Это видео недоступно.
Сожалеем об этом.
This video card had a broken fan. Here's how I fixed it
HTML-код
- Опубликовано: 30 авг 2022
- EVGA XC 20 series cards were notorious for having one or both fans getting stuck at 100%. In this video we diagnose and fix WHY this happens!
Check out the new H7 series of cases from NZXT at nzxt.co/JayzH7
Get your JayzTwoCents Merch Here! - www.jayztwocen...
○○○○○○ Items featured in this video available at Amazon ○○○○○○
► Amazon US - bit.ly/1meybOF
► Amazon UK - amzn.to/Zx813L
► Amazon Canada - amzn.to/1tl6vc6
••• Follow me on your favorite Social Media! •••
Facebook: / jayztwocents
Twitter: / jayztwocents
Instagram: / jayztwocents
SUBSCRIBE! bit.ly/sub2Jayz...
I was a mechanic for 20yrs......it was ALWAYS an enjoyable experience to disassemble, clean, repair the internals of anything without needing to replace the entire part........theres something so satisfying about it.
SPECIALLY when it comes to repairing something electrical that isnt produced any more.
I thought I was the only one that enjoyed vehicle maintenance and repairs.
@@Jdmorris143 dooooood........when something is discontinued, not found on months of internet searches, or nonexistent in a salvage yard........and ya fix it......cant describe in detail enough for someone to understand......when ya know, ya know
Getting old things spiffy again is the best. I took some old tools and dropped em in vinegar and sanded the rust pits down and they look great and the quality is awesome.
@@robertboicourt It's like a religious experience
As an electrician/electro tech, i admire your troubleshooting process. Sometimes people doesn't realize that just by opening their faulty hardware and they could save hundreds of $.
Great video as always!
As a DIY builder, this sort of simple troubleshoot and diagnose type video is simply invaluable to me, (and I'm sure many others) who may experience issues with our components. Helps us wrap our brains around simple issues that we may not have thought of or deducted with mechanical thinking. Thanks, as always for the informational video!
This is precisely my absolute favorite kind of hardware content. Good stuff.
yaaz yaaz true
Jay, I think you have discovered a problem that many EVGA owners are experiencing. Not only with the 2000 serries, but 3000 serries also. I built a system that has an EVGA 3090, and now that user is experiencing overheating issues becuase of a faulty fan.
Is it the far left fan and its just started with that problem?
I changed that fan on a zotac rtx trinity but I also got told that there's a resistor in the fan hub that goes faulty that makes the fan go 100%.
@@adelestevens Yes, closest fan towards the outputs of the card. When it worked fan would be going 100%. But now that fan is dead, card started to smell like hot wiring. I had the customer send the card in for RMA today.
When I got my 2080 XC Ultra decades ago I believe it was the first time they intoduced the silent fan shit and after quite a while I got annoyed with hearing the fan(s) constantly kick and stop. To remedy I bought a hybrid kit and liquid cooled it. Fast forward to 3k series, same thing, traded that 2080 for a 3080 FTW3 and immediatly bought a hybrid kit for it haha. And lastly, just this year I moved into a hybrid 3090 Ti FTW3. I can never go back to an air cooled card.
I just RMA'D a EVGA 3090 with a dead fan closest to the output as well.
@@renegade_patriot I wonder how EVGA is solving this problem 🤔 Seems like cheap wiring, and bending of the wire in the cards seem to be a common factor.
That fan is nuts. It highlights how nice your audio usually is. Your noise gate that’s so smooth now accentuates that intense background noise.
I went through 3... yes THREE 2080ti FTW3 cards from EVGA and all 3 had an RPM issue with Fan 1 similar to the card you repaired here. I love EVGA and their customer service is second to none, but I think the issue you found on that card is present across many of their 2000 series SKUs.
Heat and voltage stress on cables & electronics always perplexed me. Funny thing was, I learnt about it from testing analog set top boxes with AV connections, because those old style boxes pushed not only video and sound, but also a small trickle charge which killed those things fast. At least digital cable boxes only push a charge thru because of MOCA bridging.
your use of the nonexistant word "learnt" made me ignore everything else you said
@@diatrymafire7243 "tell everyone what an ass you are without using the words."
Learnt is absolutely a word and past participle of learn. But go off?
@@diatrymafire7243 when an um-actually moment turn against you 🤓 that's why I learnt long time ago to not be the smart ass
@@Mr.Anders0n_ 1. dialect has no bearing on proper english but due to dialect it is a word just not proper but I chose to drop the conversation. 2. your comment turns you into the smart ass and negates itself.
Personally, he was not to blame for the "0db" technology.
I would look for the guilty ones in the design and/or assembly department.
Such crushing, twisting and stretching of wires, especially thin ones, is not a good idea.
Idk who came with idea like this to push in those wires and plugs like this.
EVGA seems to have some issues with their fans. My previous 2060 super EVGA card had both fans clipping the heatsink making an awful noise and they eventually failed outright. My current EVGA 3080 ti has no problems so far but it seems they need to pay better attention to their fan design and QA.
Ive had multiple generations of EVGA GPU's and none of em have had fan issues. /shrug
I had to replace my middle fan on this unit; bearing let go after 3 mo.
Yes this is a clear case of bad QC . and for something so simple like a fan cable, unacceptable - ASUS with issues with their HERO board, upsidedown caps, etc, it's clear that quality control has taken a big hit, for many companies - eventually as companies cut on costs and cut corners, the shit will hit the fan (sorry had to 😁 ).
@@jimmydandy9364 I bet 3 double bacon cheeseburgers that there was a serious scab issue after the pandemic. untrained workers.
@@krushandbashgaming Fortunately this is nothing major and an easy fix, it is not exploding or catching fire 😁 But none the less, companies skip out on the quality control, I mean how the fuck can companies not detect a bloody cap installed backward FFS, lack of QC, they assume most of the market buying this are complete ignorants - today it's PWM wires coming loose, tomorrow it could be fan blades coming off and and flying out the PC case and slicing everything on site, or maybe fans literally falling off :P It's gonna get far worse over the next coming years - You should see some of the cheap arsed power boards and circuit boards they put even on expensive TVs.......................Good thing Jay fixed this issue hopefully it will be quite helpful to others, that's why I am such a "fan" of this channel 😁
Don't know if anyone else noticed, you can actually hear the fan every time Jay speaks or types. You may need good headphones to really pick up on it because it's subtle, but once you do hear it you can't unhear it XD Phil did a great job noise gating it for the most part. I only realized it because when Jay cut to working on the card, I couldn't hear it anymore
Also noticed this too, thought it was maybe because of aircon but makes sense with it being the gpu, was quite distracting hearing it cut in and out
That happened to me with 3080TI. Had to wait three updates courtesy of the Precision X software before an update fixed it. It was when they added support for the Kingpin card. Both fans ran 100 percent but games ran without gpu supporting them. Had to use a rollback driver for a while.
Hey Jay, hope you see this. Because of Fast Boot and other shenanigans a "Restart" is actually a more complete shutdown than the "Shutdown" option. So unless you switch the PSU off and back on it's smarter to do that.
Yep - restart does a full windows reboot. To do a classic "Full shutdown and cold boot" you need to choose restart in windows, then hit the power button during POST to power off fully, then power off the power supply for a minute to give capacitors a chance to drain.
Or, disable Fast Boot via Windows-R then typing powercfg.cpl and turning off Fast Boot there.
@@raymarquardt5157 Yup! I've disabled Fast Startup and the system still boots plenty fast enough. No perceivable difference. Using an NVMe on PCIe gen 3. It probably only makes sense if someone is still booting off a hard disk... we see all the confusion that this feature has caused, and almost every tech youtuber making a video about it. "YoUrE tUrNiNg OfF yOuR Pc WrOnG!" lol
For those who need a little more guidance, after running powercfg.cpl (in Run, or CMD, powershell, whatever), click "choose what the power button does" (on the left). Checkbox is here, but you may have to click "change settings that are unavailable" (requires administrator) to make changes. Windows 10. For further guidance, there is the googles.
Just hold the Shift key while clicking shutdown.
Turning off the PSU is going to do nothing to Fast Boot, the whole tech working behind it is hibernating the kernel while shutting down userland. Since its a hibernation, not a standby, the system losing complete power is completely unrelated
I turn PC off and flip the PSU off (easy to get to). Get some coffee, snack hit the head and anything else I want to do while I'm up. Go back flip it on and it boots perfectly every time.
Love how your test bench is called "THESTBENTH". Too funny Jay keep the content coming.
Whenever Jay talks, you can hear what sounds like a fan. Now you've heard it, you can't unhear it.
You're welcome.
I had this happen to my MSI1080ti Gaming X Twin Frozr a few years ago, exactly the same issue with one fan always running at 100%.
I contacted MSI because it was still under warranty but they wouldn't RMA it because I bought it used, and MSI don't take products back from the public in Europe, you have to send it back to the shop you bought it from, and I had it shipped from the UK to Ireland and so couldn't take it back to the shop.
MSI, were terribly rude about it too; I contacted them several times, and each time as soon as I mentioned that I didn't buy it new they just deleted my ticket! 🤬
So I just wired a motherboard connector on to the fan and ran it from the motherboard instead with Speedfan. And shortly afterwards I got a waterblock for it.
But damn, I wish this video had been released a few years ago, it was probably an easy fix like this!
@JayzTwoCents, Windows 10 shutdown is actually sleep/hibernation. They changed how Windows starts up to allow for "Fast Boot". Restart actually shuts down the system. You have to disable hibernation in order for shutdown to truly shutdown. I have been meaning to comment the last time you said that but this video reminded me.
I deep cleaned my GPU recently. Evga GTX 1070 SC.
During reassembly I didnt fully connect the fan wires. And under standard running vibrations it came loose, but not fully loose. So for that (unknown at the time) reason, my fans were spinning up to the best part of 5000rpm, then it reconnected at returned to my fan curve. After a bit of trial and error with fixes, MSI Afterburner and such, I checked the loose connecter. Very easy fix if I knew the issue from the start. Gpu out, use a clip puller, like what you use to remove fender linings on cars, to push the connector fully in. And boom. Fixed
I agree with you when it comes to not using the 0db mode. I like the fact that the card gets less dusty and the fan isn't wearing out, but I don't like the idea of SMD components on the PCB getting cooked sous-vide style.
I also find the fan spinning up and slowing down at random intervals more distracting than a fan quietly spinning at a barely audible RPM.
That happened to me too, with my 3070 FTW. I changed it for a ROG for the same price. The performance is slightly better, but the FTW looked much better.
Dave's garage did a really interesting video about how modern power states have made a restart cleaner than a shutdown and reboot. I still do shut down and reboot, because that was always the best way to get a full fresh state, but apparently there is so much quick start stuff built in to Windows and system bios now that a restart is actually better.
Came here to say exactly this.
6:52 A restart is actually the better way to do this. When fast boot is enabled (which it mostly is by default), system memoy just get dumped on your os drive so it can boot faster the next time you power it on, while a reboot actually clears out everything and starts fresh.
You can force windows to do the same thing on shutdown if you hold shift while clicking on shutdown in the start menu (just like you can bring up the recovery menu by holding shift while clicking reboot).
I don't think that was caused by thermal cycling. I'm not an expert, but I have seen hundreds, if not thousands of failed crimps, and I'm pretty sure this is one. It's likely just a case that a spec of dust or something got in between the jaws of the crimping tool and caused it to not quite apply the correct amount of force.
How's that a bad crimp? It doesn't seem it slid out from where the sheath is crimped, but just broke at the pin. It looks like it was probably damaged during installation due to strain and then it was only a matter of time till it let go entirely.
bad whirring. glad you found out what was going on. You think for the cost they would have better cables.
I had this exact same issue on an Evga 2070. Issue was solved by just pushing the fan connector in again properly as it seemed that is somehow slowly worked itself loose which caused the issue. The connector was still plugged in just not enough apparently.
This reminds me of a time I was having an issue diagnosing a ridiculously complicated 3-way active stereo system I installed in my 2012 Mustang.
It turns out one of my speaker wires had a weak connection INSIDE the wire insulation. On the outside it looked almost perfectly fine except a small bulge where the issue was. Once I stripped the wire back I saw that I had managed to get the wire spool that had the main wire splice during manufacturing. You could see where some machine had botched twisting 2 lengths of copper wire together before sending it through the machine to add the insulation.
Supervideo. Just 1 thing. Voltage does not create heat in a cable/wire. Amps combined with resistance does. So if there is a bad connection (meaning: high resistance) that point will heat up because of the amps flowing through it. There will be a voltage increase over that point and a voltage decrease at the end of the cable/wire. You maybe could be able to see the hotspot at the breakpoint of the wire with a thermal cam. I can imaging it will be hard on an graphics card.
"IT WAS AGAINST TEH TWIST!" all subs watching did not even flinch and understood PERFECTLY!
The cable is fixable if you can solder without melting the connecter possible if you know what your doing along with some RTV to close off the bare end of the solder joint. But good that you found the problem with the video card.
Good stuff, Besides the Humour and occasional Saltyness, this is what makes us enjoy Mr. J.
I had this issue with my 2070S Black. Difference was that an unplug and replug of the main fan connector is what fixed it for me.
Similar problem with a zotac rtx 3090 trinity.
The card was OK when I bought it and several months later one fan went to 100%.
Just changed the one fan(again the far left one) and the problem was solved.
Chatting to another user and it turned out this was a common fault caused by a resistor inside the actual fan hub.
But by this time the broken fan was in the trash and I couldn't test this out.
Best thing about tearing the card down was being able to put much better thermal pads and paste in and dropping the temps by 10°!
I have an EVGA FTW3 2080 Super and have had the first fan starting at 100% for like the past year or so, Def gonna look into this.
So, if EVGA has a bunch of replacement parts, do they have an end-user accessible part number for the Y-harness? That would be handy to have.
Always check the most common fail point first, the wires.
Loved this video! These are the kinds of videos that made me love your channel. You've helped me so much with my pc over the last 2 years. Thanks Jay!
I have had the same experience with car ECUs the constant heat cycles dry out the solder joint and the cable just pulls out of the connector.
I'm saving every Jay's troubleshooting video. Very valuable stuff here.
God find Jay! I think vibrations from the fan spinning might have contributed more than heat cycles. Also, starting from zero rpm is going to generate larger vibrations until the fan is up to speed.
The true cause of the problem though. Bad Design or assembly. It never should have been designed to secure the harness at the highest stress part of the cable. It secure it on the connector (preferred) or secure it away from the connector (recommended). Securing it on the wires next to the connector wouldn't pass inspection where I've worked in the past.
See International Electronics assembly standard IPC/WHMA-A-620 Wire Harness Reliability.
This was actually very helpful as I had a card doing the exact same thing...Thanks Jay🙂👍
Just had a similar issue on a Strix 1070 i got for £90, At around 64% fan speed or any custom movement with AB and a clicking noise then eventual failure happened, i could plug it in to another PC and it would be fine, so after some contact cleaner and connector cycling making sure for no pinches..all seems well so far, but i think the fan making a clicking noise is showing it may be EOL as pushing the wires like in the video could cause changes.
OMG, Jay finally fixed something first try-ish.
Good video. Jay, want to do another video where you fix or replace the connector? PLEASE! PLEASE! PLEASE!
The bearing seized on fan 1 on my EVGA 3080ti, tapping the centre hard got it spinning again, but of course it was noisy as hell. Took it back to the store for a full refund and I just had to wait a couple of weeks for a replacement.
Well, I was kind of hoping for a simpler fix, but at least it's still under warranty. EVGA's customer support has been super awesome in the past.
Interestingly, some video cards have a half-zero-dB mode. Not sure if that's still a thing or if it's just in ancient tech like my Sapphire Radeon R9 270X. When it's under light load, its fan closest to the display ports (fan 1, I'll call it) goes at some low speed, while its other fan (fan 2) is off. I spent a while worrying that my fan was broken, actually looked into getting replacement fans and figuring out the process, before, during that research, finding that fan 2 is *supposed* to be off when it's not under heavy load, and the way you check if it's dead is to check it during boot.
I literally just replaced my far right fan on my EVGA 3070 Ti as it was beginning to make terrible noises, sounded like the bearing were dying. I could see my temp's were going up as the fan eventually fully died as I was waiting on the replacement. All happened in the span of 3 days between hearing the faulty fan and it fully dying (not spinning) and installing the replacement.
Jay:
I wouldn't thought you might have tried using that fan speed program you showed us a few months ago....if you could speed the fan up and down with that, that may have given you more diagnostic information.
Honestly the noise gate was completely unnecessary. I know alot of dumb people comment on youtube videos complaining about fan noise in the audio. So im contributing to the other side of this issue. The noise gives context for the space you are in. Noise isnt a bad thing if it’s part of the content.
Also an expander would be more gentle than a noise gate. A -10db expander plug in would probably sound pretty good. Especially with a slow release.
Engines for vehicles are like that wire. If you have an aluminum header and an iron block the 2 will heat up at different rates. That’s why the heat cycle for a vehicle is more important than miles, that’s why semi trucks last forever because basically their engine is always on.
In recent times I've had a EVGA 1080 Ti FTW3 that had one of the fans stuck at 100%, got it from B-Stock like that and RMA'd it. I've also had a EVGA GTX 1660 Ti that the fans would "pulse" at idle, moving in and out at 0 RPM causing a clicking type sound to persist since fans make a bit of noise when going from 0 RPM to spinning. Not major issues but reading some of the other comments it seems maybe EVGA doesn't have the best QC when it comes to fans on their cards
You should zoom in with camera post production when you are pointing out small parts. The camera is in focus but that zoom would make it so much better.
Love the videos thanks for all the glorious content. I would be so dumb if you didn't make this channel.
I can't stand to use the 0dB mode on my 2070 super because the fans like to spin up, then stop, then spin up, then stop. It's much more annoying than having a quiet minimum RPM level.
Hey, Jay
Can you please make an updated video (tutorial) on how to flash your bios on Nvidia and AMD cards?
Many graphics cards now that are on second hand market have the minig bios on them and gamers would like to see a step by step guide on how to return their cards, that they have bought from a miner, to a factory bios state.
Thank you.
😂😂❤️ What are the odds,
I had the exact same problem on a car at work today, power wire to the friver door moduler was broken inside the conector 😂😁🎉
Grate video btw 😁
We like it so much when you stay focused!!!
Having the same issue on a brand new ASUS Proart 4070, fan #1 going 100% from first boot and all the time no matter the software troubleshooting. Doesn't matter if in BIOS, desktop, any GPU workload or idle, temps were fine. Fans 2 and 3 responded to MSI Afterburner control normally, but fan 1 kept going 100%. HWInfo said fan 1 is getting sent a PWM signal, that signal went up and down with the Afterburner slider but no RPM change. Clean driver update? Nada. Q vs P vBIOS? Zip.
Well within the return period still, so not taking chances opening it up and instead doing the RMA song and dance.
Zero RPM mode causes less of a temperature change than running it at say 37% idle. The fans will turn off once it goes below 50 then the GPU will passively stay at that temperature until its being used again. Looking at my GPU which does not do 0 rpm I can get 26 degrees celcius on it. The delta is about 40 degrees while if it had 0 rpm mode the delta would only be 20 degrees.
IMO the problem is with Molex. They do not advice to solder the wires because normally it is not required. But in some cases soldering the wires removed any potential problem.
Soldering does require more time and precision so obviously you want to avoid it. With these small wires it really should be done though.
The secondary problem is that the wire was simply not crimped correctly which could have been caught if there was enough quality checking. Lightly pulling the wire after crimping will ensure it is actually good.
The issue is that the temperature you're referring to is only the GPU core - the rest of the board without airflow can get as hot as it wants, as long as that GPU is under 50C those parts will get no airflow. Also the fan itself - a bearing is a part that wears in motion, but running it constantly will keep the fan's lubricant in motion and dispersed, creating the boundary layer between metal/plastic parts that optimally prevents the wear - 0db/RPM modes put the fan into a stop start cycle that allows the parts to settle into contact with each other, then upon restart grind against each other until operating conditions are reached. Depending on usage, this could result in more wear than continual operation. Such stop/start wear is why you're recommended to let a car reach operating temperature before you start slamming the throttle on a track day for example - you can seriously damage parts if you don't let the car reach operating condition before pushing it.
I have this problem with my Gigabyte Aorus 2080 Super, it's driving me mad. I think the warranty period has run out, but I can't do without it for a couple of months anyway as I need it for business so that was never really an option. Unfortunately I don't have another similar card lying about so those cables might be hard to come by! Very interesting video though, it clearly looks like it's the problem!
I've had a similar problem with my EVGA 3050 running fine for hours to then suddenly crash windows and spin both fans at 100%. I swapped in my old 980 and it was fine, no issues. Put the 3050 back in and after awhile, crash and fans at 100%. I've updated drivers, checked bios, set nVidia control panel to Max Performance, done the same in Windows yet still the crashes persist. There are times when it will be good for hours, then crash 4 times in 30 minutes. This is not a gaming rig, just my photo editing (Photoshop, Lightroom, CaptureOne) machine. Any clues would really be appreciated.
This video was really useful, I'll check the hardware just in case. Thanks Jay & community! ; )
I had similar problem. It was bad RAM. Crashed randomly after 5min or 30min. Sometimes 3x in 5 min, sometimes 1x after 1h and than ok for 2h than 3x in 10min... It was nerv eating expirience.
Fans at 100% after a crash is some kind of protection. AMD GPUs also have it.
@@AlexisSLO this makes sense if part of the driver was in an area of RAM that corrupted or got unloaded, but then Is expected a blue screen. I think we need to know what "crash" means.
It's unlike any other crash I've had. Screen freezes for a second, then completely black as if it was about to restart. Then both fans max out. I'll do a system reset and all is good for awhile, could be an hour, could be 8, but a crash is inevitable. I installed the studio drivers, no change. I have not pulled the card to look if there's any crimped wires as Jay found here, as the situation sounds very different. Afraid to tinker too much, but could crimped wires cause this scenario?
@@DougEvans try to underclock (with afterburner) the gpu or ram see if it crashes out.
My last 4 cards have been EVGA, two only on purpose coincidentally, but the 1070ti even with custom fan controls never actually activated the second fan from new. As I recall it was *supposed to kick on at a given temperature threshold, but the dust build-up that I had to clean periodically said otherwise. No amount of fiddling with wires or board trickery ever fixed it. Card still worked fine though.
My typical solution with broken GPU fans:
Ghetto-mod two 92mm fans to the gpu. Plug them into the motherboard and either control them wit fancontrol or just set them to 60-70% in the bios.
Works well, makes the card a lot more quiet.
I also have an adapter from the typical mini fan header to normal 4pin, but I only use those if I have to. Like the one here, those break quite easily.
Whenever I do firmware updates I hook up the computer (and other device, if not battery powered) to my UPS
Brings a lot of piece of mind... I have the UPS for my little server, but it isn't a bad idea to get one for a high-end desktop (cleaner power and no hard-shutdowns to prevent corruption and data loss).
Great video Jay… love the troubleshooting process you go through.
14:00 I think Jay has not thought this thorugh (like often in technical issues it seems): if the fans stop and the cards stays at e.g. 45°C -50 °C, then the temp range will only be 40-45°C (assuming 90°C max. temp), whereas keeping the fans constantly running an having the card in idle at e.g. 30°C, the temp range would 60°C.
So Jay wants to keep the fans running to have lower idle temp to reduce the temp range due to an increase to max power draw. Yeah, sounds perfectly logic as shown above, because 40-45°C increase (OK, technically correct it would be 40-45K) are more than 60K increase...
no dude, there are other components on a GPU that give out heat like MOSFETs and Capacitors that don't have sensors on them and that don't have contact with the heatsink for passive cooling, that's why he has his fans turned on low all the time, it has nothing to do with core temps.
I always hate having to troubleshoot these sorts of failures. Thanks for doing this.
My room isn't capable of handling 55C idling, so I use a linear fan curve. 0% at 0C, 100% at 100C, and a straight line between. Idle is around 35C now
My 2070 super is having a similar problem - one of the fans sounds god awfully loud after a few hours of gaming. Fairly certain it's from the bearing and not the wiring - hopefully the replacement fans I bought do the trick
I had a similar issue where a wire for the gpu fan had actually completely broken off, but I fixed it by simply de-pinning the broken wire, stripping a bit of the sheathing off the good part and re-pinning it back into place.
I feel myself lucky for not having any fan problems with video cards for as long as I remember.
Water cooled cards needs no fan 🤣
Actually, for Windows 10, shutdown is less effective than restart (unless you have fast start disabled).
awesome discovery! Good to know it was a simple fix.
I love this type of videos. Once again I learned something. Thank you!
You should pull the terminal out, and solder the wire back into the terminal and see if you can fix it. I’ve done this multiple times on many different types terminals, with 100% permanent fix
I had the same issue with a PWM wire on an old Arctic Freezer 7 Pro CPU cooler. I thought for years it was the header on the motherboard until I swapped it out because the fan noise was getting to me. That's when I noticed the wire just hanging off the side of the plug. I also love the difference in sound in the new work area. Spectacular.
Awesome troubleshooting, fair play. And it was rather therapeutic seeing you do it too. Cheers 👍
I got a GTX 980ti that has _never_ had the fans ever spin. Ever. My solution was honestly to just form a "box" of three case fans (two along the length and one on the end) to force air past the non spinning fans. I never had it fail, so I t seems to have worked, though now that I'm building another computer, and have it out, I might as well test to see if the headers are even getting power, or if the three GPU cans will even spin when powered directly.
0db mode is great for reducing the dust build up as well though.
my old gtx 460 without 0db had to be disassembled and cleaned about once a year where as my 1070ti with 0db mode looked nearly brand new when I sold it to a friend 2 years after purchasing and keep in mind too my computer never gets turned off due to my old habits from back before SSD's when i didnt want to wait 5 min for windows to boot. And that i have cats...
Isn't 0 db on by default?
restart is better then shut down since windows 8. Because of fast boot function, if you didn't disable it shut down will not clear your memory. Check in Task Manager up time.
I thought no one would say this. Restart is always better than shut down since win8 unless the user disabled fast boot. 😐
Wow, so weird to see the exact GPU I'm using just randomly featuring on the Thestbenth! :o
"I can just ask EVGA for a new cable" that's great and all, but what if you don't have that option. In my case I could solder it. But not everyone can do that either.
The way you described 3-pin and 4-pin fan control was incorrect.
4-pin PWM control fans in two ways (usually exclusively): PWM Duty Cycle signal (fixed 12v power with digital 0-100% signal for fan) OR they use DC voltage control (mostly legacy support).
3-pin can only use DC control or they're stuck at a fixed voltage (usually 12v).
Fun fact: Some PWM fans actually do not work if DC controlled.
thanks Jay, this really helped me out, reseated the connection and it working perfectly again
I would guess it was a bad crimp job, I see that very often while fixing stuff.
Worst case was something that had been running for years but started saying the backup battery was missing. Opened it up and the wire just fell out of the connector, it wasn't crimped at all, don't know how the wire stayed in there for so long, I just had to put the wire back in and crimp it properly.
I have learned more from watching you than some other RUclips videos that try and do the same thing.
I'm curious as to what test bench that is. It looks really sturdy from how he moved it around by tugging on the top. It looks like a Praxis, but it's not angled.
Question would a gpu be able to damage components ?
for example i have a gpu that my friend says wont work or show any sign of life , would it be a bad idea to test it / slot it into my working system and cheek ?
For me won't fix it like this, directly deshroud and put my favourite fans, job done
good video however for total rookies like myself- it's hard to find a really good video on instructions on how to completely remove the GPU from the start of the motherboard, step by step- shutting the computer off, etc..
Can you help? My gtx 1070ti has no display. Even its fans and lights are running.
This makes me really glad I tossed a 1st gen corsair water block on my 2080ti when I got it. Yes the block squirted me and the card once, but I torqued that fucker down hard and it's behaved just fine ever since, as has the GPU. Took a lot of work manually tuning all the fans how I wanted them to run, but that's what 2 commander pros are for (one came with my 1000D, the other i had already). 5 additional temp probes and 4 separate fan profiles later I got everything tuned pretty solid though :)
I love these type of videos Jay very informative
little tip: on windows 10 and 11, fast startup takes effect when shutdown is used, but does a clean start without that cache when restart is used.
Restart is actually better than Shutdown. Things have changed since windows 7. This is explained by former Microsoft developer Dave from Dave's Garage. Shutdown, does a similar process to hibernate or sleep, and stores active applications in a written file to open the exact states as they were. Restart, turns off everything and closes all applications and opens fresh copies of everything. After learning this, I tested it out, and yes, restart is way better than shutdown now.
Yeah it's been like this since Windows 8
Wonderful. Thanks for doing such a good job on the fix of this card.
3070 ftw3 fan stop all day sits at pretty 37c, even while watching movies the cooler is a beast for the spec card lol
This issue with the GPU reminds me of how I had a Logitech webcam fail on me in a similar fashion. The cable that went from the board to the camera sensor was crushed in the tiny case and after years of use it started going bad (it'd randomly make the image Matrix-green tinted) but for quite a while loosening or tightening the tiny screws of the plastic shell of it would make it work properly.
I presume that in both cases over time the pressure damages the wires inside and causes contact problems.
My EVGA 2070Super has this exact problem. One fan likes to sit at 100% for a very long time. My current solution is to not turn my computer off since that seems to prevent the fan from having a stroke.
These are the kind of videos that I like, because I can learn from them.
"It's amazing what happens when the wires are fine" - That has to be a life lesson. Of... some sort.