Computer not working: 😡😩😖 Make a video about it: 👉😎👉 📝• Also a note I forgot to mention for those not aware: The reason it was stuck at 100C is that is the temperature limit at which the CPU will begin to throttle itself down to however low power is required to stay below that temperature limit. If the CPU is stuck at 100C at idle, it probably means that the cooling has been stopped altogether. The last time where it shut down is if it goes too far above the limit even while throttling, the CPU will shut off to protect itself.
I know about engine's... The type that uses dinosaur squeezings.. (I have also used game engines 🐱) Anyway... With an engine the general practice is to back flush the whole system or take certain pipe's or parts off and run water in the reverse direction.. interior heaters are often clogged.. which entails connecting a hose pipe to it.. and run till the water is clear. Using the same logic .. take all the bits off and flush them through.. or take outside and gently flush with the garden hose.. a jet wash is too extreme 🤯 also an airline to clean the fins .. wear goggles and mask. If the bits stick to a magnet you could use that to grab the swarf.. and you could put a magnet at some point in the circuit to catch bits.. but obviously well away from anything that it will affect. Thank you for your videos, hope this is a little help in return. Kindest regards Bella 🐈 & Mark. (ps I have a question about something else)
Sounds just like the issue my Dad's MSI desktop had recently. It was also overheating every once in a while too. We had tried cleaning it out and re-applying thermal paste, but it only helped for a few days. We ended up replacing the entire AiO water cooler and were amazed how much better the temperature was.
This is why AIOs are risky for cpu cooling. If your going water cooling route its better off to go custom loop or for budget cooling just get standard cooler to avoid these types of issues lol.
@@Matcc Well, it was a pre-built computer that just came with the AiO cooler. My dad really likes the computer otherwise, and the replacement cooler has worked out well.
MSI had been having issues with the coolant inside the pump/radiator having some kind of growth that was clogging the cold plate fins and tubes. Greg Salazar has a couple of videos of him tearing them down and showing the problem. I had the same issue recently with a Corsair pump/radiator. I have an RMA from them to send it back, so I'm sending it back. I've since gone back to Cooler Master AIOs. I have been using them since the beginning with ZERO issues. Don't know why I started using Corsair, other than reputation, and probably was on sale on Amazon.
Asetek has a patent on a broad swath of AIO designs so no one can legally build AIOs with the pump in the block but Asetek so Asetek sells their coolers to everyone else as an "ODM" and that is how you get EK and several other brands all having these clogging issues lately.
No it doesn't, thankfully to the best of my knowledge their patent only covers North America that broadly. So there are other AIO cooler designs in other markets, particularly Asia.
@@-eMpTy- I never said it did but due to Asetek's patent on that pump location all AIOs are built by them and just sold under other brands with different housing designs on the outside of the pump.
Greg Salazar would be proud, he documented a lot of these. MSI claimed it was a supplier qc process that caused the clogging and it has been addressed. Many brands went through that factory so a lot of CLCs from 2021/2022 had issues.
Greg would be partially proud. He wouldn't be proud about Thio sayin "I am not gonna bother with warranty" . He HAS to, it's not a matter if he wants to. It's a tech channel and saying that on camera, isn't exactly the best example of how you should manage those things.
@@statendrei5 The root cause found by Igor's lab is allegedly that they used excess flux in the aluminum radiator construction, specifically units that needed a manual second pass during QA checks.
@@t3amb4sh Thio Joe didn't recommend avoiding an RMA, simply stated he probably won't go down that route. I'm sure he has more important things to spend his time on.
Not saying getting a liquid cooler is bad, but this is the exact reason why i always go with tower coolers. There's just nothing like strapping a big block of metal to your CPU. If you see the fan spinning, its definitely working. Especially when I'm building for someone else I just don't want to have to worry about the liquid, or pump, or any leaks.
This video actually helped me realize why my laptop was overheating. While my laptop isn't water cooled, it did actually help me realize that I should probably check the inside of the fan and turns out it was breathing through a carpet.
The video is a GREAT example of how sharing "learning experiences" can be as beneficial as teaching the "right way". Using the infrared thermometer during the earlier parts of the learning experience might have produced a shorter timeline for the final "solutiion" but being willing to share the mistakes keeps me coming back to this channel!! GOOD STUFF!!
They would probably RMA replace it for you for free, as the AIO coolers really should last a lot longer before the coolant needs to be flushed and replaced. I know Gamer's Nexus has done some pieces on how AIO liquid coolers have had cases where the plastic parts inside will degrade/desolve overtime and then start clogging things up. Maybe that's what is happening?
My first (and only) AIO died in a similar manner. It clogged up and the pump died. I run custom water cooling now, and keep an eye on my reservoir for gunk like you had. This setup has been going strong for years now, and I'm only still using it because I had the parts. If it starts dying, I'll replace it with an air cooler.
I have a tower cooler on my i5-13600k. Even on full load it barely goes above 60 degrees. Just a Noctua 15 with only one of the fans attached (the other doesnt fit on with my ram). Works awesome, so much easier than water cooling
I got the noctua nh-d15 and Ryzen 9 5950x and it never goes above 65 even when benchmarking. Idles in windows at 30. I'm never go back to water cooling. Too many things that can go wrong.
Yeah because that's just a 13600K, not comparable to a 13900KS. Also the NH-D15 basically costs as much as a water cooler 😂. There are much cheaper options.
Sure but for longevity you can't beat aircoolers. My old nhd-15 lasted me two pc builds then I sold it years later to a man that used it in his build lol. The thing just wouldn't die and this was still with the original fans that came with it 😂. Had the thing for like 7 years@@b4ttlemast0r
@@federicocatelli8785he is using a clc, but there's nothing stopping you from draining and replacing the fluid. Some models even have a fill port in the radiator. Mine has a small screw covered with a sticker.
That's why i don't recommend water cooling when i get asked for a build. Even a techie may forget this can be an issue, i don't want to diagnose it over a phone with somebody that gets anxious just removing the side panel of the case.
We've all had problems like this and they're usually easy to solve. Some are the more difficult ones are where say the 5v line drops slightly which can randomly crash your machine. PSU problems where the voltages are slightly out can be a nightmare to diagnose without having a spare PSU to test against. There are some more interesting questions that I hoped would have been answered in the video. 1) Did you refill the AIO once you'd opened it? 2) You cleaned some debris off the fins, but do you have any idea how much debris is left in the coolant? Opening the AIO probably let in air into the system, so how are dealing with air blocking the water flow? All are FAR more relevant questions than you finding a relatively easy problem.
@@LTNetjakYeah I know AIOs have some air, but you've the potential to introduce a lot more. This is why I personally wouldn't disassemble one. There's probably a sweet spot of air to water for it to work well, but I don't know enough about the dynamics of sealed systems like these.
@@rafflesnh In all honesty, I don't think it would have taken me so long to have figured out this problem. And yes, I would have probably trashed it for quite a few reasons. Firstly as I said, you don't know how much debris is still in the liquid. You could be re-opening it in a few weeks to clean it again. Secondly, you don't know where the debris came from. It could be from the inside of the radiator corroding; it could be the plastic or even the original water being dirty. You just don't know. Thirdly, AIOs are not user serviceable. Opening it, refilling it, all have an impact. You could be introducing more air, you don't know what the materials that everything is made of, so what you refill it with may attack or degrade the components or seals. Fourthly, changing the fluid may not fix the underlying problem and may be just buy you a small amount of time before you're back taking it apart again. So yes, I'd have trashed it.
This is why I'm gonna stick to old good air coolers as long as possible. They are simpler constructively and therefore more reliable. Good work figuring this one out though!
Yeah. I briefly used an AIO liquid cooler several years ago, and had the pump die. I went right back to using a trusty Noctua air cooler instead of replacing it with another AIO. Just so many more points of failure in a water cooling setup.
Had NH-D15 for six years on 5820k. Then upgraded to 5900x with Arctic liquid Freezer II 420. I would not go back to Noctua even if they paid me. While it might be able to get to the temps Arctic achieves, it would be at the cost of significant noise. Arctic gets me 60°C while being practically silent. And it has 6 years of warranty, proof that they stand by their products.
@@TheTeflon490 I have a 5600x. With the stock stealth cooler, it used to idle at around 65 degrees and would hit 96 degrees under load. This was considered "typical and by design" by AMD, so I left it. Interestingly they say the 5600x can get hotter than more powerful chips without throttling: 90C (5800X/5900X/5950X) and 95C (5600X). Sometimes my mouse would stop working when the PC was under heavy load (and thus very hot). I'd turn it around to unplug the mouse, and the case, USB port and mouse connector would be insanely hot. Sometimes If I'd restart after heavy load, I'd get a CPU overheating message before posting. But I thought nah don't worry about this one, she runs hot. Finally had enough of my room being a furnace and all these issues, so I got the Thermalright Phantom Spirit (newer version of Peerless Assassin). I have its fans running at around 500 RPM, so it's more silent than my case fans (650 RPM). Overall it's very quiet. Idles at like 40 degrees, hits 60 under load. By far the loudest and most annoying noise that my computer produces is the coil whine from the GPU when I scroll up and down in a web browser.
I would recommend that the cooler be deep cleaned; lower temperatures may increase the longevity of the cpu. Edit: also you should look into where the gunk came from, so that it does not come back too soon.
This is why I have my RGB set to change color based on Temperature. I had to replace my cooler when I noticed that the lights on the Pump Became more yellow more often and eventually stayed yellow. Recently it changed to Red, so I looked into what was going on and just like yours, it had gunk built up in the pump. It was bound to happen since the closed looped cooler was about 5 years old at that point.
I do something similar. The top right of my keyboard has zones for CPU and GPU temps. Set a gradient based on what my usual idle temps are and the max temp it usually gets to.
I also have a story of a weird computer problem, which took me even longer to figure out than yours and still haunts me a little. What I experienced was that whenever I booted up the computer, after it loaded into Windows, the bottom quarter of my screen would start flickering like crazy, and the flickering would become even worse when that bottom quarter was supposed to display something dark. That effect would then gradually fade away over the course of half an hour. Everything else worked fine and I benchmarked it a ton but found no performance drops so naturally I assumed it had to be the monitor or the cable. After I had also tried a few different cables but found no diference, I tried out the monitor on other input sources we had in the house. Surprisingly, it worked just fine on everything else I tested it with, think a laptop, a different PC, the TV reciever. The issue would only occur when the monitor was plugged into my PC, and then only while Windows was loaded and not in the BIOS. So, I also tried to plug some other monitors into my PC and they all worked fine too. You might see how this had me at a complete loss, so I ended up not doing anything about it for an entire year or so, but leaning ever more heavily into the theory that the problem was my graphics card after all. What was the solution? My monitor turned out to be the problem, but the reason I couldn't replicate the issue anywhere else was that it would only occur when the monitor was being operated at 144 FPS. All the other sources I had only ran at 30 or 60 FPS. So in conclusion, things like this happen to everyone. Computers are just too complex to not overlook anything.
try turning of freesync or gsync whichever one you have, for me that is usually something that causes problems and i don't actually notice any improvement from having it on so i just disable it everytime
@@psilon6999 I didn't mention it but after I had identified my monitor as the problem, I was luckily able to RMA it. The replacement has since worked fine, even with freesync. I'm pretty sure though that I did try turning it off on the broken monitor, to no avail.
Yeah already after reading half of your comment I thought it sounds like "half broken" monitor that cant keep up with refreshing fast enough, because of the "bottom half" issue. Very usual problem with monitors actually, thats one way monitors die out over the years.
My Corsair AIO's pump died after a few years so I never bothered with water-cooling after that. Large Noctua heatsinks perform just as well and are quieter, will last longer and won't cost much to replace the fans if they ever die. The pump was always noisy as well, blegh.
Awesome video, I have an AIO and I was wondering what would be some early signs of failure. This I wasn't expecting, why would there be gunk in the lines if it's all enclosed from the start. This is good to be aware of!
They add biocides and anti-corrosion agents to prevent this, but it appears some units accidentally don't get enough of those. Your only sign will be increasing temperatures. The issue here, is that if it goes on too long, the pump will start being damaged, as it depends on fluid flow for cooling and lubrication.
Corrosion, for one. Most AIOs have copper heat plates and aluminum radiators. Sure, they have corrosion inhibitors in the liquid, but inhibiting isn't equal to prevention.
@@Steamrick Good to know. More reason to either stick to air for no/no maintenance, or full custom loop and regular draining. AIO seems worst of both worlds, outside of ease of install.
Greg Salazar has found more of this failing watercoolers in the past.... Not a specific brand tho, bcuz lot of them have the same hardware unfortunatly...
Thio is highkey baller with his PC. 2 4090s??? The gunk in the loop looks a lot like some kind of biofilm and i think something was growing in there, grew enough to cause you issues, and then fried itself on the cooling plate fins. This happened to my AC unit, but it also happens in PC cooling loops. Jayztwocents has several videos on how to clean and flush water cooling loops. You may want to clean your cooling plate with a soft bristle tooth brush and get some biocide to put in the coolant so it remains clean (i used detol to clean my stuff but i am not sure if the chemicals in there are safe for PC hardware).
@@antoniohagopian213 you don't need SLI to utilize 2 GPUs though. You can give them two separate tasks, like say for virtualization, or say if you're editing videos, a lot of software can utilize multiple cards without SLI or Cross-fire
Not sure if it's biofilm, i think it's manufacturer related, since most closed loop systems will be using a mixed-metal safe chemical-based coolant that has excellent thermal properties to increase the thermal efficiency of the cooler, compared to using plain water. I've seen this stuff before, and i think it might be due to how AIBs go to major manufacturers for the construction of their products, and that manufacturer probably uses cheap coolant which will gunk up the microfins on the coldplate.
When the pump started giving out on my previous AIO I simply went back to air cooling. I'm only on a R5 5600 anyway, and I usually wear headphones, so my current dual tower cooler is probably overkill already. And I must admit, after reapplying the paste I'd probably try to see if there were any problems with the current BIOS with that CPU. Idk if unclogging would've occurred to me, especially since that is such a reputable brand.
I had the same problem with my thermaltake AIO just a year ago. When you started describing the one tube being hot, the other tube cooler and the pump still working that's exactly what happened to me. I didn't open mine up to see if it was clogged, but moving it around would fix it sometimes. What would cause it to 'act up' again was actually shutting my PC off and when I turned it back on the pump would instantly be having issues. As long as I kept my PC turned on, the pump would work most of the time but problem went away when I installed a new AIO. Now I know what to keep an eye out for! Thanks Thio.
Don't worry, I've done stupider. A couple months ago I upgraded the RAM in my server. Had to rotate the cooler since the last RAM slot conflicted with it, but whatever. When I turned on the computer, everything was fine. Perfect upgrade. But, I was playing on the Minecraft server it hosted for about 30 minutes before it crashed. I figured it could be RAM instability since it was 2 different SKUs (two different sets, but both Vengeance LPX 2x8GB, for a total of 32 GB). I decided I would investigate more if it did it again. Which it did, about 10 minutes later. So I took the old RAM out, noticing it was a bit hot, and then started it again. It immediately powered down and began to boot loop. Luckily, though, I was booting with the side panel off. This made the real issue very obvious: I forgot to plug the CPU fan back in after upgrading. The thing was hitting overtemp protections and hard crashing. Stupidity remedied in 5 seconds, and everything is working fine now.
I had a similar problem with a Cooler Master AIO. To be fair, the cooler was at least 10 years old. My PC kept shutting down. I knew it wasn't cooling the CPU so I didn't want to deal with the issue. It felt like the pump was working as I could feel it. I just purchased a new cooler and my problem was solved. In your case I would not think it was clogged by it's young age. Thanks for figuring it out. It could save people a lot of frustration.
whenever one tube is significantly warmer than the other, more so in idle, that's a telltale sign of pump blockage. I would honestly contact EK to get a new unit. they are very good AIOs and those early blockages are usually water quality issues anyway, so, an unlucky occurrence this time. All brands have this happen, without exception.
You can scrub the fins with an old toothbrush to clean them out. But the fact that you got corrosion on this unit means that you may continue to get it. Hopefully you replaced the coolant with something that has a decent anti-corrosive which might help, but I'd keep a backup cooler ready to go.
I prefer air coolers. The best ones like Noctua can cool just as well or nearly as well as liquid especially for non-overclocked CPUs. They're far more reliable and nearly foolproof. They still cool if the fan dies which is easily replaceable. And they don't leak causing a path of destruction.
I’d definitely tell EK even if you don’t plan on RMA it , find out what they use as coolant . Definitely seems like that AIO don’t like mixed metals , not the first I’ve heard about with EK . Same thing happened to that EK-MSI mono block motherboard combo . Maybe it’s the plastic/coolant they use
I have now replaced one defective MSI and one failed unknown brand liquid cooler, both for paying clients, and both with Noctua fan based coolers. Both now have significantly better cooling than when the liquid coolers were brand new. I know some people like liquid coolers, but I am not one of them.
Many years ago back when I was running an i7-960, those bad boys ran hot. Got a corsair aio and my temps were good for a while, but over a 2 year period I had to RMA it 4 times due to pump failures. In the end, the replacement one they sent me leaked and killed the mobo and cpu, which they gave me a £60 reimbursement for because it was old and not worth a lot. I'm sort of glad we don't need water anymore to get decent temps unless you're doing some really niche overclocking!
I would guess that your coolant has contaminants in it, or something in the loop eroded and clogged it up. Did you put in new coolant for this second go-around?
Taking a while to find source of issue is not dumb. It's dedication! Glad you found the cause. After watching this video I'm glad that my CPU can be cooled with just beefy tower :P
As an aircraft mechanic, this video really shows how useful filters are, also, check to make sure your pump is alright, cause metal chunks in a closed system can be an indication of a pump that has broken, usually in aircraft its in the filter, but there isn't a filter in this system it seems
I love a learning experience like this. After 50+ years in IT, I can say we all do this, sooner or later. What's this? You said it yourself: assuming we have found THE problem. These things are (nearly) always clearer in retrospect. So let's look at an ideal debugging procedure and highlight the errors (not to score points, but as a learning exercise). This applies to situations where we have already tried a guess or two, since an educated guess frequently leads to a fix. 1. Collect ALL the available evidence. Using the thermal gauge at the start would give more evidence. 2. Evaluate the evidence. For example what is the pump temperature versus the radiator temperature? Missed this step. Sometimes this leads to repeating step 1. 3. Develop a theory that explains the evidence. Made an assumption here. Assumptions are OK, but it's important to be conscious that it's an assumption. Sometimes this leads to repeating step 1. 4. Test the theory. Missed this step, again by making an assumption. Those assumptions are killers. 5. If the theory is shown to be wrong then go back to step 1 again with the new info. 6. Great, the theory was right. Fix it. Maybe with a trial fix, but then do it properly. 7. You're not finished yet. Was that the only problem? Is everything completely fixed?
I run a custom loop, all copper no mixed metals to react and gum up the loop, I don't use the ready mixed stuff, but use distilled water and a bit of biocide and a small splash of antifreeze, I check the PH and top up if needed every 3 months, I have just cleaned it out after my 1st upgrade in 5 years, and its the 1st time my loop has been cleaned out since about a year before the pandemic and my waterblock wasnt anywhere near as bad as that.. my guess is something is breaking down in the loop, maybe a seal or even the inside of the tubing or maybe cleanliness during assembly wasn't as good as it should have been... I clean mine out with a white vinegar solution flushing it out several times with distilled water, and make sure everything is absolutely spotless to prevent things growing in such a complicated loop, with watercooling you always get what you pay for and a lot of those AIO's use aluminium radiators which can react with the copper if there isn't the right amount of an active inhibitor in the coolant or if it breaks down...
The only dumb thing I heard in this whole story is that you think you were dumb for not figuring it out sooner. You are one of the best PC tech RUclipsrs, DON'T EVER SELL YOURSELF SHORT. You are f-in brilliant.
This is not that uncommon. I'm tech support for years and trust me, some problems are so basic that nobody can figure them out. And when you finally find the cause, 6 people will facepalm and say "What? I should figure this one out." My favorite example of this is when a lots of our users suddenly got a bunch of problems with Internet. Some pages were not loading, VPN was not connecting, Outlook not refreshing,... not everything for each of them, but a lot of people with weird issues where the only common thing was "Internet connection". Network guys spent two days on it, VPN team was looking, Security tried to figure out,... after some time I decided that I will check even things that cannot be changed without admin password on one of those computers and there it was, proxy settings changed. Somebody released policy update that accidentaly changed proxy settings. Nobody checked this, because proxy settings were "admin access only" at this customer.
This is why I didn't opt for AIO when I built my first gaming PC last month. I prefer the simplicity of the Air Cooler. It's easy to clean if it gets dirty. I'm using the Deepcool AK620, and it works really well.
I had a Thermaltake cooler, that I basically had to giggle the tube going to the cpu, and it would cool down, or lift it up, so it would flow better. Threw that garbage away and went with Noctua, and had no more issues.
This pretty much covers my reasoning against AIO kits and even translucent water blocks in general. With air cooling, you see the fins, you see the fan, you'll see the problem if there is any. With acrylic waterworks, plugs, buildups, and air bubbles are very visible. With everything concealed in an AIO, anything goes and you're the last to know.
In the first 30 seconds of the video as soon as he showed he have AIO I thought "that's water pump problem". Mine would sometimes just stop when set to too low RPM, so just to be sure I set up throttling temp to 78C in bios, and CoreTemp software to system shutdown at 82C. With these settings system should switch off before CPU hits 90+C.
tbh, a good high airflow case with decent coolers and fans can be almost as quiet as water cooling, while being much more reliable. i have the lian li lancool 2 case, paired with the darkrock pro4 cpu cooler. i also added another top case fan. a year later i had to open the case to add another drive and i was in shock how little to no dust there was. even when playing cpu heavy games the fans hardly ramp up, which shows much airflow there is even in idle. so much in fact that dust doesnt even has a chance to settle, its crazy.
I was quite excited to discover this video as I have been having the same BIOS Post 'CPU Overheat Error' and had to turn off the notification in BIOS to get the computer to boot into Windows. When I took my 10 year old Corsair H100i apart hoping to see a load of crud blocking the cooling vanes, I was shocked to see it was completely clean. However, the pump had been making a light rattling noise on and off whilst in use, so am now wondering if the pump is failing instead.
I had a similar problem with my recent 7950X3D rig, all sorts of crashing and then the m.2 died. Ordered a new one and sent that in to be replaced, well a month later and I'm through every part in the rig and sending the CPU in for warranty. I wanted to move my NEW m.2 drive over to my spare old rig, and when I popped the heatsink off it was wet. I had waited a few days previously before touching it so I never noticed my EK 360 basic AIO (it was pretty old, several years at least and I've used it a lot) was leaking from the bottom edge of the cold plate right onto the m.2 heatsink and slot. Amazing.
I never had that kind of problem with an AIO. But I have heard countless reports from many others who have. But recently, I had a clogging issue with a EK open loop setup where my top radiator had a chunk of trapped solder dislodge after over 3 years of use when I was doing a regular drain and flush of the system. The pump was pushing fluid just fine during the refill process. But would slow to almost nothing once the loop was almost completely full. It took me a couple of days of trial and error to finally find that clog.
My AIO from ID Cooling had the exact same problem and I was just like you, puzzled about it for a long while before noticing the pump/AIO was broken. I gave up using AIO.
Similar thing that still (20+ years later) has my missus ribbing me about it. Non computer, but similar d'oh moment. Old car, fog lights at the rear not working, I checked the fuse, feed to the switch and was at the point of sitting with my dashboard on my lap, she says "did you check the bulbs?" .... Both fail together, don't be silly...another half hour later and I have finally got to the bulbs...yeah, they were blown. This is your 'foglights' moment. Welcome to the club of the sillies!
Just some guessing from what you tell and what I can see in the video: 1 - Algae? The transparent-ish stuff that blocks the fins in the CPU-cooler look very much like algae that has been building up over time. 2 - At what temperature rating of the tubes and other materials used in the water cooling system you're using? I'm asking is because it could be cause initially (thermal paste?) that later led to another cause for the overheating. The 'gunk' could be reminisens of plastic/coating/etc. from materials that didn't handle the high temperature(s). Whatever the cause is I would've installed a filter on your water cooling system because there are very likely more 'gunk' hiding in the water cooling system. You could just use a simple gasoline filter for cars, those with clear plastic and a yellow/orange paper filter inside. They are very easy to install and with the clear plastic you can easily see if the filter is clogging up.
Every other year I disassemble my Cougar Helor 360 and clean out the deposits from the pump-head distro plate. I dumped and flushed out all the original coolant the first time and replaced it with distilled water and since then NO real deposits formed again. The key was agitating the rad for many minutes while repeatedly flushing the rad to remove all of the sediment. The original coolant was the culprit, and have used distilled water ever since...that was 5 years ago.
I had the exact same issue two days ago, with a Kelvin S36. But in my case, it happend after I installed a new PSU. I think moving my PC during the installation caused the clog, which eventually cleared by itself. But it presented itself in identical fasion: first, the super high temperatures, even after I restored stock ratio and voltages. At one point it even shut itself down due to the high temps. Then, I noticed that the block was *really* hot, even though there was no hot air coming out of the radiators. My original assumption was that there must've been some UEFI settings that changed after a Windows update, so I unplugged the pump from the MB and plugged it back in, which did the trick. But your explanation make much more sense. This Kelvin S36 has been going for nearly a decade now though, so I'm gonna start looking at a replacement rather than risking a leak if I try to clean it. Thanks for the vid, I did not expect to find a sensible explanation for what happened, especially by accident.
something that needs to be a trick of the trade when cleaning watercooling objects is cycling water through the pump while one tube is disconnected to have a open loop filteration with 1/4 rubbing alcohal and 3/4 water for 100 percent maintanance.
56 degrees Celsius isn’t scorching hot, 100 degrees Celsius is the limit. And this is why I only air coolers in my computers even both my AMD 5950X builds. LOL, using camera sensor cleaning swaps for removing the gunk in the AIO.
I had a similar thing happen the only time I tried to use liquid cooling. When I built my new-at-the-time i7-8700K gamer I tried to use liquid. The first nice cooler from Newegg was DOA. I shipped it back and just went and bought an OK one from Fry's. Less than a year later I started to get thermal throttling. I could sense through touch that the pump was moving and all of the hardware monitors looked good. Eventually (about two weeks) I found that there was gunk (corrosion or something) inside. I just returned it under warranty and bought a nice air cooler and that was the end of my problems. With an air cooler you know it is working if the fan is spinning. If you don't REALLY need extra-high-performance cooling then sometimes Old School is the Best School. I didn't have to worry about any of this with on my first PC: 4.77 MHz 8088 loaded with all 640 KB memory and a 20 MB hard drive. Ambient cooling only... Then back when active cooling first became necessary there wasn't any of this thermal throttling stuff. If you accidently turned on the PC without the heat sink you just put in a new processor.
I also had the same problem with my Aio from fractal design. When I googled it, it was a know problem and there Allready was a form online to fill out and get a V2 of that AiO. 1 week later I Allready had my V2 delivered and could keep the old broken one. It kept the fans and threw the rest away. It was the mixture of the cooling liquid. The liquid in combination with the materials inside (hoses, copper aso) lead to the buildup of this stuff inside which would clog. Probably the same as with your aio.
Watercooling is a complex system that always makes problems. Lots of interactions between materials and liquids and air and it just created problems. Still my next computer is going to be water cooled again as i really wanna try extracting the heat and pushing it outdoors.
I've been doin this building and repair thing for 22 years and I've always felt water cooling isn't worth the hassle. Thanks for reinforcing my beliefs 😆
This is why you need use DISTILLED water , also add some water-wetter / something to kill bacteria. same reason in a car radiator it clearly says use distilled water and NOT tap/rain water and mix with coolant which kills bacteria / fungus etc. Also a filter is a good idea and an indicator for flow-rate. As you noticed , tons of dirt , algae , minerals and stuff growing in the water-loop is not making the computer happy. Using just normal mixed car coolant will probably work ok. Also note a car cooling loop has way bigger passages than your cpu block, making passages this tight is asking for trouble as a grain of sand will block it up. My first water cooling system used effective 2-3mm hoses and later I upgraded to 8mm hoses , the increase in flow made it orders of magnitude better. Just drilling entry holes on my first waterblocks doubled its perfomance alone.
I've had the same problem as you, a long time ago. The computer turned off sporadically and all it did was beep indicating CPU overheating. After several tests I came to the conclusion that it was the liquid cooling system. I opened it up and it was stuck. I "flashed" it with isopropyl alcohol, i mean, i replaced all the coolant with alcohol putting the open pump/dissipator in a container filled with alcohol without the dissipator. You can't even imagine the dirt that came out. I had to replace the alcohol 3 times. After thar i replaced the alcohol with coolant for the car and it has been working for me for about 6 months.
I would absolutely try to get in touch with EK about that cooler. I'm sure they'll send you a new one and they probably want to know what the cause of the clogging on yours is
artic had that issue, they send you a repair kit within 5 days just within two emails. same day response. kit was a new plate, some liquid and a seal. important to note is they informed customer before it went all bad for most of us. so it was a replacement in advance to not build up stuff in the first place. best service ive seen in decades by an IT company asside from APC (2 days replacement batterys for the UPS just based on a serial number)
I have at times wondered how many hours I have spent over the years chasing down computer problems. I built my first PC at the end of 1983. In the era of no plug and play, dip switches on cards, limited and no sharing of interrupts, even in the first gen ISA 8 bit slots the card went into could stop some cards from working, that did persist even after the 16 bit slots were introduced a year or two later. As you may guess I continued building each new CPU release, 286, 386, 486, pentium, pentium 2 and so on. In the previous decade computers have become child's play to build. But as you can assume with this history I have spent many an hour scratching the head and clocking up unproductive time. I'm glad you found your problem, for me a replacement would be high on my budget list.
Would have gone a few steps further and spray out those fins of the cooling plate with a waterpick and flushed out the rad / refill it. Just cleaning the plate seems like the problem might be resolved now but its likely to come back at a later time. Either way, great job solving the issue.👍
I feel like you should notify EK with any serial number you have, because if there's a pattern amongst a group of coolers, they may be able to identify a larger issue. If it can identify a pattern, it could be helpful data.
I used to work for Siemens Medical, and we used a circulating pump to cool X-Ray tubes used in certain kinds of cardiac cath labs. There was a requirement to use only distilled, demineralized and deionized water in those coolers (which, in that case was mixed with ethylene glycol). The reason was that tap water is just corrosive enough to cause problems with heat exchangers, possibly the problem you had. Obviously, such a system operates at above room temperatures, and we have all seen how things deteriorate outside, with copper turning a powdery brown under the best of conditions, and then green as biology takes over. Maybe it is the cooling water? Lab grade water runs around $15 a gallon, by the way...
I sooo want you to take that AIO again and push water pressure through the rad to flush out all that gunk, then refill with proper glycol and anti corrosion/bacteria/fungus additives.
I had a similar problem, idling and working temperatures began to go higher, than one day I came home to find the system off, with a log that showed it shut down due to overheating- on idle. tried turning back on, and it went up to 90C. Checked the water loop, and it was thick green, and the pump had burned out due to being clogged with - an algae bloom! First time in 15 years of water cooled systems that that has ever happened.
*World's First Ever Angioplasty on AIO Cooler!* I did not know those things required open-heart ❤ surgery. Can't wait for your video short about the inevitable transplant procedure. Nice video Doc 👍 Also send that schmegma to a laboratory for chemical analysis please. Would be interesting and super nerdy.
I hope you saved the gunk to see what kind of solvent it takes to dissolve and wash out. The whole thing makes me thing of incompatible chemicals in the coolant or system. Who knows, the coolant might be dissolving part of the supply system.
I’ve only had one time I’ve ever had an issue with cpu temps and it was with a stock Intel cooler that had the plastic mounting mechanism go bad during a reapplication of the thermal paste. I simply replaced it with the cheapest tower cooler I could order and it has been working flawlessly since.
I would have put the cold plate into an ultrasonic cleaner with distilled water to remove the particles in the fins, then do the rubbing alcohol afterwards for any that remains. That would probably only improve the temps a couple degrees more than you saw at best, but it would be a relief to know that you actually got as much out as possible to prevent a reoccurrence.
It's all part of the learning process so don't be too hard on yourself. In '95, I have struggled to get Internet working again after trying out all kind of software/driver/resintall Window only to discovered that dialup modem inside the computer got fried somehow. Was a learning experience not to only look at the software issue but on the hardware side too. This change of my thinking helped me to solve the other hardware issue that I had too instead of wasting countless hours trying to fix the issue on the software side.
Computer not working: 😡😩😖
Make a video about it: 👉😎👉
📝• Also a note I forgot to mention for those not aware: The reason it was stuck at 100C is that is the temperature limit at which the CPU will begin to throttle itself down to however low power is required to stay below that temperature limit. If the CPU is stuck at 100C at idle, it probably means that the cooling has been stopped altogether. The last time where it shut down is if it goes too far above the limit even while throttling, the CPU will shut off to protect itself.
First reply?
Thank you for showing us and keeping us protected
3rd reply and 6th like
good for you cause you can monitize it
I know about engine's...
The type that uses dinosaur squeezings..
(I have also used game engines 🐱)
Anyway...
With an engine the general practice is to back flush the whole system or take certain pipe's or parts off and run water in the reverse direction.. interior heaters are often clogged.. which entails connecting a hose pipe to it.. and run till the water is clear.
Using the same logic .. take all the bits off and flush them through.. or take outside and gently flush with the garden hose.. a jet wash is too extreme 🤯
also an airline to clean the fins .. wear goggles and mask.
If the bits stick to a magnet you could use that to grab the swarf.. and you could put a magnet at some point in the circuit to catch bits.. but obviously well away from anything that it will affect.
Thank you for your videos, hope this is a little help in return.
Kindest regards
Bella 🐈 & Mark.
(ps I have a question about something else)
Sounds just like the issue my Dad's MSI desktop had recently. It was also overheating every once in a while too. We had tried cleaning it out and re-applying thermal paste, but it only helped for a few days. We ended up replacing the entire AiO water cooler and were amazed how much better the temperature was.
MSI has a bad history with clogging due to using Apaltek as their manufacturing OEM.
This is why AIOs are risky for cpu cooling. If your going water cooling route its better off to go custom loop or for budget cooling just get standard cooler to avoid these types of issues lol.
@@Matcc Well, it was a pre-built computer that just came with the AiO cooler. My dad really likes the computer otherwise, and the replacement cooler has worked out well.
MSI had been having issues with the coolant inside the pump/radiator having some kind of growth that was clogging the cold plate fins and tubes. Greg Salazar has a couple of videos of him tearing them down and showing the problem. I had the same issue recently with a Corsair pump/radiator. I have an RMA from them to send it back, so I'm sending it back. I've since gone back to Cooler Master AIOs. I have been using them since the beginning with ZERO issues. Don't know why I started using Corsair, other than reputation, and probably was on sale on Amazon.
@@l.i.archer5379 GamersNexus has also taken up the MSI AIO problems several times and they're a fair bit more reputable than Greg...
Surprising that EK would have this problem. You should consider to RME it to EK, so they can have a look at it for the cause.
Asetek has a patent on a broad swath of AIO designs so no one can legally build AIOs with the pump in the block but Asetek so Asetek sells their coolers to everyone else as an "ODM" and that is how you get EK and several other brands all having these clogging issues lately.
@@rockdem0n
That doesn't make much sense.
No it doesn't, thankfully to the best of my knowledge their patent only covers North America that broadly. So there are other AIO cooler designs in other markets, particularly Asia.
@@rockdem0n
I meant you comment haha. The finplate on AiOs clogging up has nothing to do with the pump location.
@@-eMpTy- I never said it did but due to Asetek's patent on that pump location all AIOs are built by them and just sold under other brands with different housing designs on the outside of the pump.
Greg Salazar would be proud, he documented a lot of these. MSI claimed it was a supplier qc process that caused the clogging and it has been addressed. Many brands went through that factory so a lot of CLCs from 2021/2022 had issues.
Glad to see this comment
Greg would be partially proud. He wouldn't be proud about Thio sayin "I am not gonna bother with warranty" . He HAS to, it's not a matter if he wants to. It's a tech channel and saying that on camera, isn't exactly the best example of how you should manage those things.
what were those things that caused the clog? glue or pump impeller shavings?
@@statendrei5 The root cause found by Igor's lab is allegedly that they used excess flux in the aluminum radiator construction, specifically units that needed a manual second pass during QA checks.
@@t3amb4sh Thio Joe didn't recommend avoiding an RMA, simply stated he probably won't go down that route. I'm sure he has more important things to spend his time on.
Not saying getting a liquid cooler is bad, but this is the exact reason why i always go with tower coolers. There's just nothing like strapping a big block of metal to your CPU. If you see the fan spinning, its definitely working. Especially when I'm building for someone else I just don't want to have to worry about the liquid, or pump, or any leaks.
Been using liquid cooling for 10+ years, never had issues. I've had regular coolers fail and literally have my processor smoke and burn a hole in it.
Same.
AIO's can be good too. My h115i pro has been running for more than 4 years now and the temps are as good as day 1.
@@scamazonprime How have you never had issues with liquid cooling? What kind of maintenance do you do?
@@SaccoBelmonte Can, but not all are. Regular heatsinks are good always if you install it properly.
This video actually helped me realize why my laptop was overheating. While my laptop isn't water cooled, it did actually help me realize that I should probably check the inside of the fan and turns out it was breathing through a carpet.
I think I might have that same problem with my laptop
When even ThioJoe don't know...
That's when it's gets real
was this caused by a copper block and aluminium radiator?
There's a known defect with AIOs, manufacturers are using insufficient biocide in the coolant that causes pathogens to grow on the cooler.
The video is a GREAT example of how sharing "learning experiences" can be as beneficial as teaching the "right way". Using the infrared thermometer during the earlier parts of the learning experience might have produced a shorter timeline for the final "solutiion" but being willing to share the mistakes keeps me coming back to this channel!! GOOD STUFF!!
They would probably RMA replace it for you for free, as the AIO coolers really should last a lot longer before the coolant needs to be flushed and replaced. I know Gamer's Nexus has done some pieces on how AIO liquid coolers have had cases where the plastic parts inside will degrade/desolve overtime and then start clogging things up. Maybe that's what is happening?
Thanks for using Celsius 👍🏻 Fahrenheit is so confusing.
Please translate for me. I don't speak Celsius
@@user-uc1ct7pq3cHe wanted Hamburger units instead of Avian Pounds.
My first (and only) AIO died in a similar manner. It clogged up and the pump died. I run custom water cooling now, and keep an eye on my reservoir for gunk like you had. This setup has been going strong for years now, and I'm only still using it because I had the parts. If it starts dying, I'll replace it with an air cooler.
Drain/Change out your water regularly (at least every 3-6 months) for pure distilled / RO / demin water, or your favorite coolant mix.
I have a tower cooler on my i5-13600k. Even on full load it barely goes above 60 degrees. Just a Noctua 15 with only one of the fans attached (the other doesnt fit on with my ram). Works awesome, so much easier than water cooling
I got the noctua nh-d15 and Ryzen 9 5950x and it never goes above 65 even when benchmarking. Idles in windows at 30. I'm never go back to water cooling. Too many things that can go wrong.
Yeah because that's just a 13600K, not comparable to a 13900KS. Also the NH-D15 basically costs as much as a water cooler 😂. There are much cheaper options.
@@b4ttlemast0r but have almost 0 noise and lifetime support!
Sure but for longevity you can't beat aircoolers. My old nhd-15 lasted me two pc builds then I sold it years later to a man that used it in his build lol. The thing just wouldn't die and this was still with the original fans that came with it 😂. Had the thing for like 7 years@@b4ttlemast0r
@@LTNetjak i think(?) 3d cache chips might be running slower thus generating less heat
Nice, did you end up putting any bio-killer in the liquid of the cooler after cleaning it?
Nah maybe i should have 🤔
Thought he was using a clc🤔
@@federicocatelli8785he is using a clc, but there's nothing stopping you from draining and replacing the fluid. Some models even have a fill port in the radiator. Mine has a small screw covered with a sticker.
@@euge963 I'm not into clc so THX for letting me know .
That's why i don't recommend water cooling when i get asked for a build. Even a techie may forget this can be an issue, i don't want to diagnose it over a phone with somebody that gets anxious just removing the side panel of the case.
We've all had problems like this and they're usually easy to solve. Some are the more difficult ones are where say the 5v line drops slightly which can randomly crash your machine. PSU problems where the voltages are slightly out can be a nightmare to diagnose without having a spare PSU to test against.
There are some more interesting questions that I hoped would have been answered in the video. 1) Did you refill the AIO once you'd opened it? 2) You cleaned some debris off the fins, but do you have any idea how much debris is left in the coolant? Opening the AIO probably let in air into the system, so how are dealing with air blocking the water flow?
All are FAR more relevant questions than you finding a relatively easy problem.
@@LTNetjakYeah I know AIOs have some air, but you've the potential to introduce a lot more. This is why I personally wouldn't disassemble one. There's probably a sweet spot of air to water for it to work well, but I don't know enough about the dynamics of sealed systems like these.
@@thewelder3538 So how would you have fixed this? Trash it?
@@rafflesnhjust refill? lol
@@rafflesnh In all honesty, I don't think it would have taken me so long to have figured out this problem. And yes, I would have probably trashed it for quite a few reasons. Firstly as I said, you don't know how much debris is still in the liquid. You could be re-opening it in a few weeks to clean it again. Secondly, you don't know where the debris came from. It could be from the inside of the radiator corroding; it could be the plastic or even the original water being dirty. You just don't know. Thirdly, AIOs are not user serviceable. Opening it, refilling it, all have an impact. You could be introducing more air, you don't know what the materials that everything is made of, so what you refill it with may attack or degrade the components or seals. Fourthly, changing the fluid may not fix the underlying problem and may be just buy you a small amount of time before you're back taking it apart again. So yes, I'd have trashed it.
This is why I'm gonna stick to old good air coolers as long as possible. They are simpler constructively and therefore more reliable. Good work figuring this one out though!
Couldn't agree more
Yeah. I briefly used an AIO liquid cooler several years ago, and had the pump die. I went right back to using a trusty Noctua air cooler instead of replacing it with another AIO. Just so many more points of failure in a water cooling setup.
Had NH-D15 for six years on 5820k. Then upgraded to 5900x with Arctic liquid Freezer II 420. I would not go back to Noctua even if they paid me. While it might be able to get to the temps Arctic achieves, it would be at the cost of significant noise. Arctic gets me 60°C while being practically silent. And it has 6 years of warranty, proof that they stand by their products.
@@TheTeflon490 I have a 5600x. With the stock stealth cooler, it used to idle at around 65 degrees and would hit 96 degrees under load. This was considered "typical and by design" by AMD, so I left it. Interestingly they say the 5600x can get hotter than more powerful chips without throttling: 90C (5800X/5900X/5950X) and 95C (5600X).
Sometimes my mouse would stop working when the PC was under heavy load (and thus very hot). I'd turn it around to unplug the mouse, and the case, USB port and mouse connector would be insanely hot. Sometimes If I'd restart after heavy load, I'd get a CPU overheating message before posting. But I thought nah don't worry about this one, she runs hot.
Finally had enough of my room being a furnace and all these issues, so I got the Thermalright Phantom Spirit (newer version of Peerless Assassin). I have its fans running at around 500 RPM, so it's more silent than my case fans (650 RPM). Overall it's very quiet. Idles at like 40 degrees, hits 60 under load. By far the loudest and most annoying noise that my computer produces is the coil whine from the GPU when I scroll up and down in a web browser.
I would recommend that the cooler be deep cleaned; lower temperatures may increase the longevity of the cpu.
Edit: also you should look into where the gunk came from, so that it does not come back too soon.
Probably in-system grease that got loose from somewhere in AIO loop. The description of the gunk seem to fit this
Thio never disappoints
This is why I have my RGB set to change color based on Temperature. I had to replace my cooler when I noticed that the lights on the Pump Became more yellow more often and eventually stayed yellow. Recently it changed to Red, so I looked into what was going on and just like yours, it had gunk built up in the pump. It was bound to happen since the closed looped cooler was about 5 years old at that point.
I do something similar. The top right of my keyboard has zones for CPU and GPU temps. Set a gradient based on what my usual idle temps are and the max temp it usually gets to.
I also have a story of a weird computer problem, which took me even longer to figure out than yours and still haunts me a little. What I experienced was that whenever I booted up the computer, after it loaded into Windows, the bottom quarter of my screen would start flickering like crazy, and the flickering would become even worse when that bottom quarter was supposed to display something dark. That effect would then gradually fade away over the course of half an hour. Everything else worked fine and I benchmarked it a ton but found no performance drops so naturally I assumed it had to be the monitor or the cable. After I had also tried a few different cables but found no diference, I tried out the monitor on other input sources we had in the house. Surprisingly, it worked just fine on everything else I tested it with, think a laptop, a different PC, the TV reciever. The issue would only occur when the monitor was plugged into my PC, and then only while Windows was loaded and not in the BIOS. So, I also tried to plug some other monitors into my PC and they all worked fine too. You might see how this had me at a complete loss, so I ended up not doing anything about it for an entire year or so, but leaning ever more heavily into the theory that the problem was my graphics card after all.
What was the solution? My monitor turned out to be the problem, but the reason I couldn't replicate the issue anywhere else was that it would only occur when the monitor was being operated at 144 FPS. All the other sources I had only ran at 30 or 60 FPS.
So in conclusion, things like this happen to everyone. Computers are just too complex to not overlook anything.
try turning of freesync or gsync whichever one you have, for me that is usually something that causes problems and i don't actually notice any improvement from having it on so i just disable it everytime
@@psilon6999 I didn't mention it but after I had identified my monitor as the problem, I was luckily able to RMA it. The replacement has since worked fine, even with freesync. I'm pretty sure though that I did try turning it off on the broken monitor, to no avail.
I had a similar problem. It turned out that i had my desktop speakers to close to the monitor. Moved the speaker problem solved.
Yeah already after reading half of your comment I thought it sounds like "half broken" monitor that cant keep up with refreshing fast enough, because of the "bottom half" issue.
Very usual problem with monitors actually, thats one way monitors die out over the years.
And that's what I dislike in liquid cooling in general... Water is an abrasive substance…
My Corsair AIO's pump died after a few years so I never bothered with water-cooling after that. Large Noctua heatsinks perform just as well and are quieter, will last longer and won't cost much to replace the fans if they ever die. The pump was always noisy as well, blegh.
You are never dumb, we are kings of tech++
c++ is my nightmare
Awesome video, I have an AIO and I was wondering what would be some early signs of failure. This I wasn't expecting, why would there be gunk in the lines if it's all enclosed from the start. This is good to be aware of!
Exactly!
They add biocides and anti-corrosion agents to prevent this, but it appears some units accidentally don't get enough of those. Your only sign will be increasing temperatures. The issue here, is that if it goes on too long, the pump will start being damaged, as it depends on fluid flow for cooling and lubrication.
Corrosion, for one. Most AIOs have copper heat plates and aluminum radiators. Sure, they have corrosion inhibitors in the liquid, but inhibiting isn't equal to prevention.
@@Steamrick Good to know. More reason to either stick to air for no/no maintenance, or full custom loop and regular draining. AIO seems worst of both worlds, outside of ease of install.
@@luzhang2982 Except that many tower coolers have heat pipes which have liquid in them that could suffer a similar fate and are not serviceable.
I hope you wrote down your experience in a forum post somewhere so that people can find it in times of need.
Greg Salazar has found more of this failing watercoolers in the past.... Not a specific brand tho, bcuz lot of them have the same hardware unfortunatly...
Thio is highkey baller with his PC. 2 4090s???
The gunk in the loop looks a lot like some kind of biofilm and i think something was growing in there, grew enough to cause you issues, and then fried itself on the cooling plate fins. This happened to my AC unit, but it also happens in PC cooling loops. Jayztwocents has several videos on how to clean and flush water cooling loops. You may want to clean your cooling plate with a soft bristle tooth brush and get some biocide to put in the coolant so it remains clean (i used detol to clean my stuff but i am not sure if the chemicals in there are safe for PC hardware).
Yes. Goo almost always has a biological origin.
2x 4090s... very impressive... I'd love to have that computer...
@@NotEventssli doesn't work on them
@@antoniohagopian213 you don't need SLI to utilize 2 GPUs though. You can give them two separate tasks, like say for virtualization, or say if you're editing videos, a lot of software can utilize multiple cards without SLI or Cross-fire
Not sure if it's biofilm, i think it's manufacturer related, since most closed loop systems will be using a mixed-metal safe chemical-based coolant that has excellent thermal properties to increase the thermal efficiency of the cooler, compared to using plain water. I've seen this stuff before, and i think it might be due to how AIBs go to major manufacturers for the construction of their products, and that manufacturer probably uses cheap coolant which will gunk up the microfins on the coldplate.
When the pump started giving out on my previous AIO I simply went back to air cooling. I'm only on a R5 5600 anyway, and I usually wear headphones, so my current dual tower cooler is probably overkill already.
And I must admit, after reapplying the paste I'd probably try to see if there were any problems with the current BIOS with that CPU. Idk if unclogging would've occurred to me, especially since that is such a reputable brand.
I had the same problem with my thermaltake AIO just a year ago. When you started describing the one tube being hot, the other tube cooler and the pump still working that's exactly what happened to me. I didn't open mine up to see if it was clogged, but moving it around would fix it sometimes. What would cause it to 'act up' again was actually shutting my PC off and when I turned it back on the pump would instantly be having issues. As long as I kept my PC turned on, the pump would work most of the time but problem went away when I installed a new AIO. Now I know what to keep an eye out for! Thanks Thio.
Don't worry, I've done stupider.
A couple months ago I upgraded the RAM in my server. Had to rotate the cooler since the last RAM slot conflicted with it, but whatever. When I turned on the computer, everything was fine. Perfect upgrade. But, I was playing on the Minecraft server it hosted for about 30 minutes before it crashed. I figured it could be RAM instability since it was 2 different SKUs (two different sets, but both Vengeance LPX 2x8GB, for a total of 32 GB). I decided I would investigate more if it did it again. Which it did, about 10 minutes later. So I took the old RAM out, noticing it was a bit hot, and then started it again. It immediately powered down and began to boot loop. Luckily, though, I was booting with the side panel off. This made the real issue very obvious: I forgot to plug the CPU fan back in after upgrading. The thing was hitting overtemp protections and hard crashing. Stupidity remedied in 5 seconds, and everything is working fine now.
So from this video and Greg’s fix or flop series is don’t get an aio they break and clogged with in months of use
I had a similar problem with a Cooler Master AIO. To be fair, the cooler was at least 10 years old. My PC kept shutting down. I knew it wasn't cooling the CPU so I didn't want to deal with the issue. It felt like the pump was working as I could feel it. I just purchased a new cooler and my problem was solved. In your case I would not think it was clogged by it's young age. Thanks for figuring it out. It could save people a lot of frustration.
whenever one tube is significantly warmer than the other, more so in idle, that's a telltale sign of pump blockage.
I would honestly contact EK to get a new unit. they are very good AIOs and those early blockages are usually water quality issues anyway, so, an unlucky occurrence this time.
All brands have this happen, without exception.
You can scrub the fins with an old toothbrush to clean them out. But the fact that you got corrosion on this unit means that you may continue to get it. Hopefully you replaced the coolant with something that has a decent anti-corrosive which might help, but I'd keep a backup cooler ready to go.
I prefer air coolers. The best ones like Noctua can cool just as well or nearly as well as liquid especially for non-overclocked CPUs. They're far more reliable and nearly foolproof. They still cool if the fan dies which is easily replaceable. And they don't leak causing a path of destruction.
I’d definitely tell EK even if you don’t plan on RMA it , find out what they use as coolant . Definitely seems like that AIO don’t like mixed metals , not the first I’ve heard about with EK . Same thing happened to that EK-MSI mono block motherboard combo . Maybe it’s the plastic/coolant they use
I have now replaced one defective MSI and one failed unknown brand liquid cooler, both for paying clients, and both with Noctua fan based coolers. Both now have significantly better cooling than when the liquid coolers were brand new. I know some people like liquid coolers, but I am not one of them.
Many years ago back when I was running an i7-960, those bad boys ran hot. Got a corsair aio and my temps were good for a while, but over a 2 year period I had to RMA it 4 times due to pump failures. In the end, the replacement one they sent me leaked and killed the mobo and cpu, which they gave me a £60 reimbursement for because it was old and not worth a lot.
I'm sort of glad we don't need water anymore to get decent temps unless you're doing some really niche overclocking!
I would guess that your coolant has contaminants in it, or something in the loop eroded and clogged it up. Did you put in new coolant for this second go-around?
Taking a while to find source of issue is not dumb. It's dedication! Glad you found the cause.
After watching this video I'm glad that my CPU can be cooled with just beefy tower :P
Ouch. Good thing it's fixed now.
Glad you figured it out. On a side note, an air cooler has fewer points of failure. Air cooling is the way to go!
You're never dumb. It's just too difficult to figure out
How did you post this before the vid was uploaded?
@@xanderplayz3446 Being a member
@@xanderplayz3446Join member subscription.
This is why i dont bother with water cooling , its cool yes. Buts its also another highly complex thing that can go wrong
As an aircraft mechanic, this video really shows how useful filters are, also, check to make sure your pump is alright, cause metal chunks in a closed system can be an indication of a pump that has broken, usually in aircraft its in the filter, but there isn't a filter in this system it seems
I love a learning experience like this.
After 50+ years in IT, I can say we all do this, sooner or later. What's this? You said it yourself: assuming we have found THE problem.
These things are (nearly) always clearer in retrospect. So let's look at an ideal debugging procedure and highlight the errors (not to score points, but as a learning exercise). This applies to situations where we have already tried a guess or two, since an educated guess frequently leads to a fix.
1. Collect ALL the available evidence. Using the thermal gauge at the start would give more evidence.
2. Evaluate the evidence. For example what is the pump temperature versus the radiator temperature? Missed this step. Sometimes this leads to repeating step 1.
3. Develop a theory that explains the evidence. Made an assumption here. Assumptions are OK, but it's important to be conscious that it's an assumption. Sometimes this leads to repeating step 1.
4. Test the theory. Missed this step, again by making an assumption. Those assumptions are killers.
5. If the theory is shown to be wrong then go back to step 1 again with the new info.
6. Great, the theory was right. Fix it. Maybe with a trial fix, but then do it properly.
7. You're not finished yet. Was that the only problem? Is everything completely fixed?
Bro casually has 2 RTX 4090s 💀
Nah one’s a 3090
@@ThioJoe Still, Pretty insane!
I run a custom loop, all copper no mixed metals to react and gum up the loop, I don't use the ready mixed stuff, but use distilled water and a bit of biocide and a small splash of antifreeze, I check the PH and top up if needed every 3 months, I have just cleaned it out after my 1st upgrade in 5 years, and its the 1st time my loop has been cleaned out since about a year before the pandemic and my waterblock wasnt anywhere near as bad as that.. my guess is something is breaking down in the loop, maybe a seal or even the inside of the tubing or maybe cleanliness during assembly wasn't as good as it should have been... I clean mine out with a white vinegar solution flushing it out several times with distilled water, and make sure everything is absolutely spotless to prevent things growing in such a complicated loop, with watercooling you always get what you pay for and a lot of those AIO's use aluminium radiators which can react with the copper if there isn't the right amount of an active inhibitor in the coolant or if it breaks down...
Nice figuring the problem, great job.
The only dumb thing I heard in this whole story is that you think you were dumb for not figuring it out sooner. You are one of the best PC tech RUclipsrs, DON'T EVER SELL YOURSELF SHORT. You are f-in brilliant.
This is not that uncommon. I'm tech support for years and trust me, some problems are so basic that nobody can figure them out. And when you finally find the cause, 6 people will facepalm and say "What? I should figure this one out." My favorite example of this is when a lots of our users suddenly got a bunch of problems with Internet. Some pages were not loading, VPN was not connecting, Outlook not refreshing,... not everything for each of them, but a lot of people with weird issues where the only common thing was "Internet connection". Network guys spent two days on it, VPN team was looking, Security tried to figure out,... after some time I decided that I will check even things that cannot be changed without admin password on one of those computers and there it was, proxy settings changed. Somebody released policy update that accidentaly changed proxy settings. Nobody checked this, because proxy settings were "admin access only" at this customer.
This is why I didn't opt for AIO when I built my first gaming PC last month. I prefer the simplicity of the Air Cooler. It's easy to clean if it gets dirty. I'm using the Deepcool AK620, and it works really well.
weird proper nouns
I had a Thermaltake cooler, that I basically had to giggle the tube going to the cpu, and it would cool down, or lift it up, so it would flow better. Threw that garbage away and went with Noctua, and had no more issues.
Good job figuring it out eventually! And thanks for the valuable info on what the issue was AND on how you figured it out!!
ThioJoe had accidentally clogged the cooler with his old CPU because he upgrades his hardware so much
This pretty much covers my reasoning against AIO kits and even translucent water blocks in general. With air cooling, you see the fins, you see the fan, you'll see the problem if there is any. With acrylic waterworks, plugs, buildups, and air bubbles are very visible. With everything concealed in an AIO, anything goes and you're the last to know.
I've seen inline flow indicators on other systems where it's critical. Something like that could be an option.
In the first 30 seconds of the video as soon as he showed he have AIO I thought "that's water pump problem".
Mine would sometimes just stop when set to too low RPM, so just to be sure I set up throttling temp to 78C in bios, and CoreTemp software to system shutdown at 82C. With these settings system should switch off before CPU hits 90+C.
This is way you're supposed to use purified water for your computer not tap water
You mean distilled water?
@@CnCDune yes
Oh thanks
Always change your coolant when you service your radiator, Joe!
anybody knows what widget he's using
It’s rainmeter but i added the temperature part to be used with hwinfo
@@ThioJoe would you plz share that skin config!! I tried but can't seem to add that Cpu Temp with hwinfo already configured correctly!!
Last
tbh, a good high airflow case with decent coolers and fans can be almost as quiet as water cooling, while being much more reliable.
i have the lian li lancool 2 case, paired with the darkrock pro4 cpu cooler. i also added another top case fan.
a year later i had to open the case to add another drive and i was in shock how little to no dust there was.
even when playing cpu heavy games the fans hardly ramp up, which shows much airflow there is even in idle. so much in fact that dust doesnt even has a chance to settle, its crazy.
I was quite excited to discover this video as I have been having the same BIOS Post 'CPU Overheat Error' and had to turn off the notification in BIOS to get the computer to boot into Windows. When I took my 10 year old Corsair H100i apart hoping to see a load of crud blocking the cooling vanes, I was shocked to see it was completely clean. However, the pump had been making a light rattling noise on and off whilst in use, so am now wondering if the pump is failing instead.
I had a similar problem with my recent 7950X3D rig, all sorts of crashing and then the m.2 died. Ordered a new one and sent that in to be replaced, well a month later and I'm through every part in the rig and sending the CPU in for warranty. I wanted to move my NEW m.2 drive over to my spare old rig, and when I popped the heatsink off it was wet. I had waited a few days previously before touching it so I never noticed my EK 360 basic AIO (it was pretty old, several years at least and I've used it a lot) was leaking from the bottom edge of the cold plate right onto the m.2 heatsink and slot. Amazing.
I never had that kind of problem with an AIO. But I have heard countless reports from many others who have. But recently, I had a clogging issue with a EK open loop setup where my top radiator had a chunk of trapped solder dislodge after over 3 years of use when I was doing a regular drain and flush of the system. The pump was pushing fluid just fine during the refill process. But would slow to almost nothing once the loop was almost completely full. It took me a couple of days of trial and error to finally find that clog.
My AIO from ID Cooling had the exact same problem and I was just like you, puzzled about it for a long while before noticing the pump/AIO was broken. I gave up using AIO.
Similar thing that still (20+ years later) has my missus ribbing me about it. Non computer, but similar d'oh moment. Old car, fog lights at the rear not working, I checked the fuse, feed to the switch and was at the point of sitting with my dashboard on my lap, she says "did you check the bulbs?" .... Both fail together, don't be silly...another half hour later and I have finally got to the bulbs...yeah, they were blown. This is your 'foglights' moment. Welcome to the club of the sillies!
Stirring up whatever cloggs it won't do it. Everything needs to be flushed.
You can see the relief on this mans face
Just some guessing from what you tell and what I can see in the video:
1 - Algae? The transparent-ish stuff that blocks the fins in the CPU-cooler look very much like algae that has been building up over time.
2 - At what temperature rating of the tubes and other materials used in the water cooling system you're using? I'm asking is because it could be cause initially (thermal paste?) that later led to another cause for the overheating. The 'gunk' could be reminisens of plastic/coating/etc. from materials that didn't handle the high temperature(s).
Whatever the cause is I would've installed a filter on your water cooling system because there are very likely more 'gunk' hiding in the water cooling system. You could just use a simple gasoline filter for cars, those with clear plastic and a yellow/orange paper filter inside. They are very easy to install and with the clear plastic you can easily see if the filter is clogging up.
Every other year I disassemble my Cougar Helor 360 and clean out the deposits from the pump-head distro plate. I dumped and flushed out all the original coolant the first time and replaced it with distilled water and since then NO real deposits formed again. The key was agitating the rad for many minutes while repeatedly flushing the rad to remove all of the sediment. The original coolant was the culprit, and have used distilled water ever since...that was 5 years ago.
I had the exact same issue two days ago, with a Kelvin S36. But in my case, it happend after I installed a new PSU. I think moving my PC during the installation caused the clog, which eventually cleared by itself. But it presented itself in identical fasion: first, the super high temperatures, even after I restored stock ratio and voltages. At one point it even shut itself down due to the high temps. Then, I noticed that the block was *really* hot, even though there was no hot air coming out of the radiators. My original assumption was that there must've been some UEFI settings that changed after a Windows update, so I unplugged the pump from the MB and plugged it back in, which did the trick. But your explanation make much more sense. This Kelvin S36 has been going for nearly a decade now though, so I'm gonna start looking at a replacement rather than risking a leak if I try to clean it.
Thanks for the vid, I did not expect to find a sensible explanation for what happened, especially by accident.
something that needs to be a trick of the trade when cleaning watercooling objects is cycling water through the pump while one tube is disconnected to have a open loop filteration with 1/4 rubbing alcohal and 3/4 water for 100 percent maintanance.
56 degrees Celsius isn’t scorching hot, 100 degrees Celsius is the limit.
And this is why I only air coolers in my computers even both my AMD 5950X builds.
LOL, using camera sensor cleaning swaps for removing the gunk in the AIO.
I had a similar thing happen the only time I tried to use liquid cooling. When I built my new-at-the-time i7-8700K gamer I tried to use liquid. The first nice cooler from Newegg was DOA. I shipped it back and just went and bought an OK one from Fry's. Less than a year later I started to get thermal throttling. I could sense through touch that the pump was moving and all of the hardware monitors looked good. Eventually (about two weeks) I found that there was gunk (corrosion or something) inside. I just returned it under warranty and bought a nice air cooler and that was the end of my problems. With an air cooler you know it is working if the fan is spinning. If you don't REALLY need extra-high-performance cooling then sometimes Old School is the Best School.
I didn't have to worry about any of this with on my first PC: 4.77 MHz 8088 loaded with all 640 KB memory and a 20 MB hard drive. Ambient cooling only...
Then back when active cooling first became necessary there wasn't any of this thermal throttling stuff. If you accidently turned on the PC without the heat sink you just put in a new processor.
I also had the same problem with my Aio from fractal design. When I googled it, it was a know problem and there Allready was a form online to fill out and get a V2 of that AiO. 1 week later I Allready had my V2 delivered and could keep the old broken one. It kept the fans and threw the rest away.
It was the mixture of the cooling liquid. The liquid in combination with the materials inside (hoses, copper aso) lead to the buildup of this stuff inside which would clog. Probably the same as with your aio.
AIOs are neat idea but the reality is they don't usually perform better than a good air cooler and you never know if you are getting gunked or not.
Watercooling is a complex system that always makes problems. Lots of interactions between materials and liquids and air and it just created problems. Still my next computer is going to be water cooled again as i really wanna try extracting the heat and pushing it outdoors.
I've been doin this building and repair thing for 22 years and I've always felt water cooling isn't worth the hassle.
Thanks for reinforcing my beliefs 😆
This is why you need use DISTILLED water , also add some water-wetter / something to kill bacteria.
same reason in a car radiator it clearly says use distilled water and NOT tap/rain water and mix with coolant which kills bacteria / fungus etc. Also a filter is a good idea and an indicator for flow-rate. As you noticed , tons of dirt , algae , minerals and stuff growing in the water-loop is not making the computer happy. Using just normal mixed car coolant will probably work ok.
Also note a car cooling loop has way bigger passages than your cpu block, making passages this tight is asking for trouble as a grain of sand will block it up. My first water cooling system used effective 2-3mm hoses and later I upgraded to 8mm hoses , the increase in flow made it orders of magnitude better. Just drilling entry holes on my first waterblocks doubled its perfomance alone.
I've had the same problem as you, a long time ago. The computer turned off sporadically and all it did was beep indicating CPU overheating. After several tests I came to the conclusion that it was the liquid cooling system. I opened it up and it was stuck. I "flashed" it with isopropyl alcohol, i mean, i replaced all the coolant with alcohol putting the open pump/dissipator in a container filled with alcohol without the dissipator. You can't even imagine the dirt that came out. I had to replace the alcohol 3 times. After thar i replaced the alcohol with coolant for the car and it has been working for me for about 6 months.
I would absolutely try to get in touch with EK about that cooler. I'm sure they'll send you a new one and they probably want to know what the cause of the clogging on yours is
Death Valley has entered your PC.
artic had that issue, they send you a repair kit within 5 days just within two emails. same day response. kit was a new plate, some liquid and a seal.
important to note is they informed customer before it went all bad for most of us. so it was a replacement in advance to not build up stuff in the first place.
best service ive seen in decades by an IT company asside from APC (2 days replacement batterys for the UPS just based on a serial number)
I have at times wondered how many hours I have spent over the years chasing down computer problems. I built my first PC at the end of 1983. In the era of no plug and play, dip switches on cards, limited and no sharing of interrupts, even in the first gen ISA 8 bit slots the card went into could stop some cards from working, that did persist even after the 16 bit slots were introduced a year or two later. As you may guess I continued building each new CPU release, 286, 386, 486, pentium, pentium 2 and so on. In the previous decade computers have become child's play to build. But as you can assume with this history I have spent many an hour scratching the head and clocking up unproductive time. I'm glad you found your problem, for me a replacement would be high on my budget list.
Would have gone a few steps further and spray out those fins of the cooling plate with a waterpick and flushed out the rad / refill it. Just cleaning the plate seems like the problem might be resolved now but its likely to come back at a later time. Either way, great job solving the issue.👍
This is one of the reasons why I just stick to air cooling.
Less to go wrong, and when it's not working, it's pretty obvious.
I feel like you should notify EK with any serial number you have, because if there's a pattern amongst a group of coolers, they may be able to identify a larger issue. If it can identify a pattern, it could be helpful data.
I used to work for Siemens Medical, and we used a circulating pump to cool X-Ray tubes used in certain kinds of cardiac cath labs. There was a requirement to use only distilled, demineralized and deionized water in those coolers (which, in that case was mixed with ethylene glycol). The reason was that tap water is just corrosive enough to cause problems with heat exchangers, possibly the problem you had. Obviously, such a system operates at above room temperatures, and we have all seen how things deteriorate outside, with copper turning a powdery brown under the best of conditions, and then green as biology takes over. Maybe it is the cooling water? Lab grade water runs around $15 a gallon, by the way...
If it starts overheating again, try giving the water block a few light taps.... If the temperature goes down, you just diagnosed it!!
I sooo want you to take that AIO again and push water pressure through the rad to flush out all that gunk, then refill with proper glycol and anti corrosion/bacteria/fungus additives.
I had a similar problem, idling and working temperatures began to go higher, than one day I came home to find the system off, with a log that showed it shut down due to overheating- on idle. tried turning back on, and it went up to 90C. Checked the water loop, and it was thick green, and the pump had burned out due to being clogged with - an algae bloom! First time in 15 years of water cooled systems that that has ever happened.
*World's First Ever Angioplasty on AIO Cooler!*
I did not know those things required open-heart ❤ surgery.
Can't wait for your video short about the inevitable transplant procedure.
Nice video Doc 👍
Also send that schmegma to a laboratory for chemical analysis please. Would be interesting and super nerdy.
Congrats for 3 Million subs!
Right on! I hope it didn't damage the CPU
I hope you saved the gunk to see what kind of solvent it takes to dissolve and wash out. The whole thing makes me thing of incompatible chemicals in the coolant or system. Who knows, the coolant might be dissolving part of the supply system.
I’ve only had one time I’ve ever had an issue with cpu temps and it was with a stock Intel cooler that had the plastic mounting mechanism go bad during a reapplication of the thermal paste. I simply replaced it with the cheapest tower cooler I could order and it has been working flawlessly since.
I would have put the cold plate into an ultrasonic cleaner with distilled water to remove the particles in the fins, then do the rubbing alcohol afterwards for any that remains. That would probably only improve the temps a couple degrees more than you saw at best, but it would be a relief to know that you actually got as much out as possible to prevent a reoccurrence.
Seems like minerals as you mentioned.
It's all part of the learning process so don't be too hard on yourself. In '95, I have struggled to get Internet working again after trying out all kind of software/driver/resintall Window only to discovered that dialup modem inside the computer got fried somehow. Was a learning experience not to only look at the software issue but on the hardware side too. This change of my thinking helped me to solve the other hardware issue that I had too instead of wasting countless hours trying to fix the issue on the software side.