Thank you for going more in-depth. I have been watching a few different repair channels, you are on another level. I feel super welcomed in your videos, Like im sitting there with you. You have taught me all kinds of things. Your ingenuity is amazing. Thanks
I subbed a few weeks ago to learn more about gfx repairs, and stayed for the commedic value xD I love how he makes it look so easy while having fun in the process.
I had issues like these with an RX590 after a bad maintenance during the time when I was experimenting with different dissipation mods. The way to fix it in my case was to simply do a proper maintenance. Consider that lowering the voltage also lowers the temperature output, and therefore the difference between edges and hotspot decreases. When the difference is higher, you get thermal stress on the material of the chip itself, causing the artifacts. So evening out the spread of good quality thermal paste could also solve the issue without the need to replace the core.
Reminds me of an old Geforce4 or 5 series I had back in the day. WoW started artifacting one day out of the blue, from then on until I could buy a new card, I kept the side of the case off and had a small desk fan with an icepack behind it blowing on my graphics card whenever I played, worked fine. Sometimes a core just gets some thermal damage over time that only shows up at higher temps, I guess.
@@santiagocastro6701 It's not a bandaid. I had that card for 2 years after that an it worked perfectly, even withstanding overclock in both GPU and memory. I only sold it because I was ready for an upgrade after those 2 years.
I first found your channel a couple years ago i think then somehow I got shuffled to the other North numnuts dude until he started to get on my nerves talking about how successful he thinks he is so I've finally found my way back to your channel. SOO much better, actually subscribed this time!
Great video. The temps have been fine on my Suprim 3090 Ti but i started undervolting. Great to see much less power being used and only lose a few fps.
No, you are wrong. He lowered the voltages across the frequency curve, not a flat number. Read both axes of the chart. How did nobody call you on this giant mistake in a year? Wow.
Informative Video, thanks. does your instruction implies that if the problem solved by reducing the core voltage/frequency then all fault is due to the gpu and no wasting time on reballing and ram chips?
I love how you kept the jokes in a video to a tasteful level. I've seen some bloggers especially in the Philippines who overdo it by spamming giggles and other meme sound effects to the point of annoyance. :/
I had artifacts like this (a bit more intense even) with an RX470 and only lowering the frequency would make it more stable, but driver would crash eventually and reset everything. It was such a pain.
Quite a few times, cleaning the video card fixed the artifacts. Only a very few times I had to replace the core. Anyway, replacing the core does the trick.
Using the polaris bios editor you could have modified the bios to permanently lower frequency and voltage and flashed that on.. frequency does not make a big difference on gpu's so few mhz lower doesn't hurt fps and it is the cheapest "permanent" fix..
@@KuntalGhosh actually I did that and tried other things too, but that card was doomed and would crash eventually. The only thing I couldn't do was replace or even lift the core because I didn't have the tools, and let's not mention a replacement. After a while it got worse and eventually stopped working. To make it even worse it was during the crypto bubble and a card like that was around $400.
Thank You 👍 I just got similar situation with RTX2080ti. After undervolting less artifacts but still there. Very high power draw , GPU-Z - 12V drops under load up to 11.3V and Pcie 12v up to 11V(not PSU fault, 800w verified)From top three vrms one heating more with normal oscillation, on another two coils little strange oscillation like with gaps, not experienced with this. Mods can't pass initial but mats not finding any faults. Where to dig next ? 😅 Dead core ?
Look at the window with the curve. Top-left. The tab that is selected is "voltage". Also, carefully look at the other line jumping around, up and down, marking a horizontal line along the frequency axis. It matches the same frequency shown on the OSD where the FPS and temperatures are.
@@EDOD_EseDelOtroDia Okay, but look at the curve now, stock curve is 700mV 1500MHz and he shifted it down to 700mV 1300-ish MHz, hence higher voltage for the same core frequency. If the core is unstable then feeding it less voltage for a given frequency would not work at all, to attain stability you would feed the core more voltage for a given core clock and it's exactly what Nortwestrepair did. It's just semantics.
@@MrSolvalou Look at the graph. Voltages 1025mv and 1250mv intersect frequency 1.8ghz at the same time. It doesn't make sense that all that range of different voltages would be applied to the same frequency at the exact same time when frecuency 1.8ghz is required. The graph is all messed up, but he actually lowered both voltage AND frequency, lowering the heat output, which was causing the artifacts from thermal stress. You can even see how the center of the chip is devoid of thermal paste precisely at the center where the hotspot is. 2:02
@@EDOD_EseDelOtroDia Nvidia cards take these VF curves as suggestions when boosting, you can see that the card goes above 1.8GHz anyway. By adjusting the curve the way he did he underclocked and overvolted the GPU but it's still gonna boost into the power limit, if you want to reduce heat output you would have to manually reduce the power limit. To undervolt with Afterburner you would raise the whole VF curve up until it's at the edge of instability and also reduce the power limit if you want stock performance with less heat.
I wish cards could come in the same format as the rest of a computer where you can just socket a GPU in there and even upgrade RAM etc, but we know all kinds of problems would come with it, apart from simply not being feasible and would be quite expensive. I didn't know before watching this video that a core could cause artifacts, I always thought it was a memory problem because every time that happens it's simply a memory replacement that fixes the problem.
I've had the same thought in the past... This, unfortunately, would drive the cost and complexity way up. Cards would also be less reliable because of the extra stress and dynamics involved.
Using sockets/slots for GPU's and DIMMs would severely limit memory bus speed - notice that current DDR5 CPU's have similar transfer rates to GPU's from a decade ago.
@@EliteRock I feel you, I do but what's the maximum frequency a CPU can reach while being on a socket? Is it 6GHz? Is it a workable DDR5 that reaches 7,2GHz stable? If it is, I don't see any unsolvable problem.
Considering how difficult (fidgety and hair-pulling frustrating) working with SMD's can be, you do make it look remarkably simple. I don't know how, but you do and it's a joy to watch. Could you perhaps share with us what solder wick you use? I know Goot is excellent but not always available. Cheers mate! 🤘
Which condition that you need to replace the core? Some of your video just need to reball the core to fix this issue, but this one you change the other core instead of reball it first?
hi, i'm a recent viewer. i really love your videos, they are very explanatory. I'm not a gpu fixer but I specialize in consoles and find many useful tips that can also be applied in my field. I would like to ask you a question. I've recently been perfecting my skills in reballing processors and bought the right hot air nozzles for this job, so my question is this: what percentage or fan power do you use to detach the core from the board? is it the same one you use to reattach it too? I ask this because I have noticed that my hot air station transfers heat better to the boards when it is at low ventilation rather than at maximum. thanks if you answer
1:10 That is not really "lowering the voltage", you are shifting the V-f curve down, which results in a combination of lower max. clock and higher voltage at lower clocks. At no point along that curve does that result in a lower voltage.
Look at the window with the curve. Top-left. The tab that is selected is "voltage". Then, carefully look at the other line jumping around, up and down, marking a whole horizontal line along the frequency. It matches the same frequency shown on the OSD where the FPS and temperatures are.
Yo can you tell me whats wrong with my pc ? - i have an rx470 4gb version nitro saphire , -i use atikmdag for driver issues because they dont appear themselves on windows so i have to use this program. - i dont think its because of the driver problem but after the pc starts and gives signal after 1 minute the screen is full of artifacts and then somethimes restart sometime gives BSOD and restarts and sometimes it just goes on worlking without restarting. The artifacts arent individualy showing they appear in a moment and then do the restart thing they are on the whole screen and they are diagonaly out ...
1:08 I think you meant to say underclocking instead undervolting. After lowering that voltage frequency curve, the GPU is now reaching a given frequency at a higher voltage than before.
I have this issue after my system was unplugged from a faulty adapter. Mine is more green dots and no vertical artifacts like that. Maybe the screen is only faulty
my 1060 6gb had some artifacts.but after restaring the pc,it does not make into the desktop.goes black screen right after the windows loading screen.what could be the issue?
How "permanent" of a band aid is undervolting (core/ram)? It is not economically viable to replace the core on my old card, but if i can keep it running undervolted/-clocked maybe a family member with less demanding games can use it for a few more years? Does it depend on which is at fault, ram or core? Speaking of ram, could vram limiting help too? I'm guessing maybe, but if the faulty chip is at the beginning of the allocation it won't help?
So what do you think was wrong with the core? The solder joints cracked within the core or it is worn out from age? If it is worn out then if i'm not mistaken it would need more volts to become stable. Lastly could a reball with the original core have worked?
I had a similar issue and I did a reball, just of the gpu core. Card was out of warranty, so I didn't have anything to lose. I was surprised that process worked. I have had no issues since and just as a fail safe did under volt card after for 2 weeks. Gradually increased voltage back to standard. Applied quality thermal compound and pads.
Seems like the gpu core suffered electrical migration where voltage going somewhere where it shouldn't that why undervoltaging works temporarily. Seems to me that voltage to core seems to be the culprit where higher voltage is not shown but explains the artifacts relation to voltage set too high or the core hates the voltage given to it after it degraded over time
my rx 6600 had artifacting since day one, lowering stuff in MSI does nothing, clean driver install neither, as long as i have no driver there is no artifacting. as soon as I install one its back. i dont know what to do
the gpu core usually costs like more then half the actual gpu price, is there any cheaper way to perform a repair like this? that doesnt involve replacing a super expensive part 💀
Out of curiosity, how did you know it wasn't a bad memory chip? Experience I presume? Didn't see you run a memory test in the beginning (but yep, you said no typical bad Micron memory). So I presume that means damage core and why it had to be physically replaced.
I have a hp pavillion laptop. It's showing this status access violation error on any web browser i use. When i gave to repair they're telling it's the issue of gpu and it's dying. Ive runned all the stress test as you showed in this video and my gpu performed very well. This artifacts issue happens only in web browsers not anywhere else. What should i do?
Had a old xfx Nvidia card that would artifact on heavy load. Noticed temp was high and fan was not really spinning, changed fan curve to be really aggressive and it was fixed lol. But it became a jet engine lmao
Thanks for the video! Very interesting. Would you expect the band aid undervolting solution to fix the symptoms for very long? Or does it only buy youa a bit more time before the gpu degrades further?
Only long term fix is replacing the core. When you first buy the gpu you should undervolt it. If you have not I would now. Undervolting saves lifespan and power.
@@zushikatetomotoshift1575 Or simply to not "buy" overclocked card that are pricier. Their lifespan is also quite low, like 2 or 3 years before core or memory dying compared to 10 year that might be possible if you undervolt it.
@@zushikatetomotoshift1575 might be true for top end cards that are pushed over the limit by the manufacturer but maybe not even then, I have a gtx 960 that I bought in 2015 or 16 and it's been working in 1 of my PCs since, still going strong, never undervolted it and even OCed it for a bit. All that said this days I undervolt my card too but not for longevity reason but for noise and temps reason.
The problem is, for higher demanding games, you will have to lower the voltage even more. Sooner or later, you will end up loosing too much performance and judging by what others were saying, its only going to get worse.
But why exactly this happanes ? We know that artifacts usually comes from VRAM but in this case looks like comes from GPU Core (didnt know it was posible) but can you explain why this happens ? Thanks
I had this issue on 2 brand new out of the box AMD R9 280X, after that I switched to nVidia and I'm never looking back - the "fix" was lowering the core clock by some small amount (~100Mhz if I recall)
@@zushikatetomotoshift1575 Half his videos are reballing cores, just seemed out of place to jump right into replacement. I saw your other comment that the GPU was dying, so that makes more sense.
Ohno, i am having similar issues, but only with few games, and it is only a green variation of the artifacts shown in the video, and that ocasionally show up covering only a line in some part of the screen, usually when i have minecraft open in full screen, and i go to exit it to a brouser or click out of the game, the artifact show up in the seccond monitor, where the game is not at, but when i change the game to the seccond monitor and i try to leave it, nothing happens. in phasmophobia it just happened in the middle of the match, in the monitor where i play this time, but once or twice in all the games i played in TABG, the same artifact show up when i open the "group tab"? and only when i open that, i am honestly confused since these are the only moments when that happens, i tried every single program to find problems, but nothing ever happens there. Should i just find a guy that will change pieces off my gpu? is there something related to temperature? should i test with a specific program? i just want some advice here ngl.
Plis help me i have a issue with my amd grafics card, games that use fxaa or taa look grany, just look awful mostly the vegetation, cant find any solution to fix these grafical artifacts
could a faulty/not strong enough power supply also cause this? I recently got a new graphics card and all of a sudden I've got loads of artifacts in most of my games, and It's been a while since I've upgraded my power supply.
My gtx 980ti is artifacting when i boot into os. And then not long after it just crashes.. I tried booting in safemode and using msi afterburner to lower voltage, but it won't even register on afternurner..
Amazing job as always! I wish you could fix my old m/b. It got pcb scratches behind it, so it turn on and off all the time. That's the only fault with it.
What if I have some flashing artifacts in one particular place of one particular game and they go away after switching to DLSS balance/perf/ultraperf?Is it a hardware issue or a software one?The game is Horizon FW and the card is EVGA 3060 XC.Undervolting/underclocking to absolute minimum does totally nothing and the card successfully passes any kind of benchmarks or stress tests
Hi tony, would like to ask what solder tip you use and solder wick and if it has a technique, becuase mine won stick the lead , i mean the lead on pcb doesnt sucked by solder wick
hmm, strange one, you lower the speed to not get artifacts, but the voltage was not sufficient, it looks like to me a old (already push GPU), OR the power delivery was not up to spec and the power controller (VRM) was damage, just saying this because the siphons looks like a bypass resistance on the power to force the card to give more power without hitting the power limit, that can push the core too much or the VRM's. Sure changing the core will fix it, but if I'm right that card will return
The thing that I've learned from this video is: "We don't do band-aids here". 👍
don't buy Gigashit
Better than factory
The cat playing the piano had me laughing. Thanks for all the great videos
Yeah! Tony's work is amazing, but, sorry Tony, cats are more fun, they got my full attention :P
Thank you for going more in-depth. I have been watching a few different repair channels, you are on another level. I feel super welcomed in your videos, Like im sitting there with you. You have taught me all kinds of things. Your ingenuity is amazing. Thanks
I subbed a few weeks ago to learn more about gfx repairs, and stayed for the commedic value xD I love how he makes it look so easy while having fun in the process.
I had issues like these with an RX590 after a bad maintenance during the time when I was experimenting with different dissipation mods. The way to fix it in my case was to simply do a proper maintenance. Consider that lowering the voltage also lowers the temperature output, and therefore the difference between edges and hotspot decreases. When the difference is higher, you get thermal stress on the material of the chip itself, causing the artifacts. So evening out the spread of good quality thermal paste could also solve the issue without the need to replace the core.
Reminds me of an old Geforce4 or 5 series I had back in the day. WoW started artifacting one day out of the blue, from then on until I could buy a new card, I kept the side of the case off and had a small desk fan with an icepack behind it blowing on my graphics card whenever I played, worked fine. Sometimes a core just gets some thermal damage over time that only shows up at higher temps, I guess.
i don't want to stir the coffee, but still is a bandaid, because still has the issue
@@santiagocastro6701 It's not a bandaid. I had that card for 2 years after that an it worked perfectly, even withstanding overclock in both GPU and memory. I only sold it because I was ready for an upgrade after those 2 years.
@@santiagocastro6701 alternative bandaid?
Maybe
I appreciate your humor in these video's. Not only do I get to learn something, you make it enjoyable too. Keep up the great work.
Nice work. love to watch your videos thank you.
I first found your channel a couple years ago i think then somehow I got shuffled to the other North numnuts dude until he started to get on my nerves talking about how successful he thinks he is so I've finally found my way back to your channel. SOO much better, actually subscribed this time!
Great video. The temps have been fine on my Suprim 3090 Ti but i started undervolting. Great to see much less power being used and only lose a few fps.
this channel deserves more subs.
I love your sense of humor. BTW you have lowered GPU Frequency not Voltage in Afterburner.
No, you are wrong. He lowered the voltages across the frequency curve, not a flat number. Read both axes of the chart.
How did nobody call you on this giant mistake in a year? Wow.
I enjoy watching you fixing.
thanks for sharing. 👍
Lov'n the cat on the piano.
Great sense of humor combined with Pro Repair. 👍
Informative Video, thanks. does your instruction implies that if the problem solved by reducing the core voltage/frequency then all fault is due to the gpu and no wasting time on reballing and ram chips?
I love how you kept the jokes in a video to a tasteful level. I've seen some bloggers especially in the Philippines who overdo it by spamming giggles and other meme sound effects to the point of annoyance. :/
2:03 Linus moment
you are amazing!!! i saw your videos all this week end and a love how you find solutions for strange issues. cool
Valley Benchmark is a real way to chill after these very tricky repairs, especially when the fixes are a success
Awesome video I just bought a brand new 4070ti super and am getting artifacts like this… frustrating but I guess I may have to send it back.
Thanks a bunch for the help on this one! The card has been running great in my fiancés build.
I had artifacts like this (a bit more intense even) with an RX470 and only lowering the frequency would make it more stable, but driver would crash eventually and reset everything. It was such a pain.
you pushed the core too hard only real fix is replacing the core what you did is only a temp fix not long term.
Quite a few times, cleaning the video card fixed the artifacts. Only a very few times I had to replace the core. Anyway, replacing the core does the trick.
Using the polaris bios editor you could have modified the bios to permanently lower frequency and voltage and flashed that on.. frequency does not make a big difference on gpu's so few mhz lower doesn't hurt fps and it is the cheapest "permanent" fix..
@@KuntalGhosh actually I did that and tried other things too, but that card was doomed and would crash eventually. The only thing I couldn't do was replace or even lift the core because I didn't have the tools, and let's not mention a replacement. After a while it got worse and eventually stopped working. To make it even worse it was during the crypto bubble and a card like that was around $400.
@@Lucius4992 you could have sold it with a very low frequency bios 🤣
Thank You 👍 I just got similar situation with RTX2080ti. After undervolting less artifacts but still there. Very high power draw , GPU-Z - 12V drops under load up to 11.3V and Pcie 12v up to 11V(not PSU fault, 800w verified)From top three vrms one heating more with normal oscillation, on another two coils little strange oscillation like with gaps, not experienced with this. Mods can't pass initial but mats not finding any faults. Where to dig next ? 😅 Dead core ?
Always a great vid to watch!
By shifting the V/F curve down you actually overvolted the core (lower frequency for each given voltage step).
Great repair as always!
Look at the window with the curve. Top-left. The tab that is selected is "voltage". Also, carefully look at the other line jumping around, up and down, marking a horizontal line along the frequency axis. It matches the same frequency shown on the OSD where the FPS and temperatures are.
@@EDOD_EseDelOtroDia Okay, but look at the curve now, stock curve is 700mV 1500MHz and he shifted it down to 700mV 1300-ish MHz, hence higher voltage for the same core frequency. If the core is unstable then feeding it less voltage for a given frequency would not work at all, to attain stability you would feed the core more voltage for a given core clock and it's exactly what Nortwestrepair did. It's just semantics.
@@MrSolvalou Look at the graph. Voltages 1025mv and 1250mv intersect frequency 1.8ghz at the same time. It doesn't make sense that all that range of different voltages would be applied to the same frequency at the exact same time when frecuency 1.8ghz is required. The graph is all messed up, but he actually lowered both voltage AND frequency, lowering the heat output, which was causing the artifacts from thermal stress. You can even see how the center of the chip is devoid of thermal paste precisely at the center where the hotspot is. 2:02
@@EDOD_EseDelOtroDia Nvidia cards take these VF curves as suggestions when boosting, you can see that the card goes above 1.8GHz anyway. By adjusting the curve the way he did he underclocked and overvolted the GPU but it's still gonna boost into the power limit, if you want to reduce heat output you would have to manually reduce the power limit. To undervolt with Afterburner you would raise the whole VF curve up until it's at the edge of instability and also reduce the power limit if you want stock performance with less heat.
So do you think that was a bad core and not the solder joints? Whta would have caused that, overvolting and copper migration?
but where can you buy a replacement core? or did you salvage it from a different card?
You earned a like for the piano playing cat synced to royalty free music
finalmente fue un reballing, muy bueno tu video, te ganaste un suscriptor
Thank you for the video and the good information that comes with it!
I wish cards could come in the same format as the rest of a computer where you can just socket a GPU in there and even upgrade RAM etc, but we know all kinds of problems would come with it, apart from simply not being feasible and would be quite expensive.
I didn't know before watching this video that a core could cause artifacts, I always thought it was a memory problem because every time that happens it's simply a memory replacement that fixes the problem.
I've had the same thought in the past... This, unfortunately, would drive the cost and complexity way up. Cards would also be less reliable because of the extra stress and dynamics involved.
Sure, if you want it, i have a video card with memory slots. It's an old ISA card.🤣
@@m8hackr60 Way up, you say? It would cost probably 10 bucks or less the part plus labor.
Using sockets/slots for GPU's and DIMMs would severely limit memory bus speed - notice that current DDR5 CPU's have similar transfer rates to GPU's from a decade ago.
@@EliteRock I feel you, I do but what's the maximum frequency a CPU can reach while being on a socket? Is it 6GHz? Is it a workable DDR5 that reaches 7,2GHz stable? If it is, I don't see any unsolvable problem.
Considering how difficult (fidgety and hair-pulling frustrating) working with SMD's can be, you do make it look remarkably simple. I don't know how, but you do and it's a joy to watch. Could you perhaps share with us what solder wick you use? I know Goot is excellent but not always available. Cheers mate! 🤘
Which condition that you need to replace the core? Some of your video just need to reball the core to fix this issue, but this one you change the other core instead of reball it first?
Keyboard cat got me. Another thumbs up for you!
hi, i'm a recent viewer. i really love your videos, they are very explanatory. I'm not a gpu fixer but I specialize in consoles and find many useful tips that can also be applied in my field. I would like to ask you a question. I've recently been perfecting my skills in reballing processors and bought the right hot air nozzles for this job, so my question is this: what percentage or fan power do you use to detach the core from the board? is it the same one you use to reattach it too? I ask this because I have noticed that my hot air station transfers heat better to the boards when it is at low ventilation rather than at maximum. thanks if you answer
1:08 I never noticed till now that Valley does not report the correct current core clock speed.
Thank you for another wonderful upload.
1:10 That is not really "lowering the voltage", you are shifting the V-f curve down, which results in a combination of lower max. clock and higher voltage at lower clocks. At no point along that curve does that result in a lower voltage.
Look at the window with the curve. Top-left. The tab that is selected is "voltage". Then, carefully look at the other line jumping around, up and down, marking a whole horizontal line along the frequency. It matches the same frequency shown on the OSD where the FPS and temperatures are.
Awesome fix and the music is the cream 😊
Didn't know the core itself can also cause artifacts, i thought it just works or it doesn't. Thanks for this 🙏 very interesting
It happens when you push the core too much. Why I recommend undervolting.
Most certainly can, fixed a ton of the HP dv6000/9000 series laptops with some in the family still working ten years later.
Same with GPU’s
@@zushikatetomotoshift1575
Undervolt all you want, they don’t fail in stock form cause someone “pushed” them too far.
Yo can you tell me whats wrong with my pc ?
- i have an rx470 4gb version nitro saphire ,
-i use atikmdag for driver issues because they dont appear themselves on windows so i have to use this program.
- i dont think its because of the driver problem but after the pc starts and gives signal after 1 minute the screen is full of artifacts and then somethimes restart sometime gives BSOD and restarts and sometimes it just goes on worlking without restarting.
The artifacts arent individualy showing they appear in a moment and then do the restart thing they are on the whole screen and they are diagonaly out ...
Impressed it worked. Nice job.
Thank you for the video
1:08 I think you meant to say underclocking instead undervolting. After lowering that voltage frequency curve, the GPU is now reaching a given frequency at a higher voltage than before.
You may be right.
This option is not available unless the voltage tweak is unlocked so I assume it was the core voltage.
I have this issue after my system was unplugged from a faulty adapter. Mine is more green dots and no vertical artifacts like that. Maybe the screen is only faulty
Can you do a vid just on how to lower the voltage?
my 1060 6gb had some artifacts.but after restaring the pc,it does not make into the desktop.goes black screen right after the windows loading screen.what could be the issue?
The cat piano stole the show for me 😆
How "permanent" of a band aid is undervolting (core/ram)?
It is not economically viable to replace the core on my old card, but if i can keep it running undervolted/-clocked maybe a family member with less demanding games can use it for a few more years?
Does it depend on which is at fault, ram or core?
Speaking of ram, could vram limiting help too? I'm guessing maybe, but if the faulty chip is at the beginning of the allocation it won't help?
So what do you think was wrong with the core? The solder joints cracked within the core or it is worn out from age? If it is worn out then if i'm not mistaken it would need more volts to become stable. Lastly could a reball with the original core have worked?
I had a similar issue and I did a reball, just of the gpu core. Card was out of warranty, so I didn't have anything to lose. I was surprised that process worked. I have had no issues since and just as a fail safe did under volt card after for 2 weeks. Gradually increased voltage back to standard. Applied quality thermal compound and pads.
Appreciate your work. What kind of station you prefer to doing this thinks. aixun t3a or aixun t3b? Thanks in advance for your opinion.
Seems like the gpu core suffered electrical migration where voltage going somewhere where it shouldn't that why undervoltaging works temporarily. Seems to me that voltage to core seems to be the culprit where higher voltage is not shown but explains the artifacts relation to voltage set too high or the core hates the voltage given to it after it degraded over time
my rx 6600 had artifacting since day one, lowering stuff in MSI does nothing, clean driver install neither, as long as i have no driver there is no artifacting. as soon as I install one its back. i dont know what to do
the gpu core usually costs like more then half the actual gpu price, is there any cheaper way to perform a repair like this? that doesnt involve replacing a super expensive part 💀
welldone. wish you 1m subs :D
If the core is the issue, where can you get a replacement??
Out of curiosity, how did you know it wasn't a bad memory chip? Experience I presume? Didn't see you run a memory test in the beginning (but yep, you said no typical bad Micron memory). So I presume that means damage core and why it had to be physically replaced.
The type of artifacts it was showing Core artifacts are different than Mem artifacts.
Both look and act different.
Also big common issue that causes this is when someone pushes too much clocks or volts or both.
I recommend Undervolting all gpus
I even recommend undervolting cpus.
00:05 seconds. Little sticker from the bottom..
Core worn out from the factory overclocking.
@@era7928Nope it's due to user not factory OC. With doing research on how stuff works you would know it's due to heat manual OC and Volts.
What thermal paste brand do you suggest for a hot GPU ?
I have a hp pavillion laptop. It's showing this status access violation error on any web browser i use. When i gave to repair they're telling it's the issue of gpu and it's dying. Ive runned all the stress test as you showed in this video and my gpu performed very well. This artifacts issue happens only in web browsers not anywhere else. What should i do?
Had a old xfx Nvidia card that would artifact on heavy load. Noticed temp was high and fan was not really spinning, changed fan curve to be really aggressive and it was fixed lol. But it became a jet engine lmao
Thanks for the video!
Very interesting. Would you expect the band aid undervolting solution to fix the symptoms for very long? Or does it only buy youa a bit more time before the gpu degrades further?
Only long term fix is replacing the core.
When you first buy the gpu you should undervolt it.
If you have not I would now.
Undervolting saves lifespan and power.
@@zushikatetomotoshift1575 Or simply to not "buy" overclocked card that are pricier. Their lifespan is also quite low, like 2 or 3 years before core or memory dying compared to 10 year that might be possible if you undervolt it.
@@era7928 You can oc a little with undervolt and have longer lifespan.
@@zushikatetomotoshift1575 might be true for top end cards that are pushed over the limit by the manufacturer but maybe not even then, I have a gtx 960 that I bought in 2015 or 16 and it's been working in 1 of my PCs since, still going strong, never undervolted it and even OCed it for a bit. All that said this days I undervolt my card too but not for longevity reason but for noise and temps reason.
The problem is, for higher demanding games, you will have to lower the voltage even more.
Sooner or later, you will end up loosing too much performance and judging by what others were saying, its only going to get worse.
you are a king man chief
Where are you located? How much would it cost to repair a Asus 3080ti RTX?
But why exactly this happanes ? We know that artifacts usually comes from VRAM but in this case looks like comes from GPU Core (didnt know it was posible) but can you explain why this happens ? Thanks
Hello im currently expirencing artifacting but i've ran furmark for a good 3 hour i didnt get a single crash what could be causing the problem
I had this issue on 2 brand new out of the box AMD R9 280X, after that I switched to nVidia and I'm never looking back - the "fix" was lowering the core clock by some small amount (~100Mhz if I recall)
cool channel. Sad I don't live in the US but in Austria.
i didnt really understand what the problem is. bad gpu passed through QA at factory ?
KITTY!🐈 Ok. Now you earned the thumbs-up.👍😉
you can probably find that card for $250 how much was the work is it worth it ?
Good luck finding it at that price working.
Here they sell for 200-250 with 1 month seller warranty 🤔
@@zetsubou3704 I will say if you want to risk it it's a good price.
Hello, how long can a video card last if the chip is repaired like in this video?
Why didnt u just reheat the old processor? may be it was just cold solder joints
GPU artifacts *exist*
owner: its the monitor, right?
GPU: ...
owner: its the monitor, right?
your workshop looks very clean sir, can u make tour video?
Would reheating the original core or reballing be worth a try before replacing?
that is only a temp fix.
@@zushikatetomotoshift1575 Half his videos are reballing cores, just seemed out of place to jump right into replacement. I saw your other comment that the GPU was dying, so that makes more sense.
i have a GPU 4070 gtx but this method not working i mean the lowring the temp part hayz
Such ASMR just by looking at the motion.
Ohno, i am having similar issues, but only with few games, and it is only a green variation of the artifacts shown in the video, and that ocasionally show up covering only a line in some part of the screen, usually when i have minecraft open in full screen, and i go to exit it to a brouser or click out of the game, the artifact show up in the seccond monitor, where the game is not at, but when i change the game to the seccond monitor and i try to leave it, nothing happens.
in phasmophobia it just happened in the middle of the match, in the monitor where i play this time, but once or twice in all the games i played
in TABG, the same artifact show up when i open the "group tab"? and only when i open that, i am honestly confused since these are the only moments when that happens, i tried every single program to find problems, but nothing ever happens there.
Should i just find a guy that will change pieces off my gpu? is there something related to temperature? should i test with a specific program? i just want some advice here ngl.
Just fixed a "dead" as is GTX 1060 6 gb, replaced thermal pads, thermal paste, did 3 x furmark test, all good. Heat dissipation is a must with a gpu.
Plis help me i have a issue with my amd grafics card, games that use fxaa or taa look grany, just look awful mostly the vegetation, cant find any solution to fix these grafical artifacts
bruh my gpu is artifacting while watching this 😐😐😐😐😐😐
Mine is doing it on startup💀
could a faulty/not strong enough power supply also cause this? I recently got a new graphics card and all of a sudden I've got loads of artifacts in most of my games, and It's been a while since I've upgraded my power supply.
My gtx 980ti is artifacting when i boot into os. And then not long after it just crashes.. I tried booting in safemode and using msi afterburner to lower voltage, but it won't even register on afternurner..
Could this issue be possibly fixed by just resoldering the chip? Maybe the old one wasn't really that bad. Idk, you should do a video on it
That's very nice but how do you move the whole voltage curve in the afterburner?
Amazing job as always! I wish you could fix my old m/b. It got pcb scratches behind it, so it turn on and off all the time. That's the only fault with it.
What if I have some flashing artifacts in one particular place of one particular game and they go away after switching to DLSS balance/perf/ultraperf?Is it a hardware issue or a software one?The game is Horizon FW and the card is EVGA 3060 XC.Undervolting/underclocking to absolute minimum does totally nothing and the card successfully passes any kind of benchmarks or stress tests
what did you do to get rid of the artifacts. i dont have money to send my pc anywhere and i get the artifacts in almost every game i play
im having same issue with gigabyte 6800xt oc, do you think its the brain on the card causing this artifact too?? are they easy to get hold of???
Are you around the Portland, OR area?
It seems to much voltage was the cause of the issue. Am I right? If so, is this a generale behaviour?
The cat was a blast !
I got an unused 2070 that stops any system it's in from booting. You think it's worth sending in for repair?
I have artifacting Even on desktop is there a fix or do i have to ditch it?
Hi tony, would like to ask what solder tip you use and solder wick and if it has a technique, becuase mine won stick the lead , i mean the lead on pcb doesnt sucked by solder wick
i think your discord got hacked. Your account is spamming my DM
superwick
@@northwestrepair yes , yesterday i changed that password, still spamming until now? Wait ill go check it.
@@motherboardrepairsPH it was yesterday. no more i think. I blocked you LOL
@@northwestrepair haha , now i think ok now, idont know how it got hacked
So the core is the culprit with a problem like that?
I have a GTX 1060 and a RTX 2080 with almost the same issue
It's not always the core for 2000 cards.
For 1000 the other way around.
@@northwestrepair thanks bro
My pc does it on the main desktop? But no games
hmm, strange one, you lower the speed to not get artifacts, but the voltage was not sufficient, it looks like to me a old (already push GPU), OR the power delivery was not up to spec and the power controller (VRM) was damage, just saying this because the siphons looks like a bypass resistance on the power to force the card to give more power without hitting the power limit, that can push the core too much or the VRM's.
Sure changing the core will fix it, but if I'm right that card will return
Not according to the stats but who knows.
@@northwestrepair , true, lets see if my theory is right or wrong, 😉
hi i have a N460 GTX cyclone ( 1024MB version ) and i keep getting artifacts and then the it crashes , anyone knows what the problem is ?
sir, may i know whats the causes of core gpu fault?, it was end user physical damage?
my 2080ti has smearing artifacts in fast moving. some have idea how to fix ?