Mate I don't know why you worry that the video is going on too long. You could do an hour one and I would watch every minute of it. As always great repair. I also repaired my first RX 470 the other day because of your video's. Thanks
Everybody is always struggling very hard to even diagnose a GPU problem, and Tech Cemetery does it as a child's play, like in every occupation there are bad, normal and good technicians, and there are extraordinary persons like you, is unbelievable your skills.
Yeah, but listen to his voice, he orates like a small child talking to the class in school, monotone, no flow, chainlinked monoclauses, he has no rubato or flexing of pitch which means he's never been around people. Pity him.
I bought a 2080 six months ago with the intent of eventually fixing it. It arrived and then I lost the courage to break out the heat gun and having a shot at it. I'm only getting around to it now and this is something like the fourth or fifth video I've watched related to a rtx20xx repair that I've watched and I think I'm going to rip the band-aid off and just do it.
Weird that those contact pads on those GDDR6 memories would just come off like that and the 2080ti is only just a few years old too all while my over a decade old GT440 is still running like a champ.
I recently stumbled upon your channel and I'm loving the content. Would you maybe consider making a guide on possible faults behind common gpu issues? I have electronics engineering background and I'd like to try to purchase some dead gpus and try repairing them and it would be useful to know what kind of issues are most promising. So far it appears to me that the easiest ones to fix would be the ones that have no fan spin and fail to boot as it would indicate a problem with supply lines (provided the core isn't dead ofc) unlike the artifacting ones which likely need a memory or bios chip replacement. What about cards that do have fan spin but no image? Could this also be supply related or most likely other issue?
Been enjoying your video’s. Always intrigued by pcb repairs like these but very hard. Did some reflow of GPU & laptops which worked surprisingly good in my 5 or so samples but feels like a blunt tool for a fine task. Do you think you could give more explanation about the practical side of things ? Software / databases you use with the schematics of the cards? How do you find them? If you get a card (undetected) what is your general process? From checking shorts to power rail and ‘powergood’ signals. Where do you get the knowledge what signals are important & where to measure them? Then also perhaps you could link to or make video’s focused on what physical skills you’ve acquired or need for these kinds of repairs (smd soldering .. etc) If this is/were on your patreon, I’d pay !
Ur videos are best. This is what I was looking for. Keep uploading your content. I have the same issue with my Msi rx 580 8gb OC GDDR5 lines on screen code 43. I will definitely try it.U r doing great. I loved it 😊
I really appreciate that you do more problems deviating from the usual "no power present". BTW - did you use the same memory module for replacement or was it a different Samsung model number. If that's the case, what do you need to look for in oder to determine whether they are compatible?
I used a higher binned variation of the same model. The old memory was a Samsung memory module rated for 14gbps, the one I used was rated for 16gbps. They have the same model number but the 14gbps memory ends in "HC-14" and the one I used ends in "HC-16"
I once heard GPUs are timing sensitive and at least in case of regular ram the timings of higher clocked memory are worse. Does this affect performance?
@@standartblitz6798 The higher clocked RAM has just more worse timing if it is clocked on higher rates.The reason is, the "timings" are simple counter values, used as "wait state" until "something" happens. In the end these values are just time periods. So if you clock on double the rate the timing counters have to be double the value to be finaly the same (!) time period. So if you calculate certain memory timings in its real time represaton in nanoseconds you will notice the time is always the same, no matter what timings are given in relation to the clock rate. This is because the chip don't need X clock cycles, they need X nanoseconds. The values are just picked to get to this timespans (as far as this is possible due of the fact you can't wait 3.1234 clock cycles).
Can we attribute this to a mix between weak PCB design and stress on the PCI Express area, or it's just a really bad GDDR6 batch? Fantastic work, as always.
Have suffered through this issue with memory modules failing one after the other. I will get back to that 2080 ti n my shelf now with a new vision of it. Once again great video. Thank you.
Hey great content here, thanks! So you we're saying that the pads had lifted and were showing as brown. Were you referring to the pads on the memory or the vc pcb? Because you would need to replace the pads if on the pcb right? Do you think they just got too hot and balled the solder? Is the brown rust? Thanks!
I believe it could be shortcut to diagnose of you diodemode all pads before putting a new ddr chip back on. Such damage happens really easilly if user had bumped PC with GPU card installed without a raiser cable. Also changing memory clock might help in revealing unstable banks.
Yo Tech. Your are a MAD MAN!! Your skills are off the charts . I only wish I could do this as good as you. Quick question on memory layout. Using MATS are the memory modules labeled the same? Or do you always go anti clock wise when identifying the bad memory module?
@@TechCemetery Ahhh I did watch that. Thats how I learned to set up MATS. I didnt pay attention to the GPU configuration when you explained. Many thanks. Learning so much from your videos. Been watching them every evening for the past 2 weeks.
What are the most common memmory modules in graphics cards, i know that E6932/E6930 mosfets are often used in gpus, but i don't know which memory ICs to buy for my repair shop, thx for the education, keep it up.
If I have a Card with a dead GDDR5 module, can I use any GDDR5 chip from any old card or is there something to look out for? Whai if i accidentally put a wrong IC? Can I replace Micron chip with a Samsung chip?
I have a legion y540 with an rtx2060 that just had artifacts recently and error 43, I wanna get it fixed but how do I know if the gpu itself is dead or it's a vram problem? Great video btw!
And here I am stressing over fixing a 970 which I burnt when using a mollex adapter, the pc didn't turn on, so I stupidly plugged it in only when it was on. Fans don't work, nothing on screen. 😰
Ihave 2070 MQ (mobile) with code 43 appeared after i repaste my laptop CPU and GPU with Liquid metal, could it be cause of the frying GPU? The GPUZ detect the graphic card but the BIOS version still unknown as seen on the GPUZ
Hey bro! Your video helpe me to understand my problem the only thing is that i have the AMD 6800XT and mats doesnt work with my video card! Any other tool? And if you find a faulty memory you have to change the bank or you can reflow it! My card i thing is saged
I have a problem with this model. When I start any game, the game crash instantly and exit. No error code. Can I know the correct thermal pad thickness for this model? I can't find any info on the net. I will test the vram with nvidia mats and mods after replacing the thermal paste and pad.
If ya'll can possibly rip this turbine cooler off and put a three fan cooler on it this card would be helped immensely. It would require a lot of adaption to properly do it with hole drilling and tapping. Probably should be done by a machinist.
They won't last at all, from what I've already seen, the general reliability of Ampere and GDDR6X in particular is atrocious, and the mining craze isn't doing Ampere any favours.
I've got a 2080ti that's artifacting but only when I put it in 4k and start to push the card hard. Is it something you think you could fix? I'm sure I could do it but I'm also sure most of the tools would cost more then buying a new card lol. It works 1440p at as high FPS as I can get it to run on ultra settings. It is only at 4k but it's quite heavy when It starts. No settings changes help only going back to 1440p
I have a 1080 evga ftw with artifacts, but when I run mats it doesn’t have any bad memories. The card post and I can go to windows, but when it tries to install driver the card crash with some vertical green lines and the video turn black.
Hi, so i have a Asus RTX 2080 Ti Turbo that i want to repair and i notice two components missing, they are ripped of the card but looks like pads are still there. They are i think resistors and i would have like to purschase new ones. Do you have a reference and a website to find some ? I don't know if i buy the same size they will work (i know nothing about those components that's why i'm asking). If you can help well i would be verry happy. Thanks
Please Does anyone know the thicknesses for thermal pads for this specific video card. I plan on buying one and replacing them as well as the thermal paste but I don't know what thickness to buy. Thanks in advance
Love the videos mate, any quick thoughts on a RX 570 that won't display when the AMD drivers are installed but will do with only the generic windows driver?
Is it safe to say that the damage is caused by GPU sag together with high thermal output (because of the blower style cooler), which both degrade the pcb and solder pads strength? And if so, could it be avoided or at least postponed by adding a sag bracket and/or changing the cooler with a better one. I have a GTX 1080, originally with a blower cooler, swapped with an EVGA ACX cooler (also flashed the bios, so it is now recognised and working as an EVGA GTX 1080 SC) and the temps dropped by 15 degrees Celsius, clocked higher than stock and also the backplate added rigidity to the pcb itself.
Yep, I believe it's safe to say the damage is caused by a combination of high temperatures and GPU sag. I also think that the damage could have been mitigated or even avoided by using either a GPU sag bracket or placing the card in a riser.
i bought a faulty 780 ti for the sake of fixing it and learning new things about GPU manufacturing I had a faulty buk so i came up with the idea of replace it buuut with different model so my card runs on sk Hynix and i solid an elpida chip from a gtx 970 and surprisingly it post but still artifacting I'm getting about a 400k write errors on both of the new chips that I solird and having 32 faulty bits so i think ineed helpp from you or at least i need guidance
My brother in law give me a RISC 51 (MSI) RTX 2060 super from aliexpress and two days later it starts to show pink horizontal artifacts all over the place. Read it on internet that it's probably in VRAM failure. Yeah, my first gpu is now dead. Now i will have to take it to the technician and pay some extras. This is so sad.
When you say you replaced the memory module, do you mean you physically replaced it with a new module? Or you just took it off to clean it/redo the solder
I have a 3070, with artifacts. Get error 43 in device manager, and 0mb ram is showing up in gpu-z. These are samsung ram, checked with multimeter. Got 22ohms. 1.37volt when turning on the card. Mats/Mods won't work with 3000 series, do you have any suggestions?
Hello brother I need your help I have a gpu p104-100 I cannot run the memory test it tells me that the card is not being recognized and the window gives me the same error 43 it recognizes it and everything but when installing the drivers it gives error 43 in Sometimes the administrator recognizes it well and it can be used but when the machine is turned off or restarted, the code 43 returns so in this gpu I use it to play although it does not have video output, window lets use it with the output of the integrated board If you can help me I would appreciate it very much
I think my board has this exact same problem, I am not a technician or anything, and I live in a small town, so I'd like to know how would I go about finding this memory banks you replaced, how much do them cost and how do I buy them and such, can anybody help me with anything?
Im having and issue with MSI 1080 gaming X , i have replaced one of the memory modules since mats told it was bad and the card pass the test now no problem. But the problem remains and it is that when the driver is installed it instantly crashes and the PC restarts if u can help me with advice ill be grateful. BTW i changed d9txs with d9vrl memory module it is Micron memory.
Try running MODS test 94 with the memory frequency maxed. MATS test runs at minimum memory speed, so if a module is borderline it will still pass MATS. Throw a scope on your power stages too, make sure your VRMs are hitting on all cylinders.
@Kids Korner guess why nvidia/amd has such a extensive QA test software even though chips are supposed to QA before the assembly?;) I mean, there are pogopin testers, capable of checking chips before soldering them on board. There are industry best soldering machines. And yet, hige software suite to test if ram shit itself after it was soldered :)
hi! I have exactly one rtx 2080ti that has a bad memory module (faulty solder) and I watch your video, at minute 7:50 you say you may need to solder the module in the empty slot, so does that mean that socket can be used? in my case for example?
@@TechCemetery well. All hope Is gone for that card 🥴 i was about to buy the neccesary stuff to solder one memory in that module leaving the damaged socket empty hoping to work. Well thank you for the answer :)
Do you do repairs my zotac rtx 2080ti randomly crashed on me while gaming and when into low resolution mode nothing changes even when I install new drivers when I check device manager it's says driver 43 error any help would be appreciated as all my repair shops keep.turning me down I am in UK but willing to ship also it is in low res on my pc but no display in other systems
Is there an AMD version of the MATS tool? I'm assuming an nVidia version will not work for an AMD card. I have an R9 270 that is artifacting due to bad memory modules, and figured this would be a good starting point
Hi guys, what is that program that shows the memory modules problems in this video? I have an RTX 2080 Gigabyte with the same green lines. The only difference that it is only at dark areas, the black areas where it shows the green lines. And it is in low resolution, windows inmediately stop the device because it causes problems (error code 43), so i can still do this test with that program?
@@nitto1983 Mats and mods is the name of the program. At other videos of this guy there is a link i think for the program, and he explains it well how to use it.
My MSI gaming x trio 2080ti have exactly the same problem like in this video, RMA it , 2 week later received the same old card and work for 2 week and have the same problem again. Now i have to RMA it again and wait for another 2 week, and who know how long it gonna work this time?. MSI RMA service suck big time
Don't soldier and don't use heat gun it is firmware issue of gpu flash the firmware first upgrade then downgrade your gpu will be fine if not the only solution is to reball
Hey there, Love your work for the community. I have the exact same issue (visually) and now would like to run a memory test.. If I squeeze the card the right way, or move the water cooling rubber pipes to push it up on the right angle it works np.. But for these times, it would be swell to do a mem test for each of these 11gb. Lets say, I have the software or rights to use for a day to help diagnose my card. For this particular card would you mind being a dear and pasting exactly what you have in your file to check this card? My card is also a RTX 2080 Ti with Samsung memory & TU102-300A-K1-A1 gpu .. Thanks man!
I have an Asus laptop which Is giving me code 43 error and motherboard has Elpida ram on it. Is it possible to change from Elpida to Samsung as I have a scrap board with same density chips in Samsung brand. Any help would be appreciated. Not very much info on the net as to whether this can be done or not
@@TechCemetery thanks so much! That's awesome news! Wanted to check with a pro before I go ahead and do it. Very skilled with microsoldering but not knowledgeable when it comes to ram
@@TechCemetery 960m Nvidia. I am wondering if it is the gpu as I replaced a shorted out mosfet on one of the gpu main rails so wonder if the gpu got kicked in the pants maybe. Also am not managing to get mats to run on laptops at all. Is there a trick to doing this? Have all the correct bios settings too.
i have a gpu thats having similar problem, can you do repair if i ship it to u ? i will love to send mine in and have it fix and do a video on it for the channel
Mate I don't know why you worry that the video is going on too long. You could do an hour one and I would watch every minute of it. As always great repair. I also repaired my first RX 470 the other day because of your video's. Thanks
What was the problem of your rx470? Did you fix it? I have one, with error 43.
@@Ayongis I have a video of it on my channel. Not an error 43 tho. Dead mosfets.
Everybody is always struggling very hard to even diagnose a GPU problem, and Tech Cemetery does it as a child's play, like in every occupation there are bad, normal and good technicians, and there are extraordinary persons like you, is unbelievable your skills.
God mode allways active 😂
2 weeks is quite a hard case :)
Yeah, but listen to his voice, he orates like a small child talking to the class in school, monotone, no flow, chainlinked monoclauses, he has no rubato or flexing of pitch which means he's never been around people. Pity him.
The effort that goes into every video is superb !
Its a shame that broken rtx20xx series cards are so costly, I'd love to work on more of them.
I bought a 2080 six months ago with the intent of eventually fixing it. It arrived and then I lost the courage to break out the heat gun and having a shot at it. I'm only getting around to it now and this is something like the fourth or fifth video I've watched related to a rtx20xx repair that I've watched and I think I'm going to rip the band-aid off and just do it.
@@mikes78 Well start by checking if its Micron Memory, becuase if its Samsung the approach is different
You sir are a talented young man. Keep up the good work.
How the hell can you solder that. Seems impossible job to me, insane! Great stuff!
Weird that those contact pads on those GDDR6 memories would just come off like that and the 2080ti is only just a few years old too all while my over a decade old GT440 is still running like a champ.
I recently stumbled upon your channel and I'm loving the content. Would you maybe consider making a guide on possible faults behind common gpu issues? I have electronics engineering background and I'd like to try to purchase some dead gpus and try repairing them and it would be useful to know what kind of issues are most promising. So far it appears to me that the easiest ones to fix would be the ones that have no fan spin and fail to boot as it would indicate a problem with supply lines (provided the core isn't dead ofc) unlike the artifacting ones which likely need a memory or bios chip replacement. What about cards that do have fan spin but no image? Could this also be supply related or most likely other issue?
Been enjoying your video’s. Always intrigued by pcb repairs like these but very hard. Did some reflow of GPU & laptops which worked surprisingly good in my 5 or so samples but feels like a blunt tool for a fine task.
Do you think you could give more explanation about the practical side of things ? Software / databases you use with the schematics of the cards? How do you find them?
If you get a card (undetected) what is your general process? From checking shorts to power rail and ‘powergood’ signals.
Where do you get the knowledge what signals are important & where to measure them?
Then also perhaps you could link to or make video’s focused on what physical skills you’ve acquired or need for these kinds of repairs (smd soldering .. etc)
If this is/were on your patreon, I’d pay !
please make a video on how to reball memory chips including blower temperature air flow , that will help a lot others
yay! a new video! cant wait to resume watching this at home
superb work - makes you wonder if desktop cases should make a comeback to prevent this happening to GFX cards...
Wow. Liked and subscribed! Please do videos only on soldering and replacing chips, memory, smd etc. You're the best!!!
Iv just found out about your channel. Great video. I like how you explain and show in great detail. Thanks
Ur videos are best. This is what I was looking for. Keep uploading your content. I have the same issue with my Msi rx 580 8gb OC GDDR5 lines on screen code 43. I will definitely try it.U r doing great. I loved it 😊
hey, another great video.. Im the guy that sent you my 10900k for possible repair, cant wait to see a vid of that if you make one!!
I really appreciate that you do more problems deviating from the usual "no power present".
BTW - did you use the same memory module for replacement or was it a different Samsung model number. If that's the case, what do you need to look for in oder to determine whether they are compatible?
I used a higher binned variation of the same model. The old memory was a Samsung memory module rated for 14gbps, the one I used was rated for 16gbps. They have the same model number but the 14gbps memory ends in "HC-14" and the one I used ends in "HC-16"
I once heard GPUs are timing sensitive and at least in case of regular ram the timings of higher clocked memory are worse. Does this affect performance?
As far as I can tell, no.
Ok, thanks for the info :)
@@standartblitz6798 The higher clocked RAM has just more worse timing if it is clocked on higher rates.The reason is, the "timings" are simple counter values, used as "wait state" until "something" happens. In the end these values are just time periods. So if you clock on double the rate the timing counters have to be double the value to be finaly the same (!) time period. So if you calculate certain memory timings in its real time represaton in nanoseconds you will notice the time is always the same, no matter what timings are given in relation to the clock rate. This is because the chip don't need X clock cycles, they need X nanoseconds. The values are just picked to get to this timespans (as far as this is possible due of the fact you can't wait 3.1234 clock cycles).
Can we attribute this to a mix between weak PCB design and stress on the PCI Express area, or it's just a really bad GDDR6 batch? Fantastic work, as always.
I would attribute it to GPU sag and stress around the PCIe bracket.
TIL that physical trauma is the fault of the memory modules' "batch". SMH.
From our experience Samsung chips have a higher tendency of ripped pads than Micron.
Glad I found your channel. Totally worth more than 10k subs.
another amazing video, and just after work too! nice.
Great job as always!
Can you recommend site for download or buy schematic or boardview for Graphic cards?
Have suffered through this issue with memory modules failing one after the other. I will get back to that 2080 ti n my shelf now with a new vision of it. Once again great video. Thank you.
Thank you for such quality content
Hey great content here, thanks!
So you we're saying that the pads had lifted and were showing as brown. Were you referring to the pads on the memory or the vc pcb? Because you would need to replace the pads if on the pcb right? Do you think they just got too hot and balled the solder? Is the brown rust? Thanks!
I believe it could be shortcut to diagnose of you diodemode all pads before putting a new ddr chip back on. Such damage happens really easilly if user had bumped PC with GPU card installed without a raiser cable. Also changing memory clock might help in revealing unstable banks.
I love your videos! Any chance you have like a Q and A about what got you started doing this?
Subscribed! Awsome content :) Entertaining and educational also ;)
If your card has these symptoms, but passes mats, what does that mean?
Hi I have been having the same issues. Did you manage to solve it?
The Tech Cemetery is fraught with empty graves, for it is home to the Techromancer!
Yo Tech. Your are a MAD MAN!! Your skills are off the charts . I only wish I could do this as good as you. Quick question on memory layout. Using MATS are the memory modules labeled the same? Or do you always go anti clock wise when identifying the bad memory module?
Counter clockwise starting from the bottom right of the GPU. You should see my MATs guide video.
@@TechCemetery Ahhh I did watch that. Thats how I learned to set up MATS. I didnt pay attention to the GPU configuration when you explained. Many thanks. Learning so much from your videos. Been watching them every evening for the past 2 weeks.
What are the most common memmory modules in graphics cards, i know that E6932/E6930 mosfets are often used in gpus, but i don't know which memory ICs to buy for my repair shop, thx for the education, keep it up.
If I have a Card with a dead GDDR5 module, can I use any GDDR5 chip from any old card or is there something to look out for?
Whai if i accidentally put a wrong IC?
Can I replace Micron chip with a Samsung chip?
Where can you buy spare parts for these cards?
great work man
Brilliant Repair!
as always great as it can be. keep up.
You are really good. Thumbs up!
Wonder where you can find replacement parts for gpu
I have a legion y540 with an rtx2060 that just had artifacts recently and error 43, I wanna get it fixed but how do I know if the gpu itself is dead or it's a vram problem?
Great video btw!
And here I am stressing over fixing a 970 which I burnt when using a mollex adapter, the pc didn't turn on, so I stupidly plugged it in only when it was on.
Fans don't work, nothing on screen. 😰
I would love to send my card in for repair, I just asked for a quote it's a very similar situation as this one
Ihave 2070 MQ (mobile) with code 43 appeared after i repaste my laptop CPU and GPU with Liquid metal, could it be cause of the frying GPU? The GPUZ detect the graphic card but the BIOS version still unknown as seen on the GPUZ
Hey bro! Your video helpe me to understand my problem the only thing is that i have the AMD 6800XT and mats doesnt work with my video card! Any other tool? And if you find a faulty memory you have to change the bank or you can reflow it! My card i thing is saged
I have a problem with this model. When I start any game, the game crash instantly and exit. No error code. Can I know the correct thermal pad thickness for this model? I can't find any info on the net. I will test the vram with nvidia mats and mods after replacing the thermal paste and pad.
good job mate.
If ya'll can possibly rip this turbine cooler off and put a three fan cooler on it this card would be helped immensely. It would require a lot of adaption to properly do it with hole drilling and tapping. Probably should be done by a machinist.
Is the damage to the modules a result of high thermals? Those blower cards sure do like to get toasty under sustained loads.
In part yes, but I do believe it's also partially as a result of the GPU sagging heavily around the PCIe slot.
@@TechCemetery Interesting, I wonder how Ampere cards will last considering how massive some of those cards are.
They won't last at all, from what I've already seen, the general reliability of Ampere and GDDR6X in particular is atrocious, and the mining craze isn't doing Ampere any favours.
Do you think the damage to the Vmem is a result of not enough heat applied when removing said Vmemory?
Not at all
@@TechCemetery It's just that I've seen some techs use 600watt stations on 10 layer boards and damage the solder points.
I've got a 2080ti that's artifacting but only when I put it in 4k and start to push the card hard.
Is it something you think you could fix? I'm sure I could do it but I'm also sure most of the tools would cost more then buying a new card lol.
It works 1440p at as high FPS as I can get it to run on ultra settings. It is only at 4k but it's quite heavy when It starts.
No settings changes help only going back to 1440p
I have a 1080 evga ftw with artifacts, but when I run mats it doesn’t have any bad memories. The card post and I can go to windows, but when it tries to install driver the card crash with some vertical green lines and the video turn black.
man the memory module and the memory bga area really gives the trypophobia vibes
Hi, so i have a Asus RTX 2080 Ti Turbo that i want to repair and i notice two components missing, they are ripped of the card but looks like pads are still there. They are i think resistors and i would have like to purschase new ones. Do you have a reference and a website to find some ? I don't know if i buy the same size they will work (i know nothing about those components that's why i'm asking). If you can help well i would be verry happy. Thanks
Please Does anyone know the thicknesses for thermal pads for this specific video card. I plan on buying one and replacing them as well as the thermal paste but I don't know what thickness to buy. Thanks in advance
Love the videos mate, any quick thoughts on a RX 570 that won't display when the AMD drivers are installed but will do with only the generic windows driver?
Memory or BIOS
@@TechCemetery Appreciate the reply, thanks and keep up the great videos!
Where did you buy the new Memory module for the GPU??
Hi! Where do you get your spare parts from?
Did u reball the chips or did u just Flux it and put it together?
Is it safe to say that the damage is caused by GPU sag together with high thermal output (because of the blower style cooler), which both degrade the pcb and solder pads strength? And if so, could it be avoided or at least postponed by adding a sag bracket and/or changing the cooler with a better one. I have a GTX 1080, originally with a blower cooler, swapped with an EVGA ACX cooler (also flashed the bios, so it is now recognised and working as an EVGA GTX 1080 SC) and the temps dropped by 15 degrees Celsius, clocked higher than stock and also the backplate added rigidity to the pcb itself.
Yep, I believe it's safe to say the damage is caused by a combination of high temperatures and GPU sag. I also think that the damage could have been mitigated or even avoided by using either a GPU sag bracket or placing the card in a riser.
i bought a faulty 780 ti for the sake of fixing it and learning new things about GPU manufacturing
I had a faulty buk so i came up with the idea of replace it buuut with different model
so my card runs on sk Hynix and i solid an elpida chip from a gtx 970
and surprisingly it post
but still artifacting
I'm getting about a 400k write errors on both of the new chips that I solird
and having 32 faulty bits
so i think ineed helpp from you or at least i need guidance
My brother in law give me a RISC 51 (MSI) RTX 2060 super from aliexpress and two days later it starts to show pink horizontal artifacts all over the place. Read it on internet that it's probably in VRAM failure.
Yeah, my first gpu is now dead.
Now i will have to take it to the technician and pay some extras.
This is so sad.
have the same issue with my rog strix 2080 ti
hello bud im having trouble with the 1070 zotac mini do you no the location of the enable for the hex as i have no enable thanks mate
Always a good and nice job...
Thank you!
hello. you are good to soldering and replacing BGA chips. for when a video where you show us how to properly replace this chips...bga chips etc...??
I'll get to it eventually.
When you say you replaced the memory module, do you mean you physically replaced it with a new module? Or you just took it off to clean it/redo the solder
New module
@@TechCemetery where do you get your parts from? Off of other cards or from a third party source?
What temperature and fan speed did you use to remove the module?
788*F and 100 for air.
I have a 3070, with artifacts. Get error 43 in device manager, and 0mb ram is showing up in gpu-z. These are samsung ram, checked with multimeter. Got 22ohms. 1.37volt when turning on the card. Mats/Mods won't work with 3000 series, do you have any suggestions?
Send it in for repair
Hello brother I need your help I have a gpu p104-100 I cannot run the memory test it tells me that the card is not being recognized and the window gives me the same error 43 it recognizes it and everything but when installing the drivers it gives error 43 in Sometimes the administrator recognizes it well and it can be used but when the machine is turned off or restarted, the code 43 returns so in this gpu I use it to play although it does not have video output, window lets use it with the output of the integrated board If you can help me I would appreciate it very much
Hi I have the exact same issue I would like to send my card in code 43 low resolution I’m in uk
I think my board has this exact same problem, I am not a technician or anything, and I live in a small town, so I'd like to know how would I go about finding this memory banks you replaced, how much do them cost and how do I buy them and such, can anybody help me with anything?
Im having and issue with MSI 1080 gaming X , i have replaced one of the memory modules since mats told it was bad and the card pass the test now no problem. But the problem remains and it is that when the driver is installed it instantly crashes and the PC restarts if u can help me with advice ill be grateful. BTW i changed d9txs with d9vrl memory module it is Micron memory.
Try running MODS test 94 with the memory frequency maxed. MATS test runs at minimum memory speed, so if a module is borderline it will still pass MATS. Throw a scope on your power stages too, make sure your VRMs are hitting on all cylinders.
@@Friend_of_the_One-Eyed_Ladies What do i need to do to make MODS run the memory to the max frequency?
@@Friend_of_the_One-Eyed_Ladies This is what MODS log tells me:
GPC Clock : 1974.0/1987.1 MHz DEFAULT
DRAM Clock : 5005.8/5005.8 MHz MPLL
Host Clock : 571.4/571.4 MHz XCLKGEN3
XBar Clock : 1847.5/1885.7 MHz DEFAULT
Sys Clock : 1708.5/1740.4 MHz DEFAULT
Power Clock : 540.0/540.0 MHz SPPLL0
Display Clock : 1325.0/1325.0 MHz DISPPLL
NVVDD : 1062 mV
Error 000000000143 : Global.SanityCheckLink PCI Express bus error
Exit 000000000143 : JsGpuTest.SetPState (test 0) PCI Express bus error
Error!
Error 000000000143 : JsGpuTest.Run (test 0) PCI Express bus error
Error 000000000143 : Object.RunGpuTests PCI Express bus error
What software did u use to id the bad mem module? Mats?
hello can you tell me what is the name of the program that tests the ram of the gpu? thank you
What temp and air flow settings do you use when replacing memory?
788*F, 100 air
788 - 32 = 756
756 x 5 = 3780
3780 / 9 = 420
@Kids Korner guess why nvidia/amd has such a extensive QA test software even though chips are supposed to QA before the assembly?;)
I mean, there are pogopin testers, capable of checking chips before soldering them on board. There are industry best soldering machines. And yet, hige software suite to test if ram shit itself after it was soldered :)
Mine is okay at low load but at high load it shuts off. If i touch the gpu while its on it will artifact and shut off. B0 has write errors
how and where do you get the usb to use for diagnosing what is the issue on the motherboard? anyone knows?
hi! I have exactly one rtx 2080ti that has a bad memory module (faulty solder) and I watch your video, at minute 7:50 you say you may need to solder the module in the empty slot, so does that mean that socket can be used? in my case for example?
Hi, I may have misspoke, I mean that the other memory modules next to the PCIe slot have to be changed.
@@TechCemetery oooh i see. Mmm so that empty module can't be used? Doesnt work at all?
The empty slot cannot be used
@@TechCemetery well. All hope Is gone for that card 🥴 i was about to buy the neccesary stuff to solder one memory in that module leaving the damaged socket empty hoping to work. Well thank you for the answer :)
Do you do repairs my zotac rtx 2080ti randomly crashed on me while gaming and when into low resolution mode nothing changes even when I install new drivers when I check device manager it's says driver 43 error any help would be appreciated as all my repair shops keep.turning me down I am in UK but willing to ship also it is in low res on my pc but no display in other systems
Yes, I take customer cards in for repair. I have an email listed in the description.
@@TechCemetery I have emailed but my gpu seems to be exact same problem as this one
Is there an AMD version of the MATS tool? I'm assuming an nVidia version will not work for an AMD card. I have an R9 270 that is artifacting due to bad memory modules, and figured this would be a good starting point
its called TServer
Is there a vram test program for amd video cards. can you share this video?
help!
what happends if errors not random?
like
ff000000 ff000000 ff000000 ff000000
i have only this for my FBIOA1 nad its have 1638708 ERROR
Hi guys, what is that program that shows the memory modules problems in this video? I have an RTX 2080 Gigabyte with the same green lines. The only difference that it is only at dark areas, the black areas where it shows the green lines. And it is in low resolution, windows inmediately stop the device because it causes problems (error code 43), so i can still do this test with that program?
hi, did you know the name of the program? I would like it too....
@@nitto1983 Mats and mods is the name of the program. At other videos of this guy there is a link i think for the program, and he explains it well how to use it.
WHAT KIND OF HEAT ARE YOU USING FOR THESE bga CHIPS?
My MSI gaming x trio 2080ti have exactly the same problem like in this video, RMA it , 2 week later received the same old card and work for 2 week and have the same problem again. Now i have to RMA it again and wait for another 2 week, and who know how long it gonna work this time?. MSI RMA service suck big time
I wish mats worked on older ATI GPUs :(
Question: If you remove the memory module, will it boot without it with less memory?
It should at least POST, but with artifacting on a UEFI Bios.
is the MODS/MATS link only for Patreon subscribers? :)
No
@@TechCemetery lol i just realised it's a link to another youtube video with the GDrive folder. thanks! great videos btw, super informative
Why is it that the memory modules near the PCIE slot are the ones that go bad?
GPU Sag and stress around the PCIe connector.
Don't soldier and don't use heat gun it is firmware issue of gpu flash the firmware first upgrade then downgrade your gpu will be fine if not the only solution is to reball
My pc do the same lane but in bios too is that the same case???
Nice work
Could I pay you to fix this 2080 that has basically the same issue?
Where can you buy gddr5x micron 1 Gb chips?
Where can I buy the broken chips for cheap? A chip on my RTX 2080 TI is broken.
I typically get mine from Taobao
Hey there,
Love your work for the community. I have the exact same issue (visually) and now would like to run a memory test.. If I squeeze the card the right way, or move the water cooling rubber pipes to push it up on the right angle it works np..
But for these times, it would be swell to do a mem test for each of these 11gb. Lets say, I have the software or rights to use for a day to help diagnose my card. For this particular card would you mind being a dear and pasting exactly what you have in your file to check this card? My card is also a RTX 2080 Ti with Samsung memory & TU102-300A-K1-A1 gpu ..
Thanks man!
I have an Asus laptop which Is giving me code 43 error and motherboard has Elpida ram on it. Is it possible to change from Elpida to Samsung as I have a scrap board with same density chips in Samsung brand. Any help would be appreciated. Not very much info on the net as to whether this can be done or not
Yes, the graphic card's RAM can likely be switched from Elpida to Samsung
@@TechCemetery thanks so much! That's awesome news! Wanted to check with a pro before I go ahead and do it. Very skilled with microsoldering but not knowledgeable when it comes to ram
@@michaelkuhn8829 You would have to change the memory straps, and then know the exact configuration if it can be supported. What GPU is this?.
@@TechCemetery 960m Nvidia. I am wondering if it is the gpu as I replaced a shorted out mosfet on one of the gpu main rails so wonder if the gpu got kicked in the pants maybe. Also am not managing to get mats to run on laptops at all. Is there a trick to doing this? Have all the correct bios settings too.
i have a gpu thats having similar problem, can you do repair if i ship it to u ? i will love to send mine in and have it fix and do a video on it for the channel
i have a geforce rtx 2080 ti gaming oc 11g
I have an email in the description.
Hi, do you also repair laptop gpu's that are artifacting ?
Yes
@@TechCemetery I sent a request on your website for a repair. I'm from Europe but willing to send the laptop for repairs. Thank you!
isn't the artifacting also caused by the gpu core (it can be), how do you know it is a memory?
It can be caused by a faulty core, you typically can't know if it's the memory or core until you attempt to swap the memory module.
Thanks for answering ☺️
nice videos i love it.
I have a 3070 with the exacte same problème !