2024 UPDATE! Watch our new investigative deep-dive into this cable standard! ruclips.net/video/Y36LMS5y34A/видео.html Sorry it took so long to put this all together! Thank you so much for the interest and patience the last few weeks on this story! USE CODE 'THISISFINE' on the GN store for 10% off through 11/23/22! store.gamersnexus.net/ - grab a GN coaster pack, PC building modmat, mouse mat, shirt, or more to support our work and get something quality in return! Watch our first video in this series! ruclips.net/video/EIKjZ1djp8c/видео.html New GN 3D Coaster Pack: store.gamersnexus.net/products/gn-drink-debug-coaster-pack-4-custom-3d-coasters-100x100mm-4x4 GN PC Building Anti-Static Modmat: store.gamersnexus.net/products/modmat-volt-large GN Red & Black Mouse Mat: store.gamersnexus.net/products/gn-red-black-hud-mouse-mat GN Blue & Black Mouse Mat: store.gamersnexus.net/products/gn-wireframe-mouse-mat
Quality takes time Steve. Plus backing up your claims isn't something you rush in the grown up world. Carolina's Finest! Hey by the way would love to see you check out places like Red Hat here in downtown Raleigh and maybe check out EPIC in Cary.
Thank you so much for your dedication and thoroughness. Everyone’s work at GN is very highly appreciated! I can’t think of another channel as doggedly pro consumer as GN.
Can't wait for LTT to make 10 times as much money off this than GN does by doing a talk to camera commentary piece using this video, since their much vaunted "labs" appears to lack the resources and manpower to do this sort of testing!
Is it possible that no one notices that this connector is badly engineered? adapters have nothing to do with it if it melts the connector close to the pcb, it would have been much better to have 3 8pin connectors
I’m no cable expert, but in regards to not properly plugged in this should be easily avoided if a single signal pin was shorter and not allow current to pass through unless it has contact..
Correct! They're looking into that for a revision, as we understand it. That's a collaborative effort involving NVIDIA and people at PCI-SIG to ratify it.
@@GamersNexus This is why UK power plugs have a shorter main connecting pin, so if a kid accidentally doesn't plug it in or unplugs it partially they don't get electrocuted and/or the device doesn't draw power. Funny how multi billion $ companies and cable experts couldn't anticipate such a simple eventuality. I'd be interested in a video as to why more cables aren't idiot proof, like USB C for instance, or better fully inserted functionality on a GPU power cable etc
@@Daniele858585 torquing the cable to the side ( common because of side panels ) also pulls the cable from the socket and a short signal pin could prevent it from getting a current if over torqued
Great job Steve and the GN team. As an engineer myself that has had to do substantial failure analysis like this, people do not realize the amount of time and effort that is spent trying to recreate failure modes. Consulting with experts and the trial and error involved wit reproducing failures all take a ton of time. Most people, even engineers, also struggle with jumping to conclusions and not following the data to consistently reproduce a failure. This often results in fixing nonexistent issues and missing the actual failure mode. Excellent, excellent work.
Also because it's hard for smart people to create dumb people problems. Sometimes the best testers are the absolute f******* numpties, rather than the super-smart engineers. Leave the latter to solve the problem.
As a fellow Engineer. 100% agree. GN is the goat for taking this on and doing proper failure analysis. I think the others are ok putting out their thoughts on it, but should have a disclaimer that goes with it indicating that it is largely an educated guess.
@Practical Crafting congrats ? yes its been known for a while but this will spread the news like wild fire and protect EVERYONE who might have had a wind fall and built their first PC .... Just because you know about it and the majority of IT does... You are saying it shouldn't be shouted from the rooftops? Wtf is wrong with you...
I was very surprised that when I just got my 4090 after this video, I plugged it in fairly hard, looked flush, and still managed to pull it out easily with a little wiggle. Glad I was able to catch this and prevent a potential burnt GPU after months/years of use and the occasional cleaning that might have loosened the 12 pin. HUGE!
@@zixerk3167 Yea, I pushed it down until it felt like it would not move any further, this was with normal plus some extra force because I saw this video ahead of time. Then I wiggled it and it came loose, so the second time I did the exact same thing but made sure to push down specifically for the latch side of the plug to make sure it hooks.
The fact that y'all guys took so long to release the video just to make sure that there's an actual problem it's something i really appreciate, love you guys
@@nuhashsarker5750 I unsubbed and just downvote jays videos now. Fluff and downright harmful rumor spreading without proper data collection. Also fairly pompous and rude to his viewers.
Really impressed by your dedication to your investigations and finding the true answers rather than theorizing. The pc community owes you guys a big thanks for everything you've done for us over the years and always mantaining your integrity!
@@GamersNexus Any interest into researching how hard it would be to have an operating system tell you when the power supply is bad or when you have a bad cable/contact situation. Also temperature sensors next to the power plugs could be useful. (also MB) I have always found that extremely lacking for such advanced computers as we have now.
Shut up. You don't even know if they're right. Won't be surprised at all if it turns out the hardware is defective and user error has nothing to do with it.
This is my very first time donating to any RUclipsr in my 10 years watching videos, and watching the entire video, being entertained and informed at the same time is an amazing accomplishment. I am well informed about the issue and have been enlightened. You guys deserve any donations for the countless hours you guys did to bring an informative and objective light into this issue. Keep it up and I will continue watching your guys' content!
Yeah continue feeding the problem and paying for these shit practices and gpus, good job making it all more and more , not only unavailable in terms of prices, but giving in to these proprietary laughable cables / power supply practices.
This whole saga is like watching an air crash investigation episode. Good work as always and im glad to know there's someone out there actually looking into things like this
@@andredavis3350 nah, it was probably because the passengers didn't return their seat backs and tray tables to the full upright and locked position. ;-)
Hi from the Land Down Under.. i am 62, been on the internet when it started... x386.... And Amiga, etc... my mate owns company that builds gaming towers for the Aussies... And he gives me the new toys to stress out, so he can give a Warranty on the Tower Built... I have a Razer 17 Pro, i7 10th gen, 8 gig RTX 3080, 32 gig ram, and the very rare 360hz screen... $6k new... With the Thunderbolt video card box i have, i plonked in the 4090, added the extra power supply... And within just doing benchmarking test, the 4090 was toasting itself... So i moved it to my Benchmark Tower, and Yep... i was looking for eggs and bacon... to cook breakfast... It's a power grabbing fire hazard...
Reminds me of the DC10 cargo door episode (where multiple design flaws led to a door appearing locked even when its not to both pilots and baggage handlers)
Imagine how much better off the world would be if areas such as politics, corporations, and other shady entities had this level of credible, detailed journalistic investigation. Unfortunately for my pipe dream you guys have found great success in the PC sector. Feels like one million subs wasn't so long ago, and pieces like this are why you're creeping on two million so quickly. Awesome work. Glad this community has you guys.
This is the first time that I was ever felt compelled to make a donation to a content creator and deservedly so. Truly amazing work from Steve and the GN team! The community is so lucky to have you guys and this video proves that no one even comes close to the dedication that you guys have to straight up factual journalism and responsible reporting.
Damn, GN going to the level and beyond of professional engineering internal failure analysis. Huge props and the quality of your work is on par with work we do for medical device failures. Also that CT, SEM and EDS testing ain’t cheap. My guess is you spent at least around $10k to have a lab do it.
Jay and Igor really forced his hand by being so irresponsible and loud with nothing to back it up. You could see how uncomfortable he was even saying that. I refuse to even watch JayzTwoBraincells after his outburst calling people dense and ignorant if they don’t listen to him. The man is a stain on the PC industry as a whole
Steve, you continue to evoke my respect and admiration with the lengths that you will go to in order to protect consumers. I deeply admire your dedication to using scientific methods to do all your testing. However, die a little inside very time you say "theory" instead of "hypothesis" in this video.
@@vespermoirai975 what do you mean "not the crap we have today" ? Nothing has changed. You don't even have to believe me. Just go and read old news stories local or national. Shit hasn't changed. Your rose coloured glasses need to be removed.
That footage of inserting the connector was really insightful. The amount of force it takes to push it in is way more than I expected. It looks so tight only half way down!
I actually plugged my adapter into my 4090 before installing the card to make sure it was fully seated. It's a pita to get in properly and I can see why some people might not get it fully seated if they've already installed the card.
When I plugged my connector, I just hold the card from the opposite side of connector with my hand and just pressed it fully so I'm not creating stress on pcie slot. And I also avoid wiggling the cable when inserting it, so the terminals inside the connector not bend themselves creating bad contact with the pins.
I can understand where the problem comes from- a lot of people are going to hesitate before pushing down that hard on a connector for fear of breaking a pin off. If you're lucky, that's a long wait for an RMA; if not, that's $1600 down the drain.
This is the best investigative journalism I have ever seen any media outlet perform in the IT industry. Usually they would just sellout (won't name names) to manufacturers or advertisers.
@@Hydranox Sure, that still doesn't require a degree in science The connector simply doesn't have enough margin of error to handle the 450 watts put through it That's why it fries when there is an inconsistent connection If I attach 4 power cords to my TV, if the power supply can handle that, it better be an overengineered connection or something's gonna start arcing
Steve, I'm so damn glad my Patreon support money is going in to content like this. Sending this stuff to get X ray/scanned probably cost a pretty penny.
The viewer who assisted with the failure analysis lab was so incredibly knowledgeable, professional, and skilled. We were absolutely blown away by the level of depth. I'm not sure that we're at liberty to really get into detail of how it all worked (his request), but we can absolutely get into detail about how awesome it was to work with him!
@@GamersNexus setting up the cable with a set of resistance meters and plugging it to a separate socket could let you plot the resistance during insertion and test a set of cables to establish the trend.
@@eric4946 Dude nobody is going to do that. They saved 1/4 of a gram of copper to make the adapter like that. If you want them to work properly, get you modular 12v cable out of your PSU and look at the connectors there. That is how you want it designed. Maximum contact will least slackness possible.
This is probably the best video on the channel. Gives me some Air Crash Investigation vibes here, where in many episodes the media, public and random "experts" on the Internet are coming up with more and more "versions" of what happened, in the meantime the investigators are honestly doing their job to determine exactly what went wrong and how to prevent it from happening again.
Thank you for existing and always accepting nothing but the truth, for doing all the hard work of gathering the data and presenting these facts. We love you guys!
@@m4c1990 Mate, please tell me you're not complaining about a donation amount? Christ i hope you're not cause if you are....... you may want to stick your head in a fridge, eat some ice, slam the door a couple times on your head (making sure all your food is correctly sealed and veggies still fresh) and chill the F... out.
Just an FYI for you monetized commenters out there, RUclips is taking 30% of the amount you're contributing. On top of that, if you're sending the monetized comment via the Apple or Android app, they are taking another 30% cut. The creators are not allowed to mention this thanks to RUclips being a super and totally awesome company. This is why creators would rather you buy merch, because you get something in return, and the tech monopolies get nothing.
In my opinion, we are all extremely lucky to have such a dedicated and thorough group as Gamers Nexus. The videos you all are putting out are fantastic.
Agreed! ALL news media sources need to do, as Steve put it, "Our due diligence as a news source." instead of just fear mongering. I'm not an nVidia fan, but at least he got to the truth instead of hopping the bandwagon. Integrity 1 - click bait 0.
After getting an aftermarket 4090 I can totally understand why people would fail to fully plug in the cable. I decided to seat the cable prior to mounting the gpu, first push on thought it was in, checked saw a 1mm gap, so pushed much harder and it clicked with no gap. The force I had to use to get the cable to fully mount I wouldn't want to do it while the gpu is in the pcie slot... In fear of damaging the motherboard. First experience with this type of connector but never had to use that much force on the old 8 pin... They basically fell in and clicked in with gravity!
Does it seem strange to anyone else that even after all this time, the PC market still hasn't really figured out connections and ports? It's like no r&d has been spent on making better physical connectors at all. There are multiple industries where a solid connection that's easy to connect; while giving the end user complete confidence that it's indeed connected is a must. I mean hell, it's like as if appliances at home keep on falling out of the receptacle. A solution is possible.
This video was the final push I needed to buy a coaster set. Thank you, GN crew, for your tireless devotion to getting the story right before you publish.
Amazing content. Thank you for doing all the hardwork and research with all your cool people. You deserve a kudos for this piece. And thank you for basing your piece on data rather than going with the ‘fear mongering tactics’.
No video stands truer than the stance I've always taken on these types of things - "wait to see what GN have got to say, because it's a given they will do the due diligence necessary". Great work as always guys. Educational, informed, clear and methodical. Just how "reporting" should be!
Honestly i think gn videos could be used on schools to teach "reporting". Not just the due diligence they give the topic, all the research, but also the professionalism, responsible reporting and hunger for truth above all else evident throughout, not to mention clear and educational presentation of complex topics
Nvidia should be paying you MILLIONS! You saved them so much money in testing and engineering costs. While also probably causing millions in loss over deserved negative press ;) Hats off to everyone at Gamers Nexus. What an amazing journey to witness as you and your team have become the gold standard of tech news, hardware review and most impressively, enterprise level testing and evaluation.
Thank you so much for that final monologue. It speaks volumes that the entire goal of this video series was consumer safety and a focus on actually getting a result rather than falling into the trap of feeling like you need to get something out. That was pretty much the best usage of constructive criticism I've ever seen of your peers without any outwardly hard feelings. Bravo.
Yeah honestly bends are normal especially for wires. Bends dont matter they can increase risk of a failure but that risk is increased if you already made a point of failure (for example not socketing the adapter completely).
Great coverage on this issue. I very much appreciate your approach in reporting data backed theory without jumping straight to making over confident claims based on your findings. This channel just keeps gaining value as a resource and advocate for consumers in the PC hardware/tech space. Thank you!
@@aa-tx7th how do you present substantial evidence to confirm your statements in 30 seconds? I think you need a better script writer and editor for your comments
You guys are pretty much the only ones I trust any more to do solid technical reporting for these types of situations. Kudos to the GN team for being amazing human beings.
I've never seen a comment section full of donations like this video. I think that shows how much the audience appreciates these detailed, failure analysis videos thats backed by rigorous lab testing and analysis.
@@Damaged7 ? The most expensive things in there are testing equiptment that they paid for. The warehouse rent also costs more than most of the hardware. A lot of the non CPU/GPU stuff like cases they buy with their own money. They also yaknow, have to pay staff. Again staff cost more than hardware in general [how many 4090's in $20k a year, which is probably quite a lot below what staff get]. They just did something expensive for testing, on the basis that it would get paid back in donations, merch buys and views etc. Paying the testing lab for electron microscope images wasn't free.
Not all heroes wear capes! You guys go above & beyond for the community. And it can't be overstated enough how appreciated your dedication to what you do, is to us. This is the kind of ethical effort that makes all of us continue to come back to this channel. Thank you for digging into this issue.
As an electrician, I am a firm believer that the pins need to be larger and the wires should be larger as well. This would create less heat to begin with and the system would be able to handle more load with less problems. This should have been addressed early on in the R&D side of things for these 4090's...
@@janglur yeah, let's say you save some dollars per cable by making it cheaper. If you sell multiple millions (I guess) of those cables, you save lot of money.
I am an electrician too, and i say, the connector is doing his job good. As you can see in the video, even with 2 out of the 6 12V pins, it dosn't melt. People just have to know what they are doing, and make sure, that they plugged in the calble correktly. And why should the wires be the problem now?
@@marcesser4218 would you run more current through 14-2 than rated? No, it would put unnecessary stress on the wires and either melt or start a fire. Seeing how GPU's do have safeguards against this, better gauge wires are needed.
As a hardware engineer, I have been _very_ interested in the root cause of this issue, as it is something that could have very real repercussions in what I do. Thank you for all your testing and reporting about this.
Thanks a lot for this! I work at an archviz company and we upgrading all our workstations to 4090's as well as most of our clients. My team actually had sleepless nights over this, great to have it finally figured out.
I love your guys deep dives into investigations of problems, companies, and reviews. You dont just pull your stuff from other reviews and always do your own reviews and put a lot of your reviewing to test way beyond what anyone else does. Your testing to the point that some engineers didn't even take into account at the time of designing or didn't care to. When you do surface contact or the coolant used in a AIO causing failure of its pumps, I'm just blown away and as a big technology guy I love your break downs.
This video was OUTSTANDING!!!! It may well be the absolute best piece that GN has ever produced. It is truly a pleasure to be able to watch thorough and competent analysis of a problem, see your proof (not just feelings or suspicions, but hard facts) by showing your work, and then provide definitive solutions on how to fix or avoid the problem in the future. You shouldn't apologize for the timeline because thorough and ACCURATE research demands time and cannot be rushed. I don't own or plan to buy a 4090 now or in the future, but the lessons learned from this situation can be applied to all computer hardware (not just GPUs). Your work here may well save someone from a catastrophic failure in the future, just from a reminder that ALL cables (not just power cables) need to be fully seated and connected to avoid problems. This was damn fine work! WELL DONE!!!
Your commitment and dedication is nothing short of amazing. You are a gold standard when it comes to hardware testing and I look forward to more of in-depth research like this.
You're 50 dollar donation would probably have fed a few families tonight. But instead you wanna simp a youtube channel and rub their ego while you hope they like you for giving them money.
@@originalityisdead.9513 What if your 4090 turns out to be DOA? I'd rather test that with the included cable 1st than wait. I assure you, RMA and warranty claims are much more longer wait.
"Late" is not how I would define this coverage. Thorough is how I would define it. I give a thanks to Steve and the Gamers Nexus team for their in-depth testing, as always.
Man, when I heard use code “THIS IS FINE” it made me really realise how much I love GN and Steve. Amazing, industry shaping work on the side of consumers, and great humour to add some levity to a bad situation. You guys are heroes
When he first said that promo code I just bust out laughing, it is so fitting, genius! I love their geeky and nerdy side coupled with strong ethics and researched reporting.
Before you even said this was the most expensive video to produce, I was wondering how much it was costing you because of all the testing. Kudos to you guys. And even though it's uncomfortable at times, you are willing to speak the truth about what you've found. There's a reason your channel is so respected.
THIS is an RCA. Thank you Steve and team. I believe all of us 4090 owners pushed a little bit harder on that connection today to make sure it's seated. Hate to say it but... user error is the most common cause of most technical issues.
The true root cause might be a design error that led to the user error. DFMEA should have caught this failure mode, especially for high current applications that are prone to use error.
@@volatilememory9338 agreed. Just want to make it clear the design specs are for sure at fault as well. Anything standardized should be tested with rigor regarding any way a user could potentially mess something up. Being that it's 2022 and this design wasn't done yesterday, could be a product of the COVID-times lack of QA availability and budgeting. Just speculation...
@@PDXnativeXPD whats next a LED light on the cable to tell people its plugged in? is dat how dim you want builders to sound? it was a user error not plugged in properly more then likely through being to eager and not having a single clue on building.
Stop apologizing for something that isn't your fault. You did the testing and show a factual result with real time proof, anyone who has opinions or hypothesis with out proof is simply speculating. You don't have to apologize. If they get mad is on them not you. Also thank you for the time spent making this information available to everyone. Keep up the good work.
6:55 imho, at that point is not user error, but design error. If you designed the connector such as being easy to improperly connect and that improper connection makes it either a fire, or a shock hazard, you should not be designing said connector in the first place.
You all have outdone yourselves on this piece. The level of rigor around in-house work, croudsourcing, and working with external experts presents an insane amount of forensic professionalism. Seriously proud and amazed at how far you all have come!
literally the first thing that crossed my mind and what i came across a few times on reddit was “let’s wait for GN” didnt disappoint, it never does. you guys are the best in depth pc hardware channel there is.
this has aged really well, but at what point do we start blaming the connector spec and not the user, because if its that hard to plug in correctly, then most users are gonna struggle.
Journalistic integrity is why I watch you guys and not a lot of your peers when it comes to trouble shooting serious issues. Thanks for the information :)
Journalistic integrity that is not always present on this channel. Like: Hating on the price of the 4090 and not mentioning that AMD and Intel are incappable of producing a competitive product. In Marketing methods like this one get used on porpuse to provoke reactions to increase your own engangement at the cost of "Journalistic integrity". (aka taking advantage of your audience) If you have the best product on the market you can ask any price. It's how capitalism works. Even twice the price would be fine. A similar example for a situation like this is Apple: They can ask for any price, because it's an Apple product. A similar PC is A LOT cheaper, but it's not an Apple product. So they often times ask 1.5x or 2x the price of a similar speced PC and people buy it anyway. Because it's an Apple product. People would buy the 4090 even for $3000 USD. What matters in this product segment is: it's the fastest. Not the price. But GN regulary fails to mention facts like this one, to make people artifically upset.
How do you declare him "neutral" and "honest" ? based on what? he is blaming people for this disaster like if people now is trying to melt their computers, it's the most absurd conclusion ever. This and calling things "sensationalism" is the equal from MSM to other people "misinformation". Now nvidia is just "not to blame", it's just people, that only want to destroy RTX4000 series cards, they must be all AMD fanboys huh? you can put all the numbers and tools, and put some guys with white coats from the MIT but that doesn't make anything more true than actual REASONABLE conclusions.
Thank you so much for this piece!! The best Investigative work on this matter and Nvidia should pay GN for this results. I am from Brazil and 4090's here are really expensive, it's great to have some facts and conclusions after all the amazing work you all have done. Well done!
Been watching y'all for years, and this is some of the greatest, most comprehensive reporting I've ever seen in the RUclips tech sphere. Keep up the good work GN Team! Would love to see more videos with the failure analysis lab in the future.
Thank you so much for the years of support! It means a lot to us that people care like that. There are channels I watch the same way, so I know what it's like on both sides and sincerely appreciate it!
That's how I feel about my 3080 and how I imagine anyone with a 30 series GPU feels I would imagine. There's just not enough value with the 40 series to justify the upgrade.
Incredible work GN!, The level of detail, honesty, quality investigation, investment and re-investment of resources and a top team is what sets GN apart!. Every video GN does sets the standard in quality unbiased reporting and a very high bar for others. Thank you GN for continuing to deliver incredible content!
I loved how deep you went into analysis. Especially enjoyed the x-ray, EM and spectroscopy results, because I currently work at a company that makes electron microscopes and spectroscopy equipment where I put EMs together and set them up + test for customers. I am proud of working there because it helps do stuff like this among other cool things. Keep up the amazing quality of content and integrity of your reporting. You and your team are the heroes we did not deserve but needed.
Gents, excellent work. However, there is little mention of a critical factor which everyone struggles silently with: The overzealous application of 'strain relief' material and far too many strain points for this nexus design. Look no further than 9:05 in the video which would make Medusa jealous! All that black may looks pretty, but it stiffens the cabling. Couple the problem with not one, or two but four strain points, each with bulky connectors, + extra useless/unused connectors and a proven unreliable latch mechanism. Your typical ATX24 cable has more conductors than this entire Medusa mess (5 x 4 cables), and generally is engineered to ride on a single, less bulkier cable bundle. Generally, the strain relief on ATX24 is not run end to end either, allowing for greater bend a flexibility and lowering the mechanical stress on the connector. Quite frankly, each of those PCI-E cables should home run to the GPU without an adapter.
Bought a shirt to support. The testing done by Steve and the team at GN is so meticulous. Coming from an Electrical Engineering background and being a PC "enthusiast", I love the extent you guys take things to duplicate and verify issues. Your work and effort deserve more likes and subs as it's rare to find such unbiased opinion and thorough testing elsewhere. Also, please restock the hoodies.
To think these guys were once working out of a house doing mostly CPU and GPU benchmarks. It's impressive to see how far you guys have come since then! Thank you for going out of your way to provide this much valuable information for the PC-building community.
Hey Steve, Patrick, Andrew and everyone else involved in this testing, When the first pictures and videos came up, I showed them to a friend of mine. He works for a big connector-supplier in the automobile sector and designs for example ECU-Connectors. He instantly told me, that the problem is in the connection between the male and female pins, not the connection from pin to cable. His words where along the lines of "The melting starts from the tip and not from the cable". He was still quiet displeased with the connectors design from a manufacturing standpoint, but his first impression of the failure is in line with your findings. We decided against publishing our opinions, since we did not have the resources to look deeper into it and without backing, our opinions would just have added fuel to the dumpster fire. So thanks for your work! Greetings from Germany and have a great day! Simmie
Scientificism at its best. Being scientific doesn't equal truth nor logic, also what is scientific about this? Just using a computer and troubleshooting is already science. The conclusion that people is to blame because they are brainwashed into destroying their own computers or they just want to for no reason is completely insulting and irrational, no matter how much science theater you put on stage.
This was really excellent reporting. As an electron microscopist, I particularly liked the failure analysis, really comprehensive. I don't own a 4090 but I'm still checking that my GPU power cable is properly seated when I get home 😂
i would say, no excuse, running that much power thru that piece of garbage... On sensitive electronics... No way, buying unsave hardware, where nowone has the solution ) its a bad gambling that NVIDIA did in my opinion...
@@realraf1 Sorry man, there is still a difference between Frankenstein wiring, and incomplete connections. You could use fucking molex and still be fine as long as you had everything plugged in all the way.
Thanks heaps for this, I have been having massive anxiety about using my 4090 since upgrading $3.5k in Australia and I did not want to lose it from a burning connector, Avoiding playing games, limiting my graphics and watching the performance monitor to keep the watts low, This video gives me some relief in the fact that I have it plugged in all the way, clipped and not bent at all so fingers crossed I don't see any issue 🤞🤞
Excellent job as always and you're the reason why I ended up with a 5950x and 6900xt based set up last year. The cost of a single 3090ti in Australia is enough to pay for my entire build and your thorough reviews steered me in the right direction. Saved a bunch of money and have a very satisfying set up for the next few years. Cheers from Down Under!!
Some folks at Nvidia will get a good night sleep after that. I've been very disappointed with the fearmongering regarding this connector when we had so little information. In the past this type of analysis would stay behind corporate doors, so it is fascinating to see the type of tools and work done to find the root cause of an issue. It is also funny that the simplest answer was the most likely, when you know what to look for (the melting and bending marks on the side of the connectors). I really like the idea of this new connector, and I hope its sturdiness can be improved upon in the short term. Having 2 or 3 8-pins in a GPU is a signal we need a better connector and should not stay as the norm, and people just refusing to consider this new connector based on flimsy theories is not great for the development of power delivery. This type of journalism deserves all the recognitions. I'm really blown away.
Thank you to everyone for the content and as always, we hope to always receive information / reviews of this caliber from GN. Always a fan and always will be a supporter of you guys.
I am a journeyman electrician and I found that your testing was sound and your conclusions reasonable. I have just purchased the 4090 knowing full well about the connector issue and you have shown that attention to detail and workmanship is the key to a successful build. As a thought to consider, when voltage goes down, current and resistance go up. When overlooking you lower the voltage. If the manufacturer used 14 gauge wire with connectors of the same rating, this could have been avoided.
the 12VHPWR connector was junk from the start - sense pins that stuck out WAY too far allowing power when the connector was not properly inserted, TOO much power going through a small amount of contact area in a CONFINED HEAT INSULATED area being the primary issues. The newer variation on that connector DID mostly fix the first issue but made the SECOND issue potentially WORSE by INCREASING THE ALLOWED POWER LIMIT even higher. AMD has been smart to IGNORE this BAD BY DESIGN connector design, I wish the NVidia AICs would do so more of the time and particularly on the HIGHER power cards.
2024 UPDATE! Watch our new investigative deep-dive into this cable standard! ruclips.net/video/Y36LMS5y34A/видео.html
Sorry it took so long to put this all together! Thank you so much for the interest and patience the last few weeks on this story! USE CODE 'THISISFINE' on the GN store for 10% off through 11/23/22! store.gamersnexus.net/ - grab a GN coaster pack, PC building modmat, mouse mat, shirt, or more to support our work and get something quality in return!
Watch our first video in this series! ruclips.net/video/EIKjZ1djp8c/видео.html
New GN 3D Coaster Pack: store.gamersnexus.net/products/gn-drink-debug-coaster-pack-4-custom-3d-coasters-100x100mm-4x4
GN PC Building Anti-Static Modmat: store.gamersnexus.net/products/modmat-volt-large
GN Red & Black Mouse Mat: store.gamersnexus.net/products/gn-red-black-hud-mouse-mat
GN Blue & Black Mouse Mat: store.gamersnexus.net/products/gn-wireframe-mouse-mat
This is fine.
Quality takes time Steve. Plus backing up your claims isn't something you rush in the grown up world. Carolina's Finest! Hey by the way would love to see you check out places like Red Hat here in downtown Raleigh and maybe check out EPIC in Cary.
Thank you so much for your dedication and thoroughness. Everyone’s work at GN is very highly appreciated! I can’t think of another channel as doggedly pro consumer as GN.
Can't wait for LTT to make 10 times as much money off this than GN does by doing a talk to camera commentary piece using this video, since their much vaunted "labs" appears to lack the resources and manpower to do this sort of testing!
Is it possible that no one notices that this connector is badly engineered? adapters have nothing to do with it if it melts the connector close to the pcb, it would have been much better to have 3 8pin connectors
I’m no cable expert, but in regards to not properly plugged in this should be easily avoided if a single signal pin was shorter and not allow current to pass through unless it has contact..
Correct! They're looking into that for a revision, as we understand it. That's a collaborative effort involving NVIDIA and people at PCI-SIG to ratify it.
@@GamersNexus I would be really interested if this can be recreated in any shape on the old 8pin adapters :)
@@GamersNexus it's so sad that people need this :X just plug that fucking cable wtf......
@@GamersNexus This is why UK power plugs have a shorter main connecting pin, so if a kid accidentally doesn't plug it in or unplugs it partially they don't get electrocuted and/or the device doesn't draw power. Funny how multi billion $ companies and cable experts couldn't anticipate such a simple eventuality.
I'd be interested in a video as to why more cables aren't idiot proof, like USB C for instance, or better fully inserted functionality on a GPU power cable etc
@@Daniele858585 torquing the cable to the side ( common because of side panels ) also pulls the cable from the socket and a short signal pin could prevent it from getting a current if over torqued
Great job Steve and the GN team. As an engineer myself that has had to do substantial failure analysis like this, people do not realize the amount of time and effort that is spent trying to recreate failure modes. Consulting with experts and the trial and error involved wit reproducing failures all take a ton of time. Most people, even engineers, also struggle with jumping to conclusions and not following the data to consistently reproduce a failure. This often results in fixing nonexistent issues and missing the actual failure mode. Excellent, excellent work.
well said, this was excellent analysis work done by GN
Failure analysis work is so cool. The work that the FA lab that helped us did was amazing. Really impressed by them and inspired in many ways!
Also because it's hard for smart people to create dumb people problems.
Sometimes the best testers are the absolute f******* numpties, rather than the super-smart engineers. Leave the latter to solve the problem.
As a fellow Engineer. 100% agree. GN is the goat for taking this on and doing proper failure analysis. I think the others are ok putting out their thoughts on it, but should have a disclaimer that goes with it indicating that it is largely an educated guess.
Half my job is destructive testing in the welding field and trying to recreate failures. Boy is that a challenge.
GN is literally irreplaceable at this point. Thank you Steve and team for your dedication 👍🏼
Right? Without them we would literally all be on fire
Thanks, Steve
They’re really putting in the work.
@Practical Crafting congrats ? yes its been known for a while but this will spread the news like wild fire and protect EVERYONE who might have had a wind fall and built their first PC .... Just because you know about it and the majority of IT does... You are saying it shouldn't be shouted from the rooftops? Wtf is wrong with you...
Always has been.
I was very surprised that when I just got my 4090 after this video, I plugged it in fairly hard, looked flush, and still managed to pull it out easily with a little wiggle. Glad I was able to catch this and prevent a potential burnt GPU after months/years of use and the occasional cleaning that might have loosened the 12 pin. HUGE!
Hi there!! Please describe how did u plugged this 12pin cable into 4090?? Did u push it to the limit ??
@@zixerk3167 Yea, I pushed it down until it felt like it would not move any further, this was with normal plus some extra force because I saw this video ahead of time. Then I wiggled it and it came loose, so the second time I did the exact same thing but made sure to push down specifically for the latch side of the plug to make sure it hooks.
So the video is just to tell us we have to make sure the adapter is plugged in?
@@ibra5o6 no. The vid is saying its safe to not plug it in fully as long as you dont angle the cable
The fact that y'all guys took so long to release the video just to make sure that there's an actual problem it's something i really appreciate, love you guys
Yeah, that's science bitch! Tech JESUS saves us again. (This is not sarcasm, I'm very impressed)
Yep, not jumping to conclusions and having patience pays off.
"You all guys" makes no damn sense...
Meanwhile JayTwoCentz made like 3 videos of fluff, with more to go.
@@nuhashsarker5750 I unsubbed and just downvote jays videos now. Fluff and downright harmful rumor spreading without proper data collection. Also fairly pompous and rude to his viewers.
Really impressed by your dedication to your investigations and finding the true answers rather than theorizing. The pc community owes you guys a big thanks for everything you've done for us over the years and always mantaining your integrity!
Thanks Steve. Whole GN team is a pinnacle of quality IT journalism in any cases, like this one. Keep it up
Thank you! We enjoy this kind of research!
@@GamersNexus We saw it in the moment when you have captured start of the melt :D
@@GamersNexus Any interest into researching how hard it would be to have an operating system tell you when the power supply is bad or when you have a bad cable/contact situation.
Also temperature sensors next to the power plugs could be useful. (also MB)
I have always found that extremely lacking for such advanced computers as we have now.
@@czcibor6436 He was so excited lol. Dozens and dozens of hours went into that shot!
Shut up. You don't even know if they're right. Won't be surprised at all if it turns out the hardware is defective and user error has nothing to do with it.
The smell of a burning 4090 smells very similar to the burnt hole in the wallet from purchasing it.
As a real estate guru and multi millionaire stock investor, can't relate.
@@whatafailedchannel3052 don't remember anyone asking about your shady business schemes
This is my very first time donating to any RUclipsr in my 10 years watching videos, and watching the entire video, being entertained and informed at the same time is an amazing accomplishment. I am well informed about the issue and have been enlightened. You guys deserve any donations for the countless hours you guys did to bring an informative and objective light into this issue. Keep it up and I will continue watching your guys' content!
simp
Same here! Long time sub but shamefully hadn't tipped GN until this story hit. More to come!!
Yeah continue feeding the problem and paying for these shit practices and gpus, good job making it all more and more , not only unavailable in terms of prices, but giving in to these proprietary laughable cables / power supply practices.
Perfect time to do it too!
I don't know how you are doing that.
This is why I'll never get tired of supporting GN. Thank you for all the work Steve and crew!
Thank you so much, Fabian!
What CRC mean?
@@lerr 15,76 Euro
@@lerr $20 roughly
@@lerr Costa Rican Colón
This whole saga is like watching an air crash investigation episode. Good work as always and im glad to know there's someone out there actually looking into things like this
except in this case the plane crashed because the passengers didn't turn off their electronic devices.
@@andredavis3350 nah, it was probably because the passengers didn't return their seat backs and tray tables to the full upright and locked position. ;-)
Hi from the Land Down Under.. i am 62, been on the internet when it started... x386....
And Amiga, etc... my mate owns company that builds gaming towers for the Aussies...
And he gives me the new toys to stress out, so he can give a Warranty on the Tower Built...
I have a Razer 17 Pro, i7 10th gen, 8 gig RTX 3080, 32 gig ram, and the very rare 360hz screen... $6k new...
With the Thunderbolt video card box i have, i plonked in the 4090, added the extra power supply...
And within just doing benchmarking test, the 4090 was toasting itself...
So i moved it to my Benchmark Tower, and Yep... i was looking for eggs and bacon... to cook breakfast...
It's a power grabbing fire hazard...
Reminds me of the DC10 cargo door episode (where multiple design flaws led to a door appearing locked even when its not to both pilots and baggage handlers)
@Graeme Johnson undervolt and or undercooked and its fine 🤷♂️
Imagine how much better off the world would be if areas such as politics, corporations, and other shady entities had this level of credible, detailed journalistic investigation.
Unfortunately for my pipe dream you guys have found great success in the PC sector. Feels like one million subs wasn't so long ago, and pieces like this are why you're creeping on two million so quickly. Awesome work. Glad this community has you guys.
Somehow I think that politicians and big companies wouldn't like that...
All it would need is that everybody would take pride in their job. One can hope.
Naive
@@fourtii8707?
This is the first time that I was ever felt compelled to make a donation to a content creator and deservedly so. Truly amazing work from Steve and the GN team! The community is so lucky to have you guys and this video proves that no one even comes close to the dedication that you guys have to straight up factual journalism and responsible reporting.
Thank you for the kind words, Joaquin! It means a lot to the team and we'll keep it up!
*waks
Same, I went right to their store. Too bad so many of the shirts are sold out!
Damn, GN going to the level and beyond of professional engineering internal failure analysis. Huge props and the quality of your work is on par with work we do for medical device failures. Also that CT, SEM and EDS testing ain’t cheap. My guess is you spent at least around $10k to have a lab do it.
Appreciate your fight for responsible reporting. Beyond admirable!
Jay and Igor really forced his hand by being so irresponsible and loud with nothing to back it up. You could see how uncomfortable he was even saying that. I refuse to even watch JayzTwoBraincells after his outburst calling people dense and ignorant if they don’t listen to him. The man is a stain on the PC industry as a whole
Steve, you continue to evoke my respect and admiration with the lengths that you will go to in order to protect consumers. I deeply admire your dedication to using scientific methods to do all your testing. However, die a little inside very time you say "theory" instead of "hypothesis" in this video.
Whoa. It's almost impossible for a company to buy this level of analysis as a service. You're giving it away as "journalism". 😲👏
Correct journalism. Not the crap we have today. It's a breath of fresh air.
i was about to say that those analysis has to be made by nvidia, not by GN......good work GN
Nvidia should totally pay for This kind of in depth testing
@@vespermoirai975 what do you mean "not the crap we have today" ? Nothing has changed. You don't even have to believe me. Just go and read old news stories local or national. Shit hasn't changed. Your rose coloured glasses need to be removed.
@@vespermoirai975 If you think that journalism today is crap then stop reading tabloids
That footage of inserting the connector was really insightful. The amount of force it takes to push it in is way more than I expected. It looks so tight only half way down!
Said the actress to the bishop.
I wouldn't have thought it needed that much force but it's good to know.
I actually plugged my adapter into my 4090 before installing the card to make sure it was fully seated. It's a pita to get in properly and I can see why some people might not get it fully seated if they've already installed the card.
When I plugged my connector, I just hold the card from the opposite side of connector with my hand and just pressed it fully so I'm not creating stress on pcie slot. And I also avoid wiggling the cable when inserting it, so the terminals inside the connector not bend themselves creating bad contact with the pins.
I can understand where the problem comes from- a lot of people are going to hesitate before pushing down that hard on a connector for fear of breaking a pin off. If you're lucky, that's a long wait for an RMA; if not, that's $1600 down the drain.
it's seriously impressive how high the quality of your investigations are at this point.
we need more of this in the tech space!
We need more journalism like this EVERYWHERE. Just a great job. This is why GN is the best.
@@lexwaldez Agreed 👍
Great thanks to the Failure Analysis LAB. Excellent work and awesome gathering of data and information and reporting from GN!!!
They were awesome!
This is the best investigative journalism I have ever seen any media outlet perform in the IT industry. Usually they would just sellout (won't name names) to manufacturers or advertisers.
Any electrician half his salt could have told you, that it's simply a matter of too much power going through too thin wires
@@Nightdare yes but this was about determining the point of error, and all the ways possible and not using speculation.
@@Hydranox
Sure, that still doesn't require a degree in science
The connector simply doesn't have enough margin of error to handle the 450 watts put through it
That's why it fries when there is an inconsistent connection
If I attach 4 power cords to my TV, if the power supply can handle that, it better be an overengineered connection or something's gonna start arcing
Oh no... RUclips now shows who paid for their comments...
@@Nightdare yes but now we know it’s not Nvidia’s fault, it’s user error
Steve, I'm so damn glad my Patreon support money is going in to content like this. Sending this stuff to get X ray/scanned probably cost a pretty penny.
The viewer who assisted with the failure analysis lab was so incredibly knowledgeable, professional, and skilled. We were absolutely blown away by the level of depth. I'm not sure that we're at liberty to really get into detail of how it all worked (his request), but we can absolutely get into detail about how awesome it was to work with him!
@@GamersNexus I concur, thank you! I'll definitely be getting some coasters!
@@GamersNexus setting up the cable with a set of resistance meters and plugging it to a separate socket could let you plot the resistance during insertion and test a set of cables to establish the trend.
@@eric4946 Dude nobody is going to do that. They saved 1/4 of a gram of copper to make the adapter like that. If you want them to work properly, get you modular 12v cable out of your PSU and look at the connectors there. That is how you want it designed. Maximum contact will least slackness possible.
@@daznis I’m not saying that’s a solution for the cable … I’m saying that’s how to test them…
This is probably the best video on the channel. Gives me some Air Crash Investigation vibes here, where in many episodes the media, public and random "experts" on the Internet are coming up with more and more "versions" of what happened, in the meantime the investigators are honestly doing their job to determine exactly what went wrong and how to prevent it from happening again.
Thank you for existing and always accepting nothing but the truth, for doing all the hard work of gathering the data and presenting these facts. We love you guys!
Wow big donation
peanuts compared to the sponsor income, ads, merch etc.
why bother?
@@m4c1990 because he appreciates the work they're doing and wanted to support this way?
@@m4c1990 Mate, please tell me you're not complaining about a donation amount? Christ i hope you're not cause if you are....... you may want to stick your head in a fridge, eat some ice, slam the door a couple times on your head (making sure all your food is correctly sealed and veggies still fresh) and chill the F... out.
@@m4c1990 because they deserve it and you should donate/pay for things/people that you enjoy.
This is extremely well done. Thoroughness of the reporting as well as the detailed analysis is top flight journalism.
Bro save your money to buy gpus or stocks lol
@@trollerprohdx3437 no he did the right thing, you can try too lol
Just an FYI for you monetized commenters out there, RUclips is taking 30% of the amount you're contributing.
On top of that, if you're sending the monetized comment via the Apple or Android app, they are taking another 30% cut.
The creators are not allowed to mention this thanks to RUclips being a super and totally awesome company.
This is why creators would rather you buy merch, because you get something in return, and the tech monopolies get nothing.
@@trollerprohdx3437 he sent 10 bucks. Showing your poor lazy self is not great.
In my opinion, we are all extremely lucky to have such a dedicated and thorough group as Gamers Nexus. The videos you all are putting out are fantastic.
Agreed! ALL news media sources need to do, as Steve put it, "Our due diligence as a news source." instead of just fear mongering. I'm not an nVidia fan, but at least he got to the truth instead of hopping the bandwagon. Integrity 1 - click bait 0.
After getting an aftermarket 4090 I can totally understand why people would fail to fully plug in the cable. I decided to seat the cable prior to mounting the gpu, first push on thought it was in, checked saw a 1mm gap, so pushed much harder and it clicked with no gap. The force I had to use to get the cable to fully mount I wouldn't want to do it while the gpu is in the pcie slot... In fear of damaging the motherboard. First experience with this type of connector but never had to use that much force on the old 8 pin... They basically fell in and clicked in with gravity!
Does it seem strange to anyone else that even after all this time, the PC market still hasn't really figured out connections and ports? It's like no r&d has been spent on making better physical connectors at all. There are multiple industries where a solid connection that's easy to connect; while giving the end user complete confidence that it's indeed connected is a must. I mean hell, it's like as if appliances at home keep on falling out of the receptacle. A solution is possible.
This video was the final push I needed to buy a coaster set. Thank you, GN crew, for your tireless devotion to getting the story right before you publish.
Truly appreciate the work y’all put in to deliver viewers informative content like this over the years. Thank you GN!
Wow! Thank you so much! We'll keep pushing to do more like this!
Amazing content. Thank you for doing all the hardwork and research with all your cool people. You deserve a kudos for this piece. And thank you for basing your piece on data rather than going with the ‘fear mongering tactics’.
No video stands truer than the stance I've always taken on these types of things - "wait to see what GN have got to say, because it's a given they will do the due diligence necessary". Great work as always guys. Educational, informed, clear and methodical. Just how "reporting" should be!
Honestly i think gn videos could be used on schools to teach "reporting". Not just the due diligence they give the topic, all the research, but also the professionalism, responsible reporting and hunger for truth above all else evident throughout, not to mention clear and educational presentation of complex topics
Nvidia should be paying you MILLIONS! You saved them so much money in testing and engineering costs. While also probably causing millions in loss over deserved negative press ;) Hats off to everyone at Gamers Nexus. What an amazing journey to witness as you and your team have become the gold standard of tech news, hardware review and most impressively, enterprise level testing and evaluation.
I don't own a 4090 and don't plan to buy one, but I still really appreciate the dedication and hard work you all put into this. Thanks!
Pour one out for their lungs. Jensen's fumes can be quite toxic.
I don't plan on buying BECAUSE of the hard work GN puts in.
Thank you so much for that final monologue. It speaks volumes that the entire goal of this video series was consumer safety and a focus on actually getting a result rather than falling into the trap of feeling like you need to get something out. That was pretty much the best usage of constructive criticism I've ever seen of your peers without any outwardly hard feelings. Bravo.
Agreed!
The fact that y’all did this without a case and managed to still get it to melt is icing on the cake
Yup, it means lack of airflow is not a primary reason (though it will make it worse)
Yeah honestly bends are normal especially for wires. Bends dont matter they can increase risk of a failure but that risk is increased if you already made a point of failure (for example not socketing the adapter completely).
Great coverage on this issue. I very much appreciate your approach in reporting data backed theory without jumping straight to making over confident claims based on your findings. This channel just keeps gaining value as a resource and advocate for consumers in the PC hardware/tech space. Thank you!
Don't need coasters, but I think the team deserves compensation for the production and investigation, and encourage others to also contribute.
at least the money for a better script writer and editor
this vid could have been 30 seconds long
@@aa-tx7th how do you present substantial evidence to confirm your statements in 30 seconds? I think you need a better script writer and editor for your comments
@@aa-tx7th In depth reporting is what this channel is about. If you want fast, go to Tech Quickie or whatever they're calling it these days.
It could of been, but there's plenty of bad videos out there already. But i'm glad it actually contains real substance and investigation.
@@aa-tx7th Okay tiktok damaged Zoomer with 1 sec attention span, lmao
You guys are pretty much the only ones I trust any more to do solid technical reporting for these types of situations. Kudos to the GN team for being amazing human beings.
Tech jesus for a good reason,
GN…. Setting things on fire for science and safety. How did we get here? Back to you Steve!
Thanks! - This is worth the wait. Slow and steady wins more races than the rush to be click bait.
I've never seen a comment section full of donations like this video. I think that shows how much the audience appreciates these detailed, failure analysis videos thats backed by rigorous lab testing and analysis.
Yep, and quite a few said they've never donated before. Really shows how appreciated these guys are.
I too want to throw money at people who have a warehouse full of high end tech they were given for free.
@@Damaged7 ? The most expensive things in there are testing equiptment that they paid for. The warehouse rent also costs more than most of the hardware. A lot of the non CPU/GPU stuff like cases they buy with their own money. They also yaknow, have to pay staff. Again staff cost more than hardware in general [how many 4090's in $20k a year, which is probably quite a lot below what staff get]. They just did something expensive for testing, on the basis that it would get paid back in donations, merch buys and views etc. Paying the testing lab for electron microscope images wasn't free.
Is called herd behavior
I assume because RUclips takes 600-900% more cut from donations compared to Patreon/Paypal.
You guys keep blowing me away with how detailed your work is. Really appreciate you all going the ekstra mile to make sure its done right.
Thanks for your dedication. Without your hard work big corporations like Nvidia could never be challenged in the past.
10 dollars
Not all heroes wear capes! You guys go above & beyond for the community. And it can't be overstated enough how appreciated your dedication to what you do, is to us. This is the kind of ethical effort that makes all of us continue to come back to this channel. Thank you for digging into this issue.
Agreed. This was excellent work, and the sort of thing that sets GN apart from other tech news channels.
As an electrician, I am a firm believer that the pins need to be larger and the wires should be larger as well. This would create less heat to begin with and the system would be able to handle more load with less problems. This should have been addressed early on in the R&D side of things for these 4090's...
funny how so called experts forget to consider basic principles of electricity in favour of aesthetics. I don't even know how the design went past QC
@@he2a 99999999:1 it was an executive marketing decision and the engineer quit over it
Source: Seen it happen dozens of times
@@janglur yeah, let's say you save some dollars per cable by making it cheaper. If you sell multiple millions (I guess) of those cables, you save lot of money.
I am an electrician too, and i say, the connector is doing his job good. As you can see in the video, even with 2 out of the 6 12V pins, it dosn't melt. People just have to know what they are doing, and make sure, that they plugged in the calble correktly. And why should the wires be the problem now?
@@marcesser4218 would you run more current through 14-2 than rated? No, it would put unnecessary stress on the wires and either melt or start a fire. Seeing how GPU's do have safeguards against this, better gauge wires are needed.
As a hardware engineer, I have been _very_ interested in the root cause of this issue, as it is something that could have very real repercussions in what I do. Thank you for all your testing and reporting about this.
I don't even have a 4090, but you guys deserve every penny for your hard work and passion!
Thanks a lot for this! I work at an archviz company and we upgrading all our workstations to 4090's as well as most of our clients.
My team actually had sleepless nights over this, great to have it finally figured out.
Wow. That must be expensive, and extremely difficult to source given the availability.
I love your guys deep dives into investigations of problems, companies, and reviews. You dont just pull your stuff from other reviews and always do your own reviews and put a lot of your reviewing to test way beyond what anyone else does. Your testing to the point that some engineers didn't even take into account at the time of designing or didn't care to. When you do surface contact or the coolant used in a AIO causing failure of its pumps, I'm just blown away and as a big technology guy I love your break downs.
This video was OUTSTANDING!!!! It may well be the absolute best piece that GN has ever produced. It is truly a pleasure to be able to watch thorough and competent analysis of a problem, see your proof (not just feelings or suspicions, but hard facts) by showing your work, and then provide definitive solutions on how to fix or avoid the problem in the future. You shouldn't apologize for the timeline because thorough and ACCURATE research demands time and cannot be rushed. I don't own or plan to buy a 4090 now or in the future, but the lessons learned from this situation can be applied to all computer hardware (not just GPUs). Your work here may well save someone from a catastrophic failure in the future, just from a reminder that ALL cables (not just power cables) need to be fully seated and connected to avoid problems. This was damn fine work! WELL DONE!!!
Your commitment and dedication is nothing short of amazing. You are a gold standard when it comes to hardware testing and I look forward to more of in-depth research like this.
You're 50 dollar donation would probably have fed a few families tonight.
But instead you wanna simp a youtube channel and rub their ego while you hope they like you for giving them money.
@@superguy7044 why buying a phone or PC from a tech companie? You could feed alot of families for that.
Shame on you
@@superguy7044 you'RE
You made 4090 owners sleep better! Thank you Steve and GN team for all the effort invested.
I never use mine as I've been waiting for my custom 12vhpwr cable.
Ironicaly you should sleep realy fucking bad, since you're part of the people that enable Ngreedia's abhorrent behavior :'D
oooo kurvaa
@@originalityisdead.9513 Superstition not science
@@originalityisdead.9513 What if your 4090 turns out to be DOA? I'd rather test that with the included cable 1st than wait. I assure you, RMA and warranty claims are much more longer wait.
"Late" is not how I would define this coverage. Thorough is how I would define it. I give a thanks to Steve and the Gamers Nexus team for their in-depth testing, as always.
Man, when I heard use code “THIS IS FINE” it made me really realise how much I love GN and Steve. Amazing, industry shaping work on the side of consumers, and great humour to add some levity to a bad situation.
You guys are heroes
When he first said that promo code I just bust out laughing, it is so fitting, genius! I love their geeky and nerdy side coupled with strong ethics and researched reporting.
Thank you Steve. That's all I can say for content like this. Mad respect!
Mad respect Navonil to you as well.
Thank you, Navonil! That means a lot to us!
Before you even said this was the most expensive video to produce, I was wondering how much it was costing you because of all the testing. Kudos to you guys. And even though it's uncomfortable at times, you are willing to speak the truth about what you've found. There's a reason your channel is so respected.
I've seen this video a few times, just wanted to comment to help boost visibility. Love what y'all do, keep it up.
THIS is an RCA. Thank you Steve and team. I believe all of us 4090 owners pushed a little bit harder on that connection today to make sure it's seated. Hate to say it but... user error is the most common cause of most technical issues.
Yes it truly is. Also why things should be developed with user error in mind. Especially with something so easily done it could be pretty common.
The true root cause might be a design error that led to the user error. DFMEA should have caught this failure mode, especially for high current applications that are prone to use error.
@@volatilememory9338 agreed. Just want to make it clear the design specs are for sure at fault as well. Anything standardized should be tested with rigor regarding any way a user could potentially mess something up. Being that it's 2022 and this design wasn't done yesterday, could be a product of the COVID-times lack of QA availability and budgeting. Just speculation...
@@PDXnativeXPD Yup 👍.
@@PDXnativeXPD whats next a LED light on the cable to tell people its plugged in? is dat how dim you want builders to sound? it was a user error not plugged in properly more then likely through being to eager and not having a single clue on building.
Stop apologizing for something that isn't your fault. You did the testing and show a factual result with real time proof, anyone who has opinions or hypothesis with out proof is simply speculating. You don't have to apologize. If they get mad is on them not you. Also thank you for the time spent making this information available to everyone. Keep up the good work.
Haven’t ever sent a tip on RUclips, but whatever you guys earn, you deserve more. This channel is truly what the pc world needs, great work!
6:55 imho, at that point is not user error, but design error. If you designed the connector such as being easy to improperly connect and that improper connection makes it either a fire, or a shock hazard, you should not be designing said connector in the first place.
You all have outdone yourselves on this piece. The level of rigor around in-house work, croudsourcing, and working with external experts presents an insane amount of forensic professionalism. Seriously proud and amazed at how far you all have come!
literally the first thing that crossed my mind and what i came across a few times on reddit was “let’s wait for GN”
didnt disappoint, it never does. you guys are the best in depth pc hardware channel there is.
Excellent, honest videos with actual testing to back it up. We need to replace the "80 plus" certs among many others with a GN Verified certification.
this has aged really well, but at what point do we start blaming the connector spec and not the user, because if its that hard to plug in correctly, then most users are gonna struggle.
I've just hooked up my 4070 and I'm still afraid I didn't connect it fully, lol.
Journalistic integrity is why I watch you guys and not a lot of your peers when it comes to trouble shooting serious issues. Thanks for the information :)
They have no integrity. They're spread eagle for the highest bidder just like all the rest.
Journalistic integrity that is not always present on this channel. Like: Hating on the price of the 4090 and not mentioning that AMD and Intel are incappable of producing a competitive product. In Marketing methods like this one get used on porpuse to provoke reactions to increase your own engangement at the cost of "Journalistic integrity". (aka taking advantage of your audience)
If you have the best product on the market you can ask any price. It's how capitalism works. Even twice the price would be fine.
A similar example for a situation like this is Apple: They can ask for any price, because it's an Apple product. A similar PC is A LOT cheaper, but it's not an Apple product.
So they often times ask 1.5x or 2x the price of a similar speced PC and people buy it anyway. Because it's an Apple product.
People would buy the 4090 even for $3000 USD. What matters in this product segment is: it's the fastest. Not the price.
But GN regulary fails to mention facts like this one, to make people artifically upset.
@@TheNerd That's a lot of mental gymnastics there bud. Thanks for your opinion.
I'm just blown away by this, the level of investigation and the integrity of you guys is truly exceptional!
If only all investigative journalists out there in whatever industry are as thorough, neutral, and honest like the GN team. Great work, guys!!!
Not neutral, objective. Neutrality is both sides BS, just an FYI.
lost all respect for jay. he just wanted to boost his own numbers
How do you declare him "neutral" and "honest" ? based on what? he is blaming people for this disaster like if people now is trying to melt their computers, it's the most absurd conclusion ever. This and calling things "sensationalism" is the equal from MSM to other people "misinformation". Now nvidia is just "not to blame", it's just people, that only want to destroy RTX4000 series cards, they must be all AMD fanboys huh? you can put all the numbers and tools, and put some guys with white coats from the MIT but that doesn't make anything more true than actual REASONABLE conclusions.
Bless these guys, its a big relieve to know that the point of error is actually really rare and most possibly just users mishandling their cables.
Fr I’m wanting to upgrade but I was nervous now I’m not so nervous
Thank you so much for this piece!! The best Investigative work on this matter and Nvidia should pay GN for this results. I am from Brazil and 4090's here are really expensive, it's great to have some facts and conclusions after all the amazing work you all have done. Well done!
Brazilian Team found same conclusion, 2 weeks ago.
ruclips.net/video/hkN81jRaupA/видео.html
Best wishes
Hi
Been watching y'all for years, and this is some of the greatest, most comprehensive reporting I've ever seen in the RUclips tech sphere. Keep up the good work GN Team! Would love to see more videos with the failure analysis lab in the future.
Thank you so much for the years of support! It means a lot to us that people care like that. There are channels I watch the same way, so I know what it's like on both sides and sincerely appreciate it!
The GN team is irreplaceable. Amazing work as always guys!
Couldn't possibly be more replaceable, actually.
Im happy with my 3070ti. Won't even consider upgrading anything on my build for many years. Just built it this year and works like a dream.
That's how I feel about my 3080 and how I imagine anyone with a 30 series GPU feels I would imagine. There's just not enough value with the 40 series to justify the upgrade.
Im still going strong with 2080ti
🤣🤣🤣🤣 whatever helps y'all sleep at night
Nothing but the utmost respect for the team at GN. You guys are in a league of your own with this kind of content.
Wow you respect victim blamers?
@@atespeach5672 That's definitely how I mean it! Especially if you're one of the victims! :)
Incredible work GN!, The level of detail, honesty, quality investigation, investment and re-investment of resources and a top team is what sets GN apart!. Every video GN does sets the standard in quality unbiased reporting and a very high bar for others.
Thank you GN for continuing to deliver incredible content!
Tech Jesus saving us all once again. I continue to be impressed by how thorough GN is.
😂😂😂😂😂😂Tech Jesus Lmaooooooo! Never realized it till now. Luv it!
Dude you're great at a lot of things, but I think Northridge Fix is right about this one.
+1
+1
I loved how deep you went into analysis. Especially enjoyed the x-ray, EM and spectroscopy results, because I currently work at a company that makes electron microscopes and spectroscopy equipment where I put EMs together and set them up + test for customers. I am proud of working there because it helps do stuff like this among other cool things.
Keep up the amazing quality of content and integrity of your reporting. You and your team are the heroes we did not deserve but needed.
Thank you for your hard work. You set the standard for what ALL types of news should follow.
You guys deserve this from every subscriber. Thanks for taking the first principles, scientific approach and staying with this story!
Thank you!
Gents, excellent work. However, there is little mention of a critical factor which everyone struggles silently with: The overzealous application of 'strain relief' material and far too many strain points for this nexus design. Look no further than 9:05 in the video which would make Medusa jealous! All that black may looks pretty, but it stiffens the cabling. Couple the problem with not one, or two but four strain points, each with bulky connectors, + extra useless/unused connectors and a proven unreliable latch mechanism. Your typical ATX24 cable has more conductors than this entire Medusa mess (5 x 4 cables), and generally is engineered to ride on a single, less bulkier cable bundle. Generally, the strain relief on ATX24 is not run end to end either, allowing for greater bend a flexibility and lowering the mechanical stress on the connector. Quite frankly, each of those PCI-E cables should home run to the GPU without an adapter.
Bought a shirt to support. The testing done by Steve and the team at GN is so meticulous. Coming from an Electrical Engineering background and being a PC "enthusiast", I love the extent you guys take things to duplicate and verify issues. Your work and effort deserve more likes and subs as it's rare to find such unbiased opinion and thorough testing elsewhere. Also, please restock the hoodies.
This is why I value GN's content above all others, be it reviews or otherwise. Absolutely stellar work, guys!
To think these guys were once working out of a house doing mostly CPU and GPU benchmarks. It's impressive to see how far you guys have come since then! Thank you for going out of your way to provide this much valuable information for the PC-building community.
Hey Steve, Patrick, Andrew and everyone else involved in this testing,
When the first pictures and videos came up, I showed them to a friend of mine. He works for a big connector-supplier in the automobile sector and designs for example ECU-Connectors. He instantly told me, that the problem is in the connection between the male and female pins, not the connection from pin to cable. His words where along the lines of "The melting starts from the tip and not from the cable". He was still quiet displeased with the connectors design from a manufacturing standpoint, but his first impression of the failure is in line with your findings.
We decided against publishing our opinions, since we did not have the resources to look deeper into it and without backing, our opinions would just have added fuel to the dumpster fire.
So thanks for your work!
Greetings from Germany and have a great day!
Simmie
As simple as the issue was, I appreciate that you guys ran multiple other tests to rule out everything else.
Your team's exhaustive and scientific approach continues to raise the standards of technology reporting, thank you for your continued hard work
Scientificism at its best. Being scientific doesn't equal truth nor logic, also what is scientific about this? Just using a computer and troubleshooting is already science. The conclusion that people is to blame because they are brainwashed into destroying their own computers or they just want to for no reason is completely insulting and irrational, no matter how much science theater you put on stage.
Rare to see such in depth and professional investigation from a third party! Appreciate what this does to help inform consumers.
imagine paying
@@vineecuador9629 imagine being cheap. Nothing wrong with supporting a good channel that you mostly watch for free
@@Intoxicated4768 imagine being mad
@@vineecuador9629🤡
@@vineecuador9629 🤡
This was really excellent reporting. As an electron microscopist, I particularly liked the failure analysis, really comprehensive.
I don't own a 4090 but I'm still checking that my GPU power cable is properly seated when I get home 😂
i would say, no excuse, running that much power thru that piece of garbage... On sensitive electronics... No way, buying unsave hardware, where nowone has the solution ) its a bad gambling that NVIDIA did in my opinion...
@@_andi_. Just plug it in right bud, you'll be fine.
@@MDxGano I have a 4080 with a daisy chain works fine so far on rm850x
@@realraf1 Sorry man, there is still a difference between Frankenstein wiring, and incomplete connections. You could use fucking molex and still be fine as long as you had everything plugged in all the way.
@@_andi_. Na, cable works fine. User error
Thanks heaps for this, I have been having massive anxiety about using my 4090 since upgrading $3.5k in Australia and I did not want to lose it from a burning connector, Avoiding playing games, limiting my graphics and watching the performance monitor to keep the watts low, This video gives me some relief in the fact that I have it plugged in all the way, clipped and not bent at all so fingers crossed I don't see any issue 🤞🤞
this deserves a video of the year award or something. the explanation, clarity, & lengths you've went thru..... fantastic.
Excellent job as always and you're the reason why I ended up with a 5950x and 6900xt based set up last year.
The cost of a single 3090ti in Australia is enough to pay for my entire build and your thorough reviews steered me in the right direction.
Saved a bunch of money and have a very satisfying set up for the next few years.
Cheers from Down Under!!
How does this even happen? They MUST have known about the issues with this before they launched the connector.
Some folks at Nvidia will get a good night sleep after that.
I've been very disappointed with the fearmongering regarding this connector when we had so little information. In the past this type of analysis would stay behind corporate doors, so it is fascinating to see the type of tools and work done to find the root cause of an issue. It is also funny that the simplest answer was the most likely, when you know what to look for (the melting and bending marks on the side of the connectors).
I really like the idea of this new connector, and I hope its sturdiness can be improved upon in the short term. Having 2 or 3 8-pins in a GPU is a signal we need a better connector and should not stay as the norm, and people just refusing to consider this new connector based on flimsy theories is not great for the development of power delivery.
This type of journalism deserves all the recognitions. I'm really blown away.
It's drama that gets clicks. of course there would be many that would take advantage of that.
Thank you to everyone for the content and as always, we hope to always receive information / reviews of this caliber from GN. Always a fan and always will be a supporter of you guys.
I am a journeyman electrician and I found that your testing was sound and your conclusions reasonable. I have just purchased the 4090 knowing full well about the connector issue and you have shown that attention to detail and workmanship is the key to a successful build. As a thought to consider, when voltage goes down, current and resistance go up. When overlooking you lower the voltage. If the manufacturer used 14 gauge wire with connectors of the same rating, this could have been avoided.
As a paranoid 4090 owner, thank you. Kind of crazy that a third party has given us more info than Nvidia themselves.
"Kind of crazy"
Nah. This happens plenty of times.
U brought 4090 on ur own so it's on u
@@8Harbinger8 Yeah. Bought it on launch and expected it to catch on fire. Definitely my fault.
@@IAMkingofkings318 lol
@@IAMkingofkings318 yep, how dare you buy a GPU and not expect a bonfire
You're not late, you took the time necessary to test extensively, you're the first (to my knowledge) to share this well tested information.
I am just blown away by the depth of the analysis. Grateful to have caring and passionate media outlets like you guys. NO CAP ❤
the 12VHPWR connector was junk from the start - sense pins that stuck out WAY too far allowing power when the connector was not properly inserted, TOO much power going through a small amount of contact area in a CONFINED HEAT INSULATED area being the primary issues.
The newer variation on that connector DID mostly fix the first issue but made the SECOND issue potentially WORSE by INCREASING THE ALLOWED POWER LIMIT even higher.
AMD has been smart to IGNORE this BAD BY DESIGN connector design, I wish the NVidia AICs would do so more of the time and particularly on the HIGHER power cards.
Sure 🤣