properties to the ones your GPU requires. They need to match in terms of switching current, Rds(on) value and switching properties. As for the voltage rating, any mosfet that can at least handle the 12V input plus some spare should do the job.
I got mine from ebay (Link: www.ebay.de/itm/20-x-NTMFS4C06NT1G-4C06N-NTMFS4C06N-NTMFS4C06-QFN8-Power-MOSFET-Single-N-Channel/183967962550?ssPageName=STRK%3AMEBIDX%3AIT&_trksid=p2060353.m2749.l2649 )
@@PhysicistAnna Hi, my display card also died recently and I am trying to fix it.
It is so sad that you have failed. I suspect that the new mosfet may not be an real one.
I check the ebay link above, the price of the mosfet is too cheap. It may be fake.
I suggest you check the new mosfet by multi-meter. Make sure the pins is behave as the real mosfet.
If the pin don't work as normal, it must be fake.
Hi Anna, I have the same problem with my graphics card, did the card work after this repair, I'm waiting for the mosfet part
Hello, thanks ffor the vid, could you give a link for the parts please i am needing capacitor for my gpu...
Hi quick question are you able to tell what size capacitors those are on the actual GPU chip around the dye...on the green part between the silver part and the actual GPU dye...and secondly will the resistance be different in all? One broke off my 2080 super and I need to replace it but not sure of the value. I'm being told it's a 01005 0402 capacitor and a 1uf should suffice? They come in various voltages can you please advise. The lowest I could find was a 2.5v as you mentioned chip uses very low voltage...would this be it uk.farnell.com/murata/grm022d80e104me15l/cap-0-1uf-2-5v-mlcc-01005/dp/2997876
Thank you very much! Your video is great!
Where did you find the datasheet?
Where can I find the data sheets to map out components
I was cheering for you because I have a 960 in the same condition and I was hoping to be able to fix it, although I'm afraid I wouldn't be able to replace SMD components so close together without doing damage T_T...
subscribed, thank you so much i am going to try repairing some gpu, very very detailed and clear explanation , appreciate your video
Same thing happened to me with the mosfets, it even ripped the copper pad off the PCB.
If you use a preheater under the board, it might be easier
Thanks to show how do you works
Hi Anna, can you remember the name of the mosfet? best regards!
There were two different mosfets used. I however used the higher rated 4c06n for the repair. The other lower rated one was the 4c10n if I'm not mistaken.
best explanation ever
Very very usefull & well explained.
hola pregunta vi este tutorial y todo me salio bien pero el mosfet driver esta malo donde lo puedo comprar me ayudas gracias
Should all the mosfets have the same resistance. Most are 3-4. One is 10, one is 75.
Ajudou muito. Desde já agradeço de vdd. Sorte minha que o vídeo tem legendas para português Brasil se não eu estava lascado.
fantastic video. Learned a lot, thank you.
A clear explanation.
you done a great job young lady !!
Very much appreciated thank you
1:56 mine shorted in the same spots too. bad mosfets
Replace the (1R0) inductor chip then replace the mosfet . that should do it. good stuff
Replacing the upper two mosfet drivers got me into windows. It seems as if the lower two are not working as I ran into some more problems especially under heavy load. Will replace them in video part 2 and hopefully that solves the issue once and for all. If you are interested in how I got there, here is part 1.5 in which I tell you my journey so far in terms of diagnosing the issue. th-cam.com/video/jXlV53GZS18/w-d-xo.html
What temp do heat-guns need to be to be able to remove components?
350-380C low air flow.
you should have used captain tape around the heating
why ?
From where can I buy the mosfets