I used to work in PCB assembly plant, and they stopped using lead solder because of the health issues that they could cause to the operators. Some bits were manual soldering and other bits were done by machines.
00:03 Fixing a 3080 TI from PNY that broke while idling 03:12 Impressed with the ease of disassembly 09:16 Discussion on the components and diagnostic process of the 3080Ti 11:56 DIY modification using sandpaper to make M3 standoff fit M2 card holes 16:58 Swapping graphics card to 3080 Ti in BIOS setup 20:17 Troubleshooting and testing GPU performance for artifacts 26:32 Assembly of the 3080Ti is smooth and enjoyable 29:03 Achieve optimal performance by ensuring hot spot and memory temperatures are similar 34:38 Showing a 3080 Ti with LED lights 37:03 Sharing information about 3080Ti fix
You should never work for free, period. I haven't seen a single YT "fixer" that offers no fix, no fee. There is a well known repair guy in the UK that charges minimum £60 per hour, plus components. This is highly skilled work, (if done properly), and your work is first class. Don't sell yourself short. Start charging a minimum price, just for looking at these things, then charge per hour. Your videos speak for themselves. Keep up the excellent work Tony 👍👍👍
@@rateater420he charges shipping and such but lewis runs a company that makes decent cash, big difference from just losing your job and hoping fixing video cards on UA-cam works out. But no fix no fee really doesn't hurt the business too badly as long as they pay the shipping.
Hmm.. I welcome the challenge; I'm quite confident that I can locate a UA-cam repair engineer who offers free services on an Nvidia GPU board. Just allow me a few mins, I'll be back, while waiting hit my S-button. 🙄
Impress myself sometime, I've got a handle on a guy by the name of, NorthridgeFix. In a video about Nvidia GPUs released five days ago with this UA-camr highlighted that certain users received both a cost-free repair and free shipping.
Your videos have literally inspired me to learn soldering. Picked up a cheapish starter kit and plan on tearing down and practicing repair on some of my old parts ill never use again. Keep up the A+ work, and i greatly admire your incredibly honest and fair business strategy. Especially these days when times are tough for the bulk of us. I see people shaming you for not charging enough, but don't seem to realize not everyone needs to be wealthy. Just enough to continue our passions and live without stress.
I had one customer who said their computer screen would go black, just by grabbing any window, and shaking it on the desktop. Sounded really odd to me. Then i saw the PSU Geek Squad put in his PC. It was an Apevia 500 watt. His problem was it couldn't handle 300 watt. The 1080 video card was starving to death, and even 2-D was enough to frack everything.😂 If all i had was a 1080 staring up at me, i would not have found anything wrong.
I find weak or insufficient power supplies in alot of computers I fix, just had an old vega 64 and it was running on a 550 watt, customer said it went black screen and the gpu fan would go 100%. 😂
i personally have an apevia 520w and i don't see how the hell it crashed. the 1080 draws 175, and any ok CPU draws anywhere from 95-135 watts, which is far under the power draw limit. I personally used an rx 580 on my PSU that draws way more than a 1080, and it ran just fine.
@@gargolgaming8101 The one I worked on had a vega 64 and a 5700g, wall power only registered 440 or so tops, 550 watt p.s. which happened to be an apevia oddly would run fine under less demanding games and black screen and fan would go 100%, replaced with a 750 and it runs fine. Power supply tested fine, I put it in my potato test computer with a 5500xt.
@@gargolgaming8101 Simple. Inline watt meter didn't lie. I changed nothing else but his PSU. I flipped him into a used Rosewill semi-modular 750, for 20 dollars, and didn't charge for install. That solved the problem. No more crashing. The apevia 500w had a strange smell, like bad clutch pads. If you got lucky, well enjoy it.
I'm pretty sure Greta was around 3 years old when RoHS was enacted. And the issue isn't lead-free solder _per se_ (of which there are multiple types - some significantly better than others), it's the fact that some manufacturers try to use the cheapest one they possibly can. Leaded solder is still allowed (even for industrial production) as long as manufacturers prove that (even good) lead-free solder isn't as reliable for their application, and commit to collecting and recycling the products they sell. A lot of medical equipment still uses leaded solder. Manufacturers who just want to sell their stuff and let others deal with the waste are (and should be) required to use lead-free.
LIke he said, unfortunately she's not the real source either. She's another consequence. You're not allowed to name the source on these platforms because they're so powerful that you can't talk about how powerful they are, while telling us that they have no power. ;) I believe Voltaire said something about that.
"A lot of medical equipment still uses leaded solder." Don't know how it it in USA but in Europe there is strong rumors it will be a ban for med equipments as well. We use lead-free for everything anyway.
@@TinMan445Not necessarily. I learned how to build a PC by simply reading and looking at pictures. I have never messed up anything but then again I know to be very careful and take certain precautions, also not to f**k with stuff if things are already working properly. These types of videos can be invaluable for learning, just you know some people shouldn't open things up or mess with stuff as they just don't have the patience or delicateness at all.
Been watching your videos for a while, more entertaining than some other content creators with "north" in their name. I know nothing about electrical repair, but I find your videos soothing and enjoy watching them. Just subscribed. Thanks for all the content.
Thankyou for making these videos my brother. I am a repair technician in Zimbabwe , and although we don't have these high power cards here, these are increasing my confidence when fixing the slower/cheaper cards.
Ty for the new found information about the hot spot and the memory running at the same temperature and that there shouldn’t be 3 deltas on the temp and if you do have it then that’s a sure indicator that you need to replace the thermal pads on the memory. And while your doing that clean and reapply thermal paste to the core. I’ve always cracked them open . Never dawned on me to use the software side of things to identify worn out thermal pads. But in my defense back in late 90’s and up to 2012 there really wasn’t that much software available or accurate for such diagnosis. Either way TY sir for my new found knowledge. There was much more that I learned but thermal pads was the easiest to share. 🍻
mine any kawpow algorythm with it, that cranks the core and memory to max. If that passes, and no visual issues in win, i'd consider it perfectly fine. Users have messed up systems, dirty dusty/greasy pcie connectors with dog fur and halfway plugged 8pins, but you are super precise to approach even a perfect card from the perspective that it might be shorted. Respect, i learned a lot again :)
i have been hear from the beginning . i mean abou3k subs. and i have watched every video. i'm repairing laptops but very interested in gpus . love to see your channel grow.
on cheap cards, the memory can go higher than hotspot temps because the memory chips aren't cooled with any cooler. anecdotally, memory chips without coolers can perform as good or better vs memory chips attached to coolers. Some cards with inadequate or weak coolers just add more heat to the memory chips
Just a little tip on these cards that you know is randomly crashing boot the card 3amp current limited if the draw is normal go straight to testing before you open it up.. In my experience in the case of cold solder joints flexing the board can temporarily fix the problem and take way more time than necessary.. Also could just be a thermal issue that could be easily diagnosed before you tear it down.. Personally I never take a card apart until it's power tested and if it passes then I'll test on the test bench to determine what the problem is.. In my opinion it's a way better than the route than you've been taking lol.. Like i said it just my opinion and everybody does things differently.. Anyways I enjoy the videos keep up the good work..
Trouble shooting step #1 no matter what is being tested is test as is to see the problem yourself... or not. Test drive the car or test the GPU? Doesn't matter, it should happen as is before any tools are out.
I also always start with a mild diagnosis and then gradually increase the level of diagnosis, but I enjoy watching this video so I can get to know more about the contains of various graphics cards.
Have you ever considered expanding outside of just GPU repairs? It seems like your advanced knowledge of circuitry would make repairing other electronics possible, such as older consoles, Mobos, and what not. You'll definitely stay busy, but more options should bring in more money, especially since you do amazing work.
Nice to see the same model that I own on your hands, and discovering good and bad things about it. Thanks as always for your video! Of course, thanks too to your funny comments, they make may day ;)
100% correct. The PNY cards are the same/identical as the Gainward, Palit & several others that use the Nvidia reference design sent out to the OEM’s. Made in same factory (obviously) & only differences are the VBIOS & stickers. I recently took my PNY 3080ti card apart to repaste/change pads & the X grace screws were literally able to be turned by light finger pressure! Once I had changed the pads/paste (used 18W/kM pads from Temu!!! & Noctua NT-H2 paste), the core temps reduced by 22C & fans speed did not ramp up automatically to 60% when on auto & now barely break 40% when gaming for a while. Hotspot delta was approaching 20C difference to the core before, but once reassembled, it is now 10-12C instead (would prefer 8-10C but happy enough). The fan connectors are a pain in the posterior I agree (I used a pair of very fine tipped tweezers to remove the 2 fan connectors before taking the HSF off but the LED connector (for the GeForce illumination) came off quite easily in comparison! The paste coverage (from new) on the core was not sufficient either as was a lot of ‘clean’ areas on both the core & contact plate where there was no paste contact & no ‘pump out’ either. Was either a Friday afternoon card or the factory needs to increase the amount dispensed onto the core/contact plate.
Funnily enough, I've been buying Gigabyte boards for 25 years (my first Gigabyte was a Riva TNT2 M64), and I've never had any issues with them. They all worked flawlessly and still work to this day (yes, I still have them, I'm a collector), although, after all of these years, I did have to replace some capacitors due to the capacitor plague from '99 - '05. The only card that died on me was an MSI GTX 760 (artifacts consistent with faulty memory ICs), but MSI was very quick to replace the card and I still have its replacement which still works after 10 years.
Gigabyte WAS sweet, back when i had a Slot-A motherboard. Now they sell motherboards that are DOA. Disappointing, would be my word of choice for Gigglebutt.
I have various cards and boards from every manufacturer and they all still work to this day. Some still running OC'd daily in family systems. I really only starting seeing a notable quality drop into the DDR4 era.
@JJFX-, all electronic devices have a 1 - 3% failure rate. That's just how it is, how it's always been, and it will continue to be true in the future as well. You can't really judge a product because you happened to receive a DOA device. I can assure you that most Gigabyte motherboards, video cards are not DOA. 🙂
@@ruxandy It could be much more than 1 - 3% in Gigabyte's case. Because I got 2 Gigabyte B650M DS3H boards IN A ROW that had a memory VRM coilwhine issue, the third one didn't though. I'm pretty sure I just got really unlucky, and I still love Gigabyte's products, but anyway, this was my experience.
@@atirta777 , coil whine is a completely separate issue altogether (and one that is very difficult to pinpoint). Since I have hundreds of computer parts, I know the issue all too well, and what I've found is that a certain device/PC component can have coil whine when used in conjunction with certain other parts, and the same device might be completely silent on another PC, with different components.
Friday morning, my old Asus EAH 5850 had a short circuit, and burned up on the PCIe bus on the electric side, my computer shut down immediately, and luckily the Motherboard, CPU, Memory and the PSU were spared, good I was at home, and could fast remove the GPU from burning further. And the PCIe slot, after cleaning, still works. Thank god. I had plans to change this GPU anyway, so now I have to speed up things. Never had a GPU blowing up that way.
You are probably flooded by work after all these videos of successful repairs , we could use someone like you here in Slovakia. Great job. I got to learn a lot about cards....
I use a RTX 3080ti and the card runs hot. So, I copper molded it and now they work perfect. I have heard the cases against copper modding. But, I did not make a slurry of Thermal Paste all over. I strategically placed it on the ram chips. The card runs at least 35% cooler.
They charge so much and still make them so cheaply because nvidia increasingly has reduced the margins for the card manufacturers. They tell them how much they are going to pay for the chips and how much they are going to sell the cards for. I think maybe because nvidia wants its founder's cards to be the best, or is wanting to take over all manufacturing eventually, or also maybe just because they can.
Hello There, it's 6am , iam in Buenos Aires, Argentina, My neighbors stores alarm was trigged at 4 am , so idk WTF i am doing , lol but love your videos
5:52 These vids are much more entertaining than I ever woulda thought 👀👏🏻 6:36 Revenge can be so sweet sometimes 😝 9:41 This is actually a real question, now we're seeing the inside of these things 🤔 and I know - because they can ✅ but hey, now we're seeing what they're actually doing 🔍👀
Hi, love the vids man, great job! on your select output problem, actually can do it via a switch, on your mobo you can force that your onboard display is always on with dedicated. then run a combiner switch to your screen so then you can switch between onboard and what ever is on the desk, i run multiple screen setups from my onboard and 3d card.
Hello Tony (x5) Great video's! I always was interested in hardware more than software but didn't have the right equipment or patience like you. Really cool that you resurrect these cards from the dead and safe them from the junkpile ever growing larger. Big probs, you have a new subscriber, Greetings, The Internet
Personally, my experience is that Furmark generaly doesn't trigger random crashes. Haven't done it for some years now, but i think i remember sometimes getting crashes that would only trigger on certain conditions, like it might only trigger in one of Heaven, or 3DMark's Fire Strike or Time Spy... and then being random, having to leave each benchmark running for hours, to see if it happens... i think they might've cut down on the testing since then, something about AMD generally taking back "unstable cards" without needing tests.
¿ but i think i remember ? What kind of English is that ? Best you undertake an extensive reading program to learn - and become proficient in the English language. I have taught students age 8 who are academically more advanced than you. Also - learn how to format text and use paragraphs. Look in a diction-ary to learn what paragraphs are. Learn why text formatting is important and why you will find it in newspapers - magazines and books. Clearly - you do not have the ability to think and analyze for yourself - You do not pay attention to detail and are mentally lazy You were programmed to be a ZOMBIE ZOMBIES have eyes but cannot see - ears and cannot hear. Your experience is limited and has taught you very little. Read more to educate yourself to a higher level. Good luck and good bye. ## Replies are ignored as I am too busy with other matters. The observation remains FACT - so get the point - Educate yourselves
I have a Gainward 3080 which is basically exactly the same as this. Idle temps are 49(GPU), 62(memory), 63(hot spot), load are 79(GPU), 86(memory), 94(hot spot). Hotspot decently higher than memory. Also my 5600x was hitting 95C under load with the stock cooler. Now with Peerless Assassin around 60C load. I imagine it's because I live in Australia where even during winter I have a fan on in my room, so ambient temps ain't great.
I have a Zotac 3080 12gb, that i think have nice temp: idle: GPU: 41ºc hotspot: 47.5ºc mem: 48ºc , with load: 69ºc-70ºc, hotspot: 79-80ºc MEM:81-82ºc, never been open, i live in Portugal, summer is very hot, today specially (39ºc outside) - Lisbon, no AC at home, maybe 27º to 28ºc at home.
I don't know if that card produced a crash off screen while testing but if I have to guess, it did not. Anyway, I had some similar issue years ago with an pov gtx 570 and an older enermax psu. While idle it worked fine, when gaming, I had to limit every game to 60fps with vsync otherwise instant freeze. Turned out after swaping card, the psu simply was dying, a bunch of caps were swollen. Changed psu and put back card, everything worked fine. I bet, who ever owns this card, maybe has psu issues or a weakening board, as in idle the gpu pulls most power from the pcie slot. Maybe the psus 5v rail is weak or one of the 12v. I would suggest, just for testing, try different psu, doesn't need to have 800+ watts, just let pc with swapped psu sit on desktop. Wait if crash or no.
Sad to hear what you said about Gigabyte video cards at the 4:28 time spot. I have had a Gigabyte 6750X for the past two years and have not had a single problem with it. What I do hate about video cards is that damn tab on the bottom side of the card that I believe is supposed to be some sort of latch to secure it to the mobo, but that is not the problem, what the problem is that movable latch on the mobo. And all the mobo's have that latch. Remove the latch or come up with a better ideal. Sorry for the rant, but the problem is not the video card, it is the stupid latch.
Weird crash symptoms on 3080(xx)? On mine it was simply that the card was either not aligned to the PCI slot or there was something funky with the edge connector. I have a MSI3080 Gaming X Trio and it would crash for no deducible reason when under load or from cold boot about 3 months after I installed it. Was about to send it back for warranty repair but figured I'd look into "variables" some more seeing as I"ve never had a beefcake GPU card before that NEEDED a support to hold it perpendicular in a slot before!!! I fixed it by cleaning the edge connector with DeOxit spray and made sure that it was reinserted and mounted "perfectly". Never had a problem since that "repair" and follow that "repair regime" every 12 months now... Go figure, eh?
Lead free solder pre dates the Thunberg she's only 20 and had been a pain in the ass for less time than the ban , leaded solder is still in use in medical and military applications due to cost , lower melting piont higher wettability and increased strength
I think it should be either with motherboard bios related or Power Supply related crashes. But usually the customer doesn't know such stuff and they end up sending the card for repair.
What's GRETA? Here in Europe we had the infamous RoHS directive twenty years ago that resulted in a tremendous downgrade on solder quality with electronic equipment since it came into force.
As a tech, I know that not finding anything wrong is sometimes worse to report to the customer then yes I found the problem and it will be a little costly to repair it, at least they have an answer.
There is a quicker way to circumvent the change of choosing the display device. It just involves getting a second testbench with a second pci-e extender so all you need to do is switch the card in the slots :) Waste of money and might not be worth the investment, but in the long run I'd say it is as it saves much precious and valuable time.
You should get some flat faced Fingernail Clippers for the plastic notches at 6:56 :) Much safer!!! And probably useful for other parts as well when you just need a clean cut.
I have a pny 3060ti that had such bad voltage drop on the pcie slot that i re routed the shunt to a separate 6 pin and added jumper wires all over the card. More caps only helped so much. But the card has a good core and decent memory so cant get everything i suppose
Krisfix is way more experienced than Northwest and he says the reason behind broken solder balls is lack of GPU bracket. He also used lead free solder, he is not against them. I would trust him rather than some tinfoil hat conspiracy.
@@hmello3250 lead free solder can be a big source of problems due to material properties, and there are many reputable scientific studies on that (some even from NASA).
Most of the GPU sag is caused by cards not having sufficient back plate fitted (if have one!). I have a PNY RTX2080 card in my wifes machine (PNY 3080ti in mine), that sags like a bitch, but on close inspection, it is the heatsink that is sagging as it is literally double the length of the PCB & not got a back plate. If the heatsink screws into the back plate in the areas where the PCB is not present, then sag is generally very low. I tested this with my PNY 3080ti (other model to the one in this video which only has lighting on the GeForce text on the side of the shroud) by checking out the sag with & without the backplate fitted & without the backplate, the PCB was level to the slot, but the heatsink was sagging a LOT. Still use anti-sag support with the 3080ti even with the backplate to be safe (I use the Lian Li support bracket kit that replaces 2 of the motherboard screws with standoffs & a metal ‘Meccano’ looking strip that a padded bracket screws on to, as cannot see it when installed unlike the pillar or I/O plate mounted ones).
Looks like the customer or someone else other than the factory tried that as PNY (& the other cards that come from same factory) do not use double stacked pads or the green pads. every PNY card I have taken apart to repaste/change pads, all have had plain grey pads & minimal paste applied (mostly not enough with several clean areas on both core & heatsink contact plate along with no pump out of the factory paste in the ‘ditch’ on the GPU core substrate. the anti-tamper sticker on the ‘X’ bracket screw is easily removed & replaced with a pair of thin nose tweezers & a little care at least 3 times before it starts to not stick or the edges crinkle.
Experience leads me to PSU is starting to fail. The card is not getting enough power and is shutting down. The other possibility is the motherboard PCIE slot is failing when the system warms up. The PSU for a RTX 3080 Ti is 850w, this is size I would recommend. This will allow for the cards power spikes. 🤓
I bought a new PSU to my son computer(850w corsair), because is corsair 650w was crashing with a 1080ti, 32 GB DDR4 and Ryzen 2600x, 3 hard disk, 1 SDD, 1 NVME, four 140 fan´s. This was after hi, swapped is 980gtx with the 1080ti. Fun fact, i thought the 1080ti was also bad, because, never imagined that 650w could not be enough for all the hardware he have in is computer, so i bought a new gpu (6750xt). the 1080ti is still good.
Question: do you have a flat rate for fixing cards? If so I guess there's probably a question that always needs to be asked and that is: Does the cost of the repair exceed the cost of replacing it with a new or better card? I certainly would just throw away or donate cards that were over a certain age or cost level rather than pay to fix them.
Hello. I love your videos, and I've began learning everything I can about GPUs over the last couple of years. I was wondering, what if you hot spot on your GPU is 93-98C but you GPU is always showing a stable temp of 68-74C max
I always watch this while eating at my break at work, now my coworker think I know how to repair electronics, even tho I dont know shit. But I managed to repair some stuff and now I get even more requests at work, is this how it starts ?
Can anyone name the the true meaning (name) behind Pax? Google just sucks delivering the meaning of pax for electronics/ graphic cards. The only thing i'm able to find is the avionic explanation of pax which i'm not interisted in.... Thanks ❤
PNY and Pine Technology are not the same or related in any way. PNY is a New York (now New Jersey) based business which makes GPUs (NVIDIA exclusive board partner), memory, storage, and other PC/video accessory products and Pine Technology owns XFX (an AMD exclusive board partner) and owns other small subsidiaries but no other well-known brands.
@@BeepBeep2_PNY actually stands for “Paris New York” where their main offices are based. Original office was the Paris one (so I was told by PNY Support when I asked which one was the first & main office). All PNY GPU’s are made using Nvidia’s reference design for the cards and made in the same factory & on same production line, as Palit, Gainward & some other brands, with the only difference between the actual cards being the VBIOS used, screen-printing on the PCB itself & any stickers placed on the cards. Everything else is identical between the cards using the same GPU cores, VRAM IC’s, SMD’s, capacitors, VRM’s, etc, etc that come off that factory line. This was also confirmed to me by PNY Support in France, but they would not confirm/deny that using a Palit or Gainward VBIOS for same PCB revision wouldn’t cause any issues (as if they did that & there was any issues, then they would be liable for telling customer that), as the Support Team member emailed me from their personal email account to tell me that rather than the official PNY one so they wouldn’t get into any trouble (& will not disclose their name either so they keep their job)
Only problem is that PNY are not an “American” company, but mostly French using a Taiwanese factory to knock them out! PNY stands for ‘Paris New York’ where their offices are (main one is in France). All of their GPU’s use the basic Nvidia reference designs as well so they basically do nothing apart from buy the cards & VBIOS, and put them in their own branded boxes. This information was given to me by PNY’s Support team from the French office.
Dude probably has an underpowered power supply! I've dealt with an issue like that trouble shooting a friends GPU that would randomly crash. Then out of curiosity, I looked at his power supply. It wasn't adequate for his system. We ordered a new one for him and installed it a few days later, problem solved.
Those random crashes might be caused by transient power spikes that these higher-end 30 and 40 series cards tend to cause. For 100 microseconds or thereabout these cards will draw 2-2.5x the usual wattage and force the PSU to restart/shutdown. This problem manifests when all the resources of the card are deployed (so for example when playing on high graphical settings with ray tracing and DLSS). I've been struggling with this problem on an EVGA RTX 3080 playing Cyberpunk in 4K high/ultra + RT + DLSS. The card is sweating like crazy but nothing really happens for 30 minutes or even 2-3 hours, and then suddenly the machine shuts down. I've sent my PSU back for RMA, but they prob won't find anything. Gonna have to buy a new 1000W+ and 80+ gold PSU (current one is just 850W 80+ bronze).
FYI for anyone reading, RMA on the PSU didn't find any faults, so I bought a 1200W 80+ Gold PSU from be quiet! and that solved the issue. Not a cheap solution, but at least there is a solution.
@@The101damnations it was an FSP Hydro K 850W, so quite low tier. I think that's not the whole story though, because when running games other than Cyberpunk (no DLSS or RT), the issue didn't manifest. Also, strangely, this EVGA RTX 3080 FTW3 I have has 3 8-pin power connectors, while all the other 3080s I've seen only have 2. So I believe this card draws more power under full load when all its resources are being utilized compared to other 3080s. It's just a theory though, I might have gotten away with a smaller high-tier PSU too, I don't know.
As that is a C tier power supply, that could explain why the transients were too much for it. That EVGA card is meant to be an overclocking beast and probably has really high power limits, which may have been the killing blow for your old psu. Undervolting my card to 0.9 mV @1920 mhz keeps my avg power draw under 300w for a small performance gain over stock. Maybe that could have helped you earlier. But good thing you have 1200w now and don't need to worry about it.
@@The101damnations Yeah, I was thinking about undervolting, but then I thought I'd bought this card to be able to use it to its fullest potential and then came the idea to really start going above and beyond in terms of wattage. Also, I have a be quiet! now with wasn't cheap so I hope I've bought a bit of future proofing with that high wattage too.
What is wrong with this PNY? It's A PNY. Every single PNY product I have ever owned has shizz the bed within its warranty period. I will never own another piano product ever.
As someone who already mentioned here, PNY doesn't make the RTX 3000 series cards but Palit, same for Gainward so you might rethink that for these PNY rebranded cards.
@@monkeyabout1297 Yes. Gainward and PNY at least for the RTX 3000 series are all manufactured by Palit, so the 3 share the same cooler with just some difference in designs on the front.
From what you said at start of video (typing this halfway through)with the fault described it sounds like the guy has Resizable Bar enabled in his BIOS as that is what these cards will do if it is enabled. I have been having a back & forth email ‘argument’ with PNY over this as I own the exact same card. They have VBIOS updates for all their cards apart from the 3080ti to enable ReBar support & according to their support, PNY have no intention of releasing an updated VBIOS to allow use of ReBar with these cards. They refuse to give an answer to why this is the case. Apparently the Palit VBIOS for their 3080ti will enable ReBar support but I am unwilling to do this with the price I paid for the card in May 22 (when the prices started coming down). They also say that no matter what the case is they will not honour any warranty if the stickers are damaged even though it is illegal to do that in most countries! When I enabled ReBar with my card, it would refuse to display anything for around an hour each time even after ReBar was disabled in the BIOS again. That made me think that they had some sort of temp killswitch built in or the capacitors held a charge for a while after booting (as leaving the card unplugged for an hour would mean it worked again). The card also runs hot even at stock settings although you can get a small OC on the card (+175 core/+450 VRAM) in Afterburner without the temps going off the scale (5-6°C increase). Anything more is completely unstable regardless of fan speeds. After the experience with PNY, I won’t be buying another one of their cards (last 2 cards were from them - 2070 & the 3080ti). Once the warranty has completely expired (will have to check exactly how long it is as PNY told me it was still covered even tho had it over a year), I will change the thermal paste & pads as I have a feeling that these will be part of the problem along with the generic cooling system they use. Even the power level slider is locked to 100% in Afterburner as they have configured the VBIOS to do so (PNY’s words) to stop people from having overheating issues. I just wish there was more 3rd party cooking solutions for GPU’s nowadays (non water cooling) as these cards could be improved greatly with a decent cooling solution fitted & a VBIOS that wasn’t so restricted. I am wondering if enabling ReBar in the PNY VBIOS caused overheating issues hence why they refuse to release an ReBar enabled version for the 3080ti? I even asked for a Beta VBIOS! PS, strange to see such a good retaining X plate on their CHEAPEST version of their cards!!!
They don’t use those green pads on all their 3080ti GPU’s. My PNY 3080ti used just plain light grey pads (until I changed them for cheap 18W/kM rated Temu pads - which work very very well!!!). What also gives it away is that some of the pads were double stacked (which I have never seen on any card with factory pads!). The anti-tamper sticker proves nothing either as they are easily lifted off the ‘X’ bracket screw & replaced like factory again with a pair of very thin tweezers & a little care.
My RTX 4070 face the same artifact and flickering problem as the AMD card you mention on chrome, especially youtube, turns out it's just a chrome bug that hasn't been fixed.
@northwestrepair I have a gigabyte gaming OC 3080 ti that only crashes in-game(apex legends, the FINALS) if you enable resizable BAR. Is there an issue with the gpu or is resizable BAR a problematic feature?
lead-free solder + ball grid array = match made in hell. Has caused nothing but problems since it's inception.
I used to work in PCB assembly plant, and they stopped using lead solder because of the health issues that they could cause to the operators. Some bits were manual soldering and other bits were done by machines.
00:03 Fixing a 3080 TI from PNY that broke while idling
03:12 Impressed with the ease of disassembly
09:16 Discussion on the components and diagnostic process of the 3080Ti
11:56 DIY modification using sandpaper to make M3 standoff fit M2 card holes
16:58 Swapping graphics card to 3080 Ti in BIOS setup
20:17 Troubleshooting and testing GPU performance for artifacts
26:32 Assembly of the 3080Ti is smooth and enjoyable
29:03 Achieve optimal performance by ensuring hot spot and memory temperatures are similar
34:38 Showing a 3080 Ti with LED lights
37:03 Sharing information about 3080Ti fix
You should never work for free, period. I haven't seen a single YT "fixer" that offers no fix, no fee. There is a well known repair guy in the UK that charges minimum £60 per hour, plus components. This is highly skilled work, (if done properly), and your work is first class. Don't sell yourself short. Start charging a minimum price, just for looking at these things, then charge per hour. Your videos speak for themselves. Keep up the excellent work Tony 👍👍👍
Monetized Video with over 13,000 views so far... he isn't working for free.
Doesn't louis rossmann do no fix no fee? Or at least that's what I saw in one of his older videos.
@@rateater420he charges shipping and such but lewis runs a company that makes decent cash, big difference from just losing your job and hoping fixing video cards on UA-cam works out. But no fix no fee really doesn't hurt the business too badly as long as they pay the shipping.
Hmm.. I welcome the challenge; I'm quite confident that I can locate a UA-cam repair engineer who offers free services on an Nvidia GPU board.
Just allow me a few mins, I'll be back, while waiting hit my S-button. 🙄
Impress myself sometime, I've got a handle on a guy by the name of, NorthridgeFix.
In a video about Nvidia GPUs released five days ago with this UA-camr highlighted that certain users received both a cost-free repair and free shipping.
The person who created those connectors that come off must be the same person who created USB 3 headers on motherboards LOL
lol right
whatever you do, the connector easily snap
Your videos have literally inspired me to learn soldering. Picked up a cheapish starter kit and plan on tearing down and practicing repair on some of my old parts ill never use again. Keep up the A+ work, and i greatly admire your incredibly honest and fair business strategy. Especially these days when times are tough for the bulk of us. I see people shaming you for not charging enough, but don't seem to realize not everyone needs to be wealthy. Just enough to continue our passions and live without stress.
I had one customer who said their computer screen would go black, just by grabbing any window, and shaking it on the desktop. Sounded really odd to me. Then i saw the PSU Geek Squad put in his PC. It was an Apevia 500 watt. His problem was it couldn't handle 300 watt. The 1080 video card was starving to death, and even 2-D was enough to frack everything.😂 If all i had was a 1080 staring up at me, i would not have found anything wrong.
I find weak or insufficient power supplies in alot of computers I fix, just had an old vega 64 and it was running on a 550 watt, customer said it went black screen and the gpu fan would go 100%. 😂
i personally have an apevia 520w and i don't see how the hell it crashed. the 1080 draws 175, and any ok CPU draws anywhere from 95-135 watts, which is far under the power draw limit. I personally used an rx 580 on my PSU that draws way more than a 1080, and it ran just fine.
@@gargolgaming8101 The one I worked on had a vega 64 and a 5700g, wall power only registered 440 or so tops, 550 watt p.s. which happened to be an apevia oddly would run fine under less demanding games and black screen and fan would go 100%, replaced with a 750 and it runs fine. Power supply tested fine, I put it in my potato test computer with a 5500xt.
@@gargolgaming8101 Simple. Inline watt meter didn't lie. I changed nothing else but his PSU. I flipped him into a used Rosewill semi-modular 750, for 20 dollars, and didn't charge for install. That solved the problem. No more crashing. The apevia 500w had a strange smell, like bad clutch pads. If you got lucky, well enjoy it.
@@gargolgaming8101 the Power on a PSU is combined power on all Rails, even 5v Rail so maybe ur 12V rail has just 300w
I'm pretty sure Greta was around 3 years old when RoHS was enacted.
And the issue isn't lead-free solder _per se_ (of which there are multiple types - some significantly better than others), it's the fact that some manufacturers try to use the cheapest one they possibly can.
Leaded solder is still allowed (even for industrial production) as long as manufacturers prove that (even good) lead-free solder isn't as reliable for their application, and commit to collecting and recycling the products they sell. A lot of medical equipment still uses leaded solder. Manufacturers who just want to sell their stuff and let others deal with the waste are (and should be) required to use lead-free.
LIke he said, unfortunately she's not the real source either. She's another consequence. You're not allowed to name the source on these platforms because they're so powerful that you can't talk about how powerful they are, while telling us that they have no power. ;) I believe Voltaire said something about that.
@@VndNvwYvvSvv - Is it the lizard people? It's the lizard people, right? I knew it! Wait, what's that hissing behind me...?
@@RFC3514 LizardPeople, delicious
"A lot of medical equipment still uses leaded solder."
Don't know how it it in USA but in Europe there is strong rumors it will be a ban for med equipments as well. We use lead-free for everything anyway.
Thank you for making these videos man. It gives me a lot more confidence working more in depth in my own PC.
A future client appears
@@TinMan445Not necessarily. I learned how to build a PC by simply reading and looking at pictures. I have never messed up anything but then again I know to be very careful and take certain precautions, also not to f**k with stuff if things are already working properly.
These types of videos can be invaluable for learning, just you know some people shouldn't open things up or mess with stuff as they just don't have the patience or delicateness at all.
Been watching your videos for a while, more entertaining than some other content creators with "north" in their name.
I know nothing about electrical repair, but I find your videos soothing and enjoy watching them. Just subscribed. Thanks for all the content.
XLR8 is pronounced as "accelerate". Cool name because it takes us back to the wildly successful Megas XLR cartoon which unfortunately got canceled.
That EVE Online add on youtube brings back so many years of memories
NW repair is just a bro helping out a bro. I have mad respect
Thankyou for making these videos my brother. I am a repair technician in Zimbabwe , and although we don't have these high power cards here, these are increasing my confidence when fixing the slower/cheaper cards.
Ty for the new found information about the hot spot and the memory running at the same temperature and that there shouldn’t be 3 deltas on the temp and if you do have it then that’s a sure indicator that you need to replace the thermal pads on the memory. And while your doing that clean and reapply thermal paste to the core. I’ve always cracked them open . Never dawned on me to use the software side of things to identify worn out thermal pads. But in my defense back in late 90’s and up to 2012 there really wasn’t that much software available or accurate for such diagnosis. Either way TY sir for my new found knowledge. There was much more that I learned but thermal pads was the easiest to share. 🍻
mine any kawpow algorythm with it, that cranks the core and memory to max. If that passes, and no visual issues in win, i'd consider it perfectly fine. Users have messed up systems, dirty dusty/greasy pcie connectors with dog fur and halfway plugged 8pins, but you are super precise to approach even a perfect card from the perspective that it might be shorted. Respect, i learned a lot again :)
i have been hear from the beginning . i mean abou3k subs. and i have watched every video. i'm repairing laptops but very interested in gpus . love to see your channel grow.
on cheap cards, the memory can go higher than hotspot temps because the memory chips aren't cooled with any cooler. anecdotally, memory chips without coolers can perform as good or better vs memory chips attached to coolers. Some cards with inadequate or weak coolers just add more heat to the memory chips
Just a little tip on these cards that you know is randomly crashing boot the card 3amp current limited if the draw is normal go straight to testing before you open it up.. In my experience in the case of cold solder joints flexing the board can temporarily fix the problem and take way more time than necessary.. Also could just be a thermal issue that could be easily diagnosed before you tear it down.. Personally I never take a card apart until it's power tested and if it passes then I'll test on the test bench to determine what the problem is.. In my opinion it's a way better than the route than you've been taking lol.. Like i said it just my opinion and everybody does things differently.. Anyways I enjoy the videos keep up the good work..
Will 17 likes do it ?
Trouble shooting step #1 no matter what is being tested is test as is to see the problem yourself... or not. Test drive the car or test the GPU? Doesn't matter, it should happen as is before any tools are out.
You do realize he is doing it not because he enjoys extra work , but for the show?
I also always start with a mild diagnosis and then gradually increase the level of diagnosis, but I enjoy watching this video so I can get to know more about the contains of various graphics cards.
Power testing first is a bit like just plugging in everything in your house to the same outlet. Eventually something bad is gonna happen lol
Have you ever considered expanding outside of just GPU repairs? It seems like your advanced knowledge of circuitry would make repairing other electronics possible, such as older consoles, Mobos, and what not. You'll definitely stay busy, but more options should bring in more money, especially since you do amazing work.
He actually does have a few videos repairing motherboards. Last one I saw he was fixing a broken DIMM slot.
Nice to see the same model that I own on your hands, and discovering good and bad things about it. Thanks as always for your video! Of course, thanks too to your funny comments, they make may day ;)
RTX 3xxx PNY, GAINWARD, PALIT is the same. Only diferent stickers on fans.
I heard same thing.
These brands are not common in the US by the way. More so in Europe
100% correct. The PNY cards are the same/identical as the Gainward, Palit & several others that use the Nvidia reference design sent out to the OEM’s. Made in same factory (obviously) & only differences are the VBIOS & stickers. I recently took my PNY 3080ti card apart to repaste/change pads & the X grace screws were literally able to be turned by light finger pressure! Once I had changed the pads/paste (used 18W/kM pads from Temu!!! & Noctua NT-H2 paste), the core temps reduced by 22C & fans speed did not ramp up automatically to 60% when on auto & now barely break 40% when gaming for a while. Hotspot delta was approaching 20C difference to the core before, but once reassembled, it is now 10-12C instead (would prefer 8-10C but happy enough). The fan connectors are a pain in the posterior I agree (I used a pair of very fine tipped tweezers to remove the 2 fan connectors before taking the HSF off but the LED connector (for the GeForce illumination) came off quite easily in comparison! The paste coverage (from new) on the core was not sufficient either as was a lot of ‘clean’ areas on both the core & contact plate where there was no paste contact & no ‘pump out’ either. Was either a Friday afternoon card or the factory needs to increase the amount dispensed onto the core/contact plate.
Funnily enough, I've been buying Gigabyte boards for 25 years (my first Gigabyte was a Riva TNT2 M64), and I've never had any issues with them. They all worked flawlessly and still work to this day (yes, I still have them, I'm a collector), although, after all of these years, I did have to replace some capacitors due to the capacitor plague from '99 - '05. The only card that died on me was an MSI GTX 760 (artifacts consistent with faulty memory ICs), but MSI was very quick to replace the card and I still have its replacement which still works after 10 years.
Gigabyte WAS sweet, back when i had a Slot-A motherboard. Now they sell motherboards that are DOA. Disappointing, would be my word of choice for Gigglebutt.
I have various cards and boards from every manufacturer and they all still work to this day. Some still running OC'd daily in family systems. I really only starting seeing a notable quality drop into the DDR4 era.
@JJFX-, all electronic devices have a 1 - 3% failure rate. That's just how it is, how it's always been, and it will continue to be true in the future as well. You can't really judge a product because you happened to receive a DOA device. I can assure you that most Gigabyte motherboards, video cards are not DOA. 🙂
@@ruxandy It could be much more than 1 - 3% in Gigabyte's case. Because I got 2 Gigabyte B650M DS3H boards IN A ROW that had a memory VRM coilwhine issue, the third one didn't though. I'm pretty sure I just got really unlucky, and I still love Gigabyte's products, but anyway, this was my experience.
@@atirta777 , coil whine is a completely separate issue altogether (and one that is very difficult to pinpoint). Since I have hundreds of computer parts, I know the issue all too well, and what I've found is that a certain device/PC component can have coil whine when used in conjunction with certain other parts, and the same device might be completely silent on another PC, with different components.
Friday morning, my old Asus EAH 5850 had a short circuit, and burned up on the PCIe bus on the electric side, my computer shut down immediately, and luckily the Motherboard, CPU, Memory and the PSU were spared, good I was at home, and could fast remove the GPU from burning further. And the PCIe slot, after cleaning, still works. Thank god. I had plans to change this GPU anyway, so now I have to speed up things. Never had a GPU blowing up that way.
You are probably flooded by work after all these videos of successful repairs , we could use someone like you here in Slovakia. Great job. I got to learn a lot about cards....
LOL I don't think Greta knew about Lead free solder. I wasn't made law until 2006, Greta was born 2003. Regardless, yes it is a pain in the arse!!
I use a RTX 3080ti and the card runs hot. So, I copper molded it and now they work perfect. I have heard the cases against copper modding. But, I did not make a slurry of Thermal Paste all over. I strategically placed it on the ram chips. The card runs at least 35% cooler.
They charge so much and still make them so cheaply because nvidia increasingly has reduced the margins for the card manufacturers. They tell them how much they are going to pay for the chips and how much they are going to sell the cards for. I think maybe because nvidia wants its founder's cards to be the best, or is wanting to take over all manufacturing eventually, or also maybe just because they can.
I have the 3080 from Pallit, and the same board, same design, memory temperatures are below the hot spot.
Added this card to my bucket list after seeing the teardown.
Hello There, it's 6am , iam in Buenos Aires, Argentina, My neighbors stores alarm was trigged at 4 am , so idk WTF i am doing , lol but love your videos
5:52 These vids are much more entertaining than I ever woulda thought 👀👏🏻
6:36 Revenge can be so sweet sometimes 😝
9:41 This is actually a real question, now we're seeing the inside of these things 🤔 and I know - because they can ✅ but hey, now we're seeing what they're actually doing 🔍👀
Hi, love the vids man, great job! on your select output problem, actually can do it via a switch, on your mobo you can force that your onboard display is always on with dedicated. then run a combiner switch to your screen so then you can switch between onboard and what ever is on the desk, i run multiple screen setups from my onboard and 3d card.
Hello Tony (x5)
Great video's! I always was interested in hardware more than software but didn't have the right equipment or patience like you. Really cool that you resurrect these cards from the dead and safe them from the junkpile ever growing larger.
Big probs, you have a new subscriber,
Greetings, The Internet
Personally, my experience is that Furmark generaly doesn't trigger random crashes. Haven't done it for some years now, but i think i remember sometimes getting crashes that would only trigger on certain conditions, like it might only trigger in one of Heaven, or 3DMark's Fire Strike or Time Spy... and then being random, having to leave each benchmark running for hours, to see if it happens... i think they might've cut down on the testing since then, something about AMD generally taking back "unstable cards" without needing tests.
¿ but i think i remember ?
What kind of English is that ?
Best you undertake an extensive reading program to learn -
and become proficient in the English language. I have taught
students age 8 who are academically more advanced than you.
Also - learn how to format text and use paragraphs.
Look in a diction-ary to learn what paragraphs are.
Learn why text formatting is important and why you will find it
in newspapers - magazines and books.
Clearly - you do not have the ability to think and analyze for yourself -
You do not pay attention to detail and are mentally lazy
You were programmed to be a ZOMBIE
ZOMBIES have eyes but cannot see - ears and cannot hear.
Your experience is limited and has taught you very little.
Read more to educate yourself to a higher level.
Good luck and good bye.
## Replies are ignored as I am too busy with other matters.
The observation remains FACT - so get the point -
Educate yourselves
I have a Gainward 3080 which is basically exactly the same as this. Idle temps are 49(GPU), 62(memory), 63(hot spot), load are 79(GPU), 86(memory), 94(hot spot). Hotspot decently higher than memory. Also my 5600x was hitting 95C under load with the stock cooler. Now with Peerless Assassin around 60C load. I imagine it's because I live in Australia where even during winter I have a fan on in my room, so ambient temps ain't great.
I have a Zotac 3080 12gb, that i think have nice temp: idle: GPU: 41ºc hotspot: 47.5ºc mem: 48ºc , with load: 69ºc-70ºc, hotspot: 79-80ºc MEM:81-82ºc, never been open, i live in Portugal, summer is very hot, today specially (39ºc outside) - Lisbon, no AC at home, maybe 27º to 28ºc at home.
Hello! Back from the UK! Your channel is amazing!!!!!
I don't know if that card produced a crash off screen while testing but if I have to guess, it did not.
Anyway, I had some similar issue years ago with an pov gtx 570 and an older enermax psu. While idle it worked fine, when gaming, I had to limit every game to 60fps with vsync otherwise instant freeze. Turned out after swaping card, the psu simply was dying, a bunch of caps were swollen. Changed psu and put back card, everything worked fine. I bet, who ever owns this card, maybe has psu issues or a weakening board, as in idle the gpu pulls most power from the pcie slot. Maybe the psus 5v rail is weak or one of the 12v. I would suggest, just for testing, try different psu, doesn't need to have 800+ watts, just let pc with swapped psu sit on desktop. Wait if crash or no.
0:11 You got me. Hello
What is the video card plugged into at 14:34? Is it available to buy? Do you have a list of equipment that you use to work on video cards?
You are a hero of mine. Keep up the good fight.
BTW, leaded solder has been banned for the most part in commercial products in the EU since 2006 with the RoHS guideline.
10:00 I'm pretty sure they simply forgot to remove the tape on those thermal pads.
Sad to hear what you said about Gigabyte video cards at the 4:28 time spot. I have had a Gigabyte 6750X for the past two years and have not had a single problem with it. What I do hate about video cards is that damn tab on the bottom side of the card that I believe is supposed to be some sort of latch to secure it to the mobo, but that is not the problem, what the problem is that movable latch on the mobo. And all the mobo's have that latch. Remove the latch or come up with a better ideal. Sorry for the rant, but the problem is not the video card, it is the stupid latch.
Yes we can hear you in Canada 🇨🇦
Weird crash symptoms on 3080(xx)? On mine it was simply that the card was either not aligned to the PCI slot or there was something funky with the edge connector. I have a MSI3080 Gaming X Trio and it would crash for no deducible reason when under load or from cold boot about 3 months after I installed it. Was about to send it back for warranty repair but figured I'd look into "variables" some more seeing as I"ve never had a beefcake GPU card before that NEEDED a support to hold it perpendicular in a slot before!!! I fixed it by cleaning the edge connector with DeOxit spray and made sure that it was reinserted and mounted "perfectly". Never had a problem since that "repair" and follow that "repair regime" every 12 months now... Go figure, eh?
In the early NASA missions, they had women calculating trajectories, they were known as computers!
Technically the plate would be an O shield, as it has no inputs.
Lead free solder pre dates the Thunberg she's only 20 and had been a pain in the ass for less time than the ban , leaded solder is still in use in medical and military applications due to cost , lower melting piont higher wettability and increased strength
Using that tension plate is cheaper than 4 small coil springs.
The difference is you can install the plate incorrectly with the legs pointing down.
I think it should be either with motherboard bios related or Power Supply related crashes. But usually the customer doesn't know such stuff and they end up sending the card for repair.
Thank you for the videos on cleaning and fixing GPU's, keep the work going dude. I would give you five likes but UA-cam allows only one.
nice work guy! youll be a lifestyle channel soon!
What's GRETA? Here in Europe we had the infamous RoHS directive twenty years ago that resulted in a tremendous downgrade on solder quality with electronic equipment since it came into force.
to me it look that the pads where not pealed by protection foil on one side...
You deserve waay more views, likes and subs, love your videos, keep on !
As a tech, I know that not finding anything wrong is sometimes worse to report to the customer then yes I found the problem and it will be a little costly to repair it, at least they have an answer.
You gave me a flash back of the hyper 212 black that sliced my finger open big time. You gotta watch those fin stacks…
There is a quicker way to circumvent the change of choosing the display device. It just involves getting a second testbench with a second pci-e extender so all you need to do is switch the card in the slots :) Waste of money and might not be worth the investment, but in the long run I'd say it is as it saves much precious and valuable time.
Hello strange craziness back @ you! And io shield. I call it al brackets! 😅
I'm here man you scared the sh!t out of me when you stuck your head under the camera.
"Witches are made of wood"😂 no short there.
can you recommend a gpu brand , i think you fixed all brands but what are the best and reliable pcb ?
absolute wizard!!!
got enough parts from donors and salvage to cobble together a working videocard?
You should get some flat faced Fingernail Clippers for the plastic notches at 6:56 :)
Much safer!!! And probably useful for other parts as well when you just need a clean cut.
The "who cares!" Should me his slogan😂
I feel bad I got a gigabyte during the pandemic.. which is the #1 card you recommend now that EVGA is gone?
Do you think the bubbles in the heat-transfer pad might have been what changed?
I have a pny 3060ti that had such bad voltage drop on the pcie slot that i re routed the shunt to a separate 6 pin and added jumper wires all over the card. More caps only helped so much. But the card has a good core and decent memory so cant get everything i suppose
I wonder if cause of how heavy coolers are nowadays that may be the reason for reball without a decent support bracket and pcb sag? Just a thought.
Yes you have to support these new cards people don't understand that PCB isn't that strong
Krisfix is way more experienced than Northwest and he says the reason behind broken solder balls is lack of GPU bracket. He also used lead free solder, he is not against them. I would trust him rather than some tinfoil hat conspiracy.
@@hmello3250 lead free solder can be a big source of problems due to material properties, and there are many reputable scientific studies on that (some even from NASA).
@@hmello3250 Cope
Most of the GPU sag is caused by cards not having sufficient back plate fitted (if have one!). I have a PNY RTX2080 card in my wifes machine (PNY 3080ti in mine), that sags like a bitch, but on close inspection, it is the heatsink that is sagging as it is literally double the length of the PCB & not got a back plate. If the heatsink screws into the back plate in the areas where the PCB is not present, then sag is generally very low. I tested this with my PNY 3080ti (other model to the one in this video which only has lighting on the GeForce text on the side of the shroud) by checking out the sag with & without the backplate fitted & without the backplate, the PCB was level to the slot, but the heatsink was sagging a LOT. Still use anti-sag support with the 3080ti even with the backplate to be safe (I use the Lian Li support bracket kit that replaces 2 of the motherboard screws with standoffs & a metal ‘Meccano’ looking strip that a padded bracket screws on to, as cannot see it when installed unlike the pillar or I/O plate mounted ones).
So maybe all this card needed was a re-paste and good to go.
Looks like the customer or someone else other than the factory tried that as PNY (& the other cards that come from same factory) do not use double stacked pads or the green pads. every PNY card I have taken apart to repaste/change pads, all have had plain grey pads & minimal paste applied (mostly not enough with several clean areas on both core & heatsink contact plate along with no pump out of the factory paste in the ‘ditch’ on the GPU core substrate. the anti-tamper sticker on the ‘X’ bracket screw is easily removed & replaced with a pair of thin nose tweezers & a little care at least 3 times before it starts to not stick or the edges crinkle.
Lead-free solder has been since 2000 and Greta is born 2003, so don't blame her 😁
Experience leads me to PSU is starting to fail. The card is not getting enough power and is shutting down. The other possibility is the motherboard PCIE slot is failing when the system warms up.
The PSU for a RTX 3080 Ti is 850w, this is size I would recommend. This will allow for the cards power spikes. 🤓
I bought a new PSU to my son computer(850w corsair), because is corsair 650w was crashing with a 1080ti, 32 GB DDR4 and Ryzen 2600x, 3 hard disk, 1 SDD, 1 NVME, four 140 fan´s. This was after hi, swapped is 980gtx with the 1080ti. Fun fact, i thought the 1080ti was also bad, because, never imagined that 650w could not be enough for all the hardware he have in is computer, so i bought a new gpu (6750xt). the 1080ti is still good.
Hello northwestrepair!
I have a Ryzen 7 pro 4750G SOC that bluescreens on UA-cam or Twitch. Anything else, works just fine.
I use it in a Plex server now. No problems.
gamers card for sure!
Question: do you have a flat rate for fixing cards? If so I guess there's probably a question that always needs to be asked and that is: Does the cost of the repair exceed the cost of replacing it with a new or better card? I certainly would just throw away or donate cards that were over a certain age or cost level rather than pay to fix them.
Damn...after the second "Hello Internet!" I couldn't hold saying "Hello!" to my cellphone XD
Hello. I love your videos, and I've began learning everything I can about GPUs over the last couple of years. I was wondering, what if you hot spot on your GPU is 93-98C but you GPU is always showing a stable temp of 68-74C max
24:50 why do you clean your, fresh paste, isn't that wastin time?
I always watch this while eating at my break at work, now my coworker think I know how to repair electronics, even tho I dont know shit. But I managed to repair some stuff and now I get even more requests at work, is this how it starts ?
600 frames on the donut I'M ROLLING ON THE FLOOR🤣
Which manufacturer makes the best gpu by design/repairability?
Can anyone name the the true meaning (name) behind Pax? Google just sucks delivering the meaning of pax for electronics/ graphic cards.
The only thing i'm able to find is the avionic explanation of pax which i'm not interisted in....
Thanks ❤
PNY is a board partner of NVidia just as BFG and Sapphire brands are. Pine Technologies (PNY) also owns the XFX performance-oriented brand.
PNY and Pine Technology are not the same or related in any way. PNY is a New York (now New Jersey) based business which makes GPUs (NVIDIA exclusive board partner), memory, storage, and other PC/video accessory products and Pine Technology owns XFX (an AMD exclusive board partner) and owns other small subsidiaries but no other well-known brands.
@@BeepBeep2_PNY actually stands for “Paris New York” where their main offices are based. Original office was the Paris one (so I was told by PNY Support when I asked which one was the first & main office). All PNY GPU’s are made using Nvidia’s reference design for the cards and made in the same factory & on same production line, as Palit, Gainward & some other brands, with the only difference between the actual cards being the VBIOS used, screen-printing on the PCB itself & any stickers placed on the cards. Everything else is identical between the cards using the same GPU cores, VRAM IC’s, SMD’s, capacitors, VRM’s, etc, etc that come off that factory line. This was also confirmed to me by PNY Support in France, but they would not confirm/deny that using a Palit or Gainward VBIOS for same PCB revision wouldn’t cause any issues (as if they did that & there was any issues, then they would be liable for telling customer that), as the Support Team member emailed me from their personal email account to tell me that rather than the official PNY one so they wouldn’t get into any trouble (& will not disclose their name either so they keep their job)
I have had asus do it too with those terminals.
PNY makes great cards- well at least from the 30 series and later-
One of the only American company's left
Only problem is that PNY are not an “American” company, but mostly French using a Taiwanese factory to knock them out! PNY stands for ‘Paris New York’ where their offices are (main one is in France). All of their GPU’s use the basic Nvidia reference designs as well so they basically do nothing apart from buy the cards & VBIOS, and put them in their own branded boxes. This information was given to me by PNY’s Support team from the French office.
Dude probably has an underpowered power supply! I've dealt with an issue like that trouble shooting a friends GPU that would randomly crash. Then out of curiosity, I looked at his power supply. It wasn't adequate for his system. We ordered a new one for him and installed it a few days later, problem solved.
Those random crashes might be caused by transient power spikes that these higher-end 30 and 40 series cards tend to cause. For 100 microseconds or thereabout these cards will draw 2-2.5x the usual wattage and force the PSU to restart/shutdown. This problem manifests when all the resources of the card are deployed (so for example when playing on high graphical settings with ray tracing and DLSS).
I've been struggling with this problem on an EVGA RTX 3080 playing Cyberpunk in 4K high/ultra + RT + DLSS. The card is sweating like crazy but nothing really happens for 30 minutes or even 2-3 hours, and then suddenly the machine shuts down. I've sent my PSU back for RMA, but they prob won't find anything. Gonna have to buy a new 1000W+ and 80+ gold PSU (current one is just 850W 80+ bronze).
FYI for anyone reading, RMA on the PSU didn't find any faults, so I bought a 1200W 80+ Gold PSU from be quiet! and that solved the issue. Not a cheap solution, but at least there is a solution.
@@lecarrotman What model was your earlier PSU? I have a MSI A750GL and it works fine for my 3080. But it is a tier A PSU on the PSU list.
@@The101damnations it was an FSP Hydro K 850W, so quite low tier. I think that's not the whole story though, because when running games other than Cyberpunk (no DLSS or RT), the issue didn't manifest. Also, strangely, this EVGA RTX 3080 FTW3 I have has 3 8-pin power connectors, while all the other 3080s I've seen only have 2. So I believe this card draws more power under full load when all its resources are being utilized compared to other 3080s. It's just a theory though, I might have gotten away with a smaller high-tier PSU too, I don't know.
As that is a C tier power supply, that could explain why the transients were too much for it. That EVGA card is meant to be an overclocking beast and probably has really high power limits, which may have been the killing blow for your old psu. Undervolting my card to 0.9 mV @1920 mhz keeps my avg power draw under 300w for a small performance gain over stock. Maybe that could have helped you earlier. But good thing you have 1200w now and don't need to worry about it.
@@The101damnations Yeah, I was thinking about undervolting, but then I thought I'd bought this card to be able to use it to its fullest potential and then came the idea to really start going above and beyond in terms of wattage. Also, I have a be quiet! now with wasn't cheap so I hope I've bought a bit of future proofing with that high wattage too.
its probaby the cheap caps on the back common issue during the shortage
What is wrong with this PNY? It's A PNY. Every single PNY product I have ever owned has shizz the bed within its warranty period. I will never own another piano product ever.
As someone who already mentioned here, PNY doesn't make the RTX 3000 series cards but Palit, same for Gainward so you might rethink that for these PNY rebranded cards.
I have a pny gtx580 that's been overclocked to over 900mhz and it works fine still.
I've been using PNY since the Pentium/K6 era (mainly SDRAM). The latest was a 1660 Super. No problems at all.
@@JCmeister9so what you are saying is that this PNY is the gainward phoenix. Sad BC I really love the look of the gainward phantom . Such sexy cards.
@@monkeyabout1297 Yes. Gainward and PNY at least for the RTX 3000 series are all manufactured by Palit, so the 3 share the same cooler with just some difference in designs on the front.
From what you said at start of video (typing this halfway through)with the fault described it sounds like the guy has Resizable Bar enabled in his BIOS as that is what these cards will do if it is enabled. I have been having a back & forth email ‘argument’ with PNY over this as I own the exact same card. They have VBIOS updates for all their cards apart from the 3080ti to enable ReBar support & according to their support, PNY have no intention of releasing an updated VBIOS to allow use of ReBar with these cards. They refuse to give an answer to why this is the case. Apparently the Palit VBIOS for their 3080ti will enable ReBar support but I am unwilling to do this with the price I paid for the card in May 22 (when the prices started coming down). They also say that no matter what the case is they will not honour any warranty if the stickers are damaged even though it is illegal to do that in most countries!
When I enabled ReBar with my card, it would refuse to display anything for around an hour each time even after ReBar was disabled in the BIOS again. That made me think that they had some sort of temp killswitch built in or the capacitors held a charge for a while after booting (as leaving the card unplugged for an hour would mean it worked again).
The card also runs hot even at stock settings although you can get a small OC on the card (+175 core/+450 VRAM) in Afterburner without the temps going off the scale (5-6°C increase). Anything more is completely unstable regardless of fan speeds.
After the experience with PNY, I won’t be buying another one of their cards (last 2 cards were from them - 2070 & the 3080ti).
Once the warranty has completely expired (will have to check exactly how long it is as PNY told me it was still covered even tho had it over a year), I will change the thermal paste & pads as I have a feeling that these will be part of the problem along with the generic cooling system they use. Even the power level slider is locked to 100% in Afterburner as they have configured the VBIOS to do so (PNY’s words) to stop people from having overheating issues.
I just wish there was more 3rd party cooking solutions for GPU’s nowadays (non water cooling) as these cards could be improved greatly with a decent cooling solution fitted & a VBIOS that wasn’t so restricted. I am wondering if enabling ReBar in the PNY VBIOS caused overheating issues hence why they refuse to release an ReBar enabled version for the 3080ti? I even asked for a Beta VBIOS!
PS, strange to see such a good retaining X plate on their CHEAPEST version of their cards!!!
I agree with mr. Wizard there is nothing wrong with the card but there is something wrong with the customer's system, that's for sure I say.
Sanding the standoffs down is a good idea.🙂
Glue the loose fan header to the board 🤣🤣
Damn, PNY using Kritical pads, they’re green like Kritical pads anyway
They don’t use those green pads on all their 3080ti GPU’s. My PNY 3080ti used just plain light grey pads (until I changed them for cheap 18W/kM rated Temu pads - which work very very well!!!). What also gives it away is that some of the pads were double stacked (which I have never seen on any card with factory pads!). The anti-tamper sticker proves nothing either as they are easily lifted off the ‘X’ bracket screw & replaced like factory again with a pair of very thin tweezers & a little care.
My RTX 4070 face the same artifact and flickering problem as the AMD card you mention on chrome, especially youtube, turns out it's just a chrome bug that hasn't been fixed.
@northwestrepair I have a gigabyte gaming OC 3080 ti that only crashes in-game(apex legends, the FINALS) if you enable resizable BAR. Is there an issue with the gpu or is resizable BAR a problematic feature?
yes i am watching you hello you 2
Great work, I wish I could do the same