To anyone looking to get an EK direct die kit... I would definitely wait another 6 months and let them sort things out. A lot of creators are hiding issues. I've had mine for a little under a week and just got stable temps with my 5th re paste/ re mount. They are worse than using the regular velocity 2 block with IHS. I wouldn't be posing this but it seems like this issue is common so definitely wait.
Ive never messed with cpus, but have been considering it. Was planning on getting the contact frame and maybe lapping the IHS. That seemed much less stressful to me, compared to de lidding. Would something like that still worth doing? Using something like nzxt elite or maybe the nucleus, considering de lidding can be much more problematic, especially with what you've brought up here.
@@Deltatwo3 lapping the top of an IHS is pretty straightforward, I’m not worried that I wouldn’t be able to do that at all. Just wondering if it would be worth it. That, combined with a contact frame for a bit better temps. Btw how else would I learn, if not by doing it? Pretty confident I can sand down the top of a flat piece of metal, especially with the tools available now. It’s the same when learning how to build your own pc for the first time. You just do it.
Given the direct collaboration between KitGuru and EKWB on this product, I expected a followup (from KitGuru) on EK's public admission that some direct die kits simply do not make sufficient contact and cause the chip to throttle. Instead of thermal improvements, many customers have received instead many wasted hours, disassemblies, and damage to their rigs. In other words, the exact opposite of the product's intended purpose! You told us how great it is, and encouraged us to buy it, now do the right thing and post a followup.
What these videos often lack, is to tell how tricky this kind of installation actually is. Even experienced builders can often find themselves re-applying the metal 3-5 times before getting the right coverage, correct contact frame pressure, and correct pressure for the block. If you're not an experienced builder, I'd suggest that you lap the CPU, use a contact frame and Kryonaut Extreme. That will still net you an 8-10 degree drop and be more consistent. I have delided many CPU's and done this for a while to say the least, and I as well as others, often don't get it right in the first try. It's often post issues, or if not that, redoing the metal so that you don't have 2, 3 or even 4 cores that sit higher than the rest. The product is fine, it's just far from as easy to use as some would like you to believe.
I got a i7 14700k and delidded it as soon as I got it home and used direct die running it at 5.9 all core stable without a sweat. Direct die is definitely the way to go with 14th gen.. well that is if you're not scared to delid a 400 dollar CPU and I did it with a 8 dollar delid tool.
@KitGuru lots of people having poor performance on the EK DD, and i think we figured it out, the cold plate is made convex which i think is wrong, convex plate is good only for proc with IHS, on direct die it should be as flat as posible
I know the 7000 series does gold with direct die cooling, but I've never been able to see a direct comparison of stocking speeds with direct cooling between AMD and Intel. I know it's a niche market, but I'd still love to see an apples-to-apples comparison and see just how crazy the performance increase would be for both across the board. I'd also like to see if EK would remake the sub ambient cooling mod they had for previous Intel CPUs. I'd love to see just how fast this thing can get with direct die sub ambient cooling.
I remember the first time I delided my Intel core i7 4790k cpu in 2014, back then we didn't have these CPU "Delid tools", I had to use a metal vice, a block of wood & a rubber hammer to get the IHS off. 😮💨
Interesting idea. When I removed the Intel ILM from my motherboard to install acontact frame, the ILM jumped out like a spring and bounced back on the socket. I was sure I had damaged some pins but everything was ok (a miracle). I don't know why the ILM did a "jack in the box"; probably the spring on the arm. So yep, next time I will protect the socket!
I think this is a good safeguard for the more clumsy or inexperienced of us but in my experience I have found that alot of these components are alot more durable and robust than we give them credit for.
almost wants to make you embrace the horror of de-lidding and faffing about with liquid metal. Almost... (but 20C improvement is hard to argue with!) Bit of a disappointment with the de-lidding tool though. I think that's entirely Der Bauer's design from watching his channel where he was developing it. Another very informative in-depth video James.
According to EK the delid tool fault is an isolated issue. It's going back to EK for investigation. Looks like it could be a manufacturing defect but not sure yet.
I delidded a 13th gen CPU and found it was easier with a heat gun. I wore some gloves and heated the thing for 15-30 seconds, tightend the bolt 1-3 threads, put it back on the heat for 15-20 seconds, checked the screw. Eventually the screw would move more freely, and going back and forth adding a little bit of heat and screwing the heat spreader I was able to get it off without over wrenching anything.
I’ve got a 13900ks and thermal grizzly 12/13th gen contact frame and a 420mm Corsair AIO. I’m definitely not advanced enough to be doing this, but it’s fun to watch!
Thank you for this great video! I'm currently using an Arctic Liquid Freezer II AIO 420mm and wish I would have an EKWB Quantum Velocity² Direct Die Cooler. The heat transition from Die to IHS to the AIO is that bad that it nearly doesn't matter how big the Radiator is. The liquid doesn't get any warm and I can keep it very quiet. But on the other side the thermals of my Core i9-13900KF quickly raise up to 100 degree Celsius, no matter how fast the fans are spinning. The weak point is the IHS. I can't get stable beyond 6.00 GHz or 5.6+ GHz All-Core and 300+ Watts result in 100+°C.
This is not as easy as it seems. Do the following. Keep your 420, get the Thermal Grizzly lapping tool and a contact frame. Lap the CPU, use Conductonaut and the contact frame. Getting rid of the nickel layer makes a big difference. One important thing if you go this route. Apply the metal. assemble and leave for a week. Disassemble again and wipe both the IHS and the plate on the block free of metal. Now the copper surfaces on both the IHS and block should have a layer of an alloy which is a mix the metal and the copper. Now reapply a new layer of metal to both the block and IHS. This way you get proper contact, and your application will last for as long as a quality paste. This way you should see 75-80% of the performance benefits of the DD solution.
I went direct die on my 9900k several years ago, and I dropped like 17c...I think I will do it again on my next system...It's just something fun to do, and it doesn't hurt that you can overclock more!
@@toseltreps1101I saw a video of 7000 amd cpu on idle using 50w while 10w for Intel we need to stop only siding with 1 firm because if we all went to amd it will be same as Intel as you can see 14900k is literally the same as 13900ks lol meaning no improvement at all
It's GOLD!... then they apparently cheap out on the delid tool? THAT is the bit you HAVE to get right; NOT the ruddy aesthetics! Not the kind of poncing about you want to see from a company in our current economic straights! If this kind of mod makes a noticeably difference to an energy bill while taming the somewhat ridiculous heat-load Intel have saddled themselves with in recent gens, we might be interested. But wasting time and budget on useless bling while your essential tool craps out? Not a good optic; I'd LIKE to give this direct die de-lid thing ago... But I think I'll look at what others might be up to, tah very much EK. Another fine video, James, that and the actual testing went well. Which is the important part as far as you and Kit Guru are concerned. looking forward to what you bring us next!
reviewers hiding ther real temps and problems, now lots of complain from DD users and EK is still in denial, EK DD losses to competitors by around 15c to 20c if you ever manage to make it work and not kill your proc
I have the kit, and temps are kinda off. 77 on core 0 and 95 on 2 cores. Had to sand down the direct die frame to 4.h mm to get 85, i now sanded it down to 4mm and lapped the cold plate. Will probably assemble it tomorrow to see if i got it up to par now
@@MrFlipcas it's better than i thought it would be, 72 degrees in CB23 after the 10 minute test. 4mm direct die frame + lapping cold plate really did it
@@MrWhizeGuy really appreciate the feedback. I just delided my cpu but yet to install the waterblock. I’ll be using PTM 7950 instead of Liquid Metal for better tolerance. I’m wondering, in case I need to sand down the contact frame, what # grit sandpaper did you use?
love EKWB, but I have to be 100% honest , sometimes their quality control seems to go astray. This seems like an easy fix mind you, but it might take some time. ill hold fire on ordering the kit. Glad you folllowed up - I liked derbaurers video, but as he is financially invested, I wonder if he will cover every single issue like you did. But he had a different kit is that right?
Der8auer has publicly stated (in one of his videos) that he doesn't make money on the product, which is the same deal he had with Lian Li for the O11 dynamic. I am not a lawyer or familiar with German/EU laws, but I'm relatively sure he'd be breaking the law stating he doesn't receive money for a product while receiving money for a product. That being said he certainly is reputationally invested as he's name is prominently displayed on the product due his contributions with the design and validation of the product.
@@fuglynugget I certainly never meant to be saying he was doing anything sinister I watch all his content and admire the guy immensely. I just think James seems to have put a lot more work into this kit than anyone else, finding all the issues and reporting them (still!). Seems like quite a massive issue if its not a one off. James never said it was a 'global issue' but that aluminum screw doesnt seem to be ideal for the job. Would stainless steel not prove more robust?
The delid tool stripping is DEFINITELY a durability issue. @JayzTwoCents just posted his video with this kit, and he had a similar issue. The threads didn't get chewed up like yours, but the tool probably couldn't be used again after 1 CPU went through it.
Yeah mine also had the issue, and the IHS didn't want to come off even after 10 passes. I kept going and my die eventually just came off with the solder lol. I wish I just stopped and ordered a different tool, but I didn't expect the tool to be that bad honestly so I had misplaced faith with continuing to try.
youtube wil probably automod this but yeah even some of the PCB came up with it cdn.discordapp.com/attachments/660082151810662421/1083843178140029021/IMG_8950.jpg
@@KitGuruJames At least it's not all bad, EK accepted fault and refunded my 13900KS purchase yesterday. Unfortunately my chip was from Igorslab, when they tested 200 CPUs, and was 7th best, so I can't replace it with a prebinned one. Gotta hope for the lottery.
I'd suggest inspecting the thread of the bolt for rough edges or damages and adding a bit of grease to it before putting into the delidding tool. Would be a good idea to look for a longer bolt with really smooth thread surfaces as a replacement. And press against the slider with your other hand while using the tool. It should survive a bit longer when having less resistance to overcome.
Finally, a video that shows the price of items tested. All I have to do is convert EUROS to Dollars. I'm sure I can find an app. for that. Shipping charges for items ordered from Europe are unbelievable. I ordered a torque screwdriver from EK recently. It cost $20.99, and DHL shipping was $25.90. The screwdriver weighs 90g. Does EK have US dealers?
Personally I'd mask the block leaving an area corresponding to the die (+ maybe 1mm) to be coated with LM, using the absolute minimum needed to wet it and allow the larger application on the die to flow over it readily - the tape allows you to vigorously work a tiny amount of LM across the area you want coated. Otherwise It's effectively two applications of LM.
I have found much better results when I roughly the same amount of LM on the die and the Block...But it really doesn't take that much! Most people tend to over do it with LM...
You're talking a 23-27 degree drop. Not possible all things being equal. Could you imagine a 14900K or 13900KS running 325 watts at 65-67 degrees? Nope.
i ran into the same problem with their delid tool. the glue on the sides of the ihs wouldnt break free. i sharpened a guitar pick with sandpaper to get in between the ihs to cut the glue and it finally popped off.
Same problem here. The glue just wouldn't budge. Had to pop it off with a very thin plastic "scalpel". Unfortunately I get really bad temps now. 75-80c in BIOS. 100C with any type of load. 😢 Contact pressure appears to be bad, even with the screws torqued to the max. I re-seated the block twice already and cannot figure out what's going wrong.
@@MrDuneTwo yep there was a lot of glue under the side tabs of the ihs and the tool doesn't push it far enough to break it free. I have yet to install the block but ive been reading everywhere that a lot of people are having those contact issues. did you hear anything back from ek?
Has anyone had any response from ek about the high temp issues? I'm having the same issues in getting up to 100 degrees Celsius whereas kitguru gets 75?
How many times did you try andedo your liquid metal? Have you any prior experience with direct die and liquid metal cooling? I ask because it is not as easy as it seems. Even experienced builders can often find themselves reapplying metal several times before it's right.
i'd personally apply like half the liquid metal. seemed like a lot to my eye through the camera. it really only needs a 0.01mm layer, and that stuff likes to squeeze out if there's excess
have 2 of these direct die coolers, cannot wait to test them. i was gonna use them on 13900k but now im definitely gonna use the 13900ks thank you for th vid :D
Considering the weight of some fittings and tubes, does that affect the contact if the tubes are pulling to one side? Thinking especially in a vertical mount rather than the test bed laying down
crazy the gains...**proceeds to show next to zero numbers** what about the cinebench r23 difference? how about pushing to 5.8-9-6.0 in cb to really see gains?
Hab ich mit einem 13900k gemacht. Hat viel zu viel Strom verbraucht :) Ist jetzt mein Zweitrechner. Der neue ist R9 7950X3D und ne 4070. Da passt das mit dem Strom. Aber der Kühler von EK ist der Hammer, bringt aber nichts wenn man normale Wärmeleitpaste verwendet.
I just bought one of the upgrade kits for my 13900KS because I wanted to keep my acetal block. Was excited to use the tool as it was gonna be my first time delidding and Debauer made it look so simple. Now I'm not so sure. Think I'll sit on it to see other reviews first and make sure it's not a manufacturing error...
@@rock962000 it was always out of stock. Finally picked up a rockitcool kit this weekend. Just gonna use that. A bit nervous now cuz now I'm hearing the direct die kit is having contact issues for some people and making temps worse. Something about the plates being convex. Man I'm so nervous about this lol maybe I'll just stick to my regular block for a while.
@@Rashimotosan That's so weird, I must have been lucky cause it was in stock when I bought it and really soon after almost everything on their website was out of stock.
James, love your voice - accent is so unusual ! im french though. This kit looks lovely and nice to see EKWB appreciating all of kitgurus help in the last 2 months with the letters etc. I like the kit, but like others, im a bit concerned ordering it due to the stripped screw - that seems like the screw needs beefed up to deal with more torque pressure. Let us know if you get any word from EKWB on the fixes
Do you need warm/heat up the processor before you start deliding process ?? (when deliding 13900KS processor) or is it only EK Delid Tool needed ?? many thanks Bartek BIG fan of you !
I was wondering this too. My tool should be in next week and I'm nervous about all these snapped screw/stripped threads issues... I was planning to use my heat gun eapecially for the first few passes to loosen up a bit.
Yep i have one of these. Ended up reseating it 8 times before I got it half working right and even then its not contacting one set of the ecores correctly and they thermal throttle. I have reached out to EK with no response so far. Also had to ditch the foam as it couldnt be reused after reseating so many times and was getting in the way.
Is this a in combination with a water pump in the CPU water block or with still be connected to the water pump in other words is it liquid cool or water cooling.?????
Hey KitGuruTech. I have done the same thing to my KS last week. I also stripped the threads out of my EK . I think they need to use a longer screw when engaging pressure on the threads. Its just partially in there when you start applying pressure. I hope that is all it is. I did go out and get a new m6 x 1.0 screw from the hardware store that was hardened. But my top slider threads were stripped as well. I was able to retap it but it and finish but still the beginning threads are gone. It took me over 12 times. I also noticed on my two (2) 13900KS there was a lot more solder than The 13900K i've seen on other delid videos and 12 series CPUS as well. Also. i tried to use the liquid metal to take off the old residue. I left it on there for 10+ mins each time but there was too much solder on mine. I had to finally razor blade it off. took 5 sharp blades.
@@KitGuruJames I did. They first were going to send me a screw but I had to show them that my delid slider screw hole was also stripped. So they told me to RMA it. Just started that today
@@3dprintingdentist599 ah.. mine just arrived today. and it's going to be used on a 13900ks. does this look like a one off manufacturing defect or you reckon it's a fault in design? should i just reach out to them directly or attempt to use it (would it cause irreparable damage to the ks)
@@Violentmercury I bought longer screws and had to retap the top. I think the longer screws will help with the engagement of more threads when a lot of torque is applied. The screw stripping should not cause issues to the actual cpu itself.
@@StuartBoyer So you just took the foam off and the Temps were what you would expect? i also used a generous amount of liquid metal but apparently something is wrong.
@@TinyChrisTV Original piece of foam they gave me was mis cut and had gotten in the way. removing it helped some but its still overheating. EKWB has admitted that they have some issues with the die guards and I just now got the replacement one but they still are arguing hard that the miscut coldplate isnt an issue when people doing pressure fit test show clearly it is which appears to be my case too.
@@StuartBoyer Can you tell me how exactly you got into contact? Normal Support Ticket? I did remove the foam Piece and the temps are now somehow worse... is the contact frame the issue or their coldplate? Because i already had a Satin Velocity2 and just ordered the Upgrade Kit. But i should have assembled that correctly.
Nice video, been waiting for something like that for quite a long time. All this work and money spent for only 14 degrees. Unless there's a much sheaper way to do this, it's not worth it.
Der8auer was interviewing an Intel thermal engineer who suggested that under load the CPU should run up to 90C and mange it’s boost clocks around there. If the chip under load is sitting lower then it will be leaving some performance on the table. Have you pushed this setup for additional performance?
I think you have to look at this as improving the temperature headroom and being limited by some other factor. Running at 90C isn't going to unlock performance.
@@antoniog8276 yeah agree, this is the point I didn’t make clearly. If there’s now more temperature headroom then there’s likely more performance to unlock. I was interested if they were tuning for that headroom or just leaving stock.
Great video james, love how you always cover all the points and are completely transparent. its invaluable in a world today based on percentages and sales figures.
Those gains are immense, but it seems like right a ball ache atm to get everything working. might be worth leaving it a while until they find tune this
It's come a long way since the first prototype that I looked at. The installation process is very easy now. However the thermal improvement could depend on how good the stock CPU to IHS contact is too.
@@stickyfingers9710 I tested one of the late prototypes with a 12900K and didn't see as good improvement. however the contact frame design has had a tweak since then.
doesn't look like the screw goes all the way in before it starts pulling on the IHS, so at least one of the problems is the short screw they include. I'm sure it delids fine for most people, but still that is quite the fuckup. impressive result thermally though.
I wonder if POST and memory bus issues are sometimes caused by not dressing the coolant pipes carefully enough to ensure they're not exerting any force on the block.
No. There's kind of a rule of thumb when using contact frames of a good quality. Finger tighten to zero lash. Then tighten with a screwdriver 90 degrees.
What TIM did you use on the aio and other block? Maybe worth seeing what it's like non de lidded with liquid metal. I know you aren't meant to but I tried it once on a 7700k and dropped like 10 degrees.
@James Dawson ok cool. Thank you for answering. I suppose it does make sense to test it against a more typical set up that some one would use. Great video by the way. Think im still too scared to delid a cpu though lol.
Did they fix the gap issue from the prototype w/12900k? I’ve noticed a few people are having gap issues on some groups with the 13900k. Could be user error but also could just have been a tolerance off
I heard some reports of die contact issues too. I thought this was fixed but we will have to wait and see what the outcome is. I believe EK is looking at this now.
I don't think 6GHz all core will be possible with this CPU and direct die cooling. 5.8ghz was not easy to get stable, that's why I had to back off to 5.7 for reliability.
Thats crazy, would've never realized we would come to a point of time where people are now removing IHS for custom water cooling, doesn't this void the warranty of the CPU?
Yes, and its been going on for ages - I did my first deliding over 20 years ago That said, it isn't without its risks - many people have lost their precious cpu's doing so over the years
Hello, very cool video, but what interests me the most is how did you get the CPU so clean, you could almost think your CPU has never seen any glue :D I've already beheaded one or the other CPU, but I always had quite a problem cleaning the CPU after beheading it because the glue still sticks well and you don't want to scratch it with great force. Maybe someone here has a tip on how to do it in one way or another.
@@SmalltimRYes, okay, but also the circuit board itself, you can't see any adhesive residue and you can't See scrape a scratch from the adhesive anywhere
@@jensbiernot312 Liquid metal will also remove residue as well. That said, there should be no scrapes or scratches on the die. The assembly process itself, is none-destructive, and the die material is much harder than the solder.
I tried direct die cooling way back in the athlon days,,,, mind you having the coolant splash directly on the silicon of the CPU not just a regular setup minus heat spreader. Besides back then CPUs did not have a heat spreader, well at least not AMD i cant remember if Intel had that. Heat spreaders i dont really get with today's big chips, you was much more likely to crush a tiny athlon CPU back in those days. It worked pretty well but it did entail quite a bit of DIY / improvisation.
It is compatible, but i had better gains on 13th gen. However I was using one of the last prototypes for 12 th gen testing and there have been some changes to the die guard since then.
Really impressive direct die cooling. I'm ordering one right now, wondering Koolance Quick Disconnect are those, really would like to copy your setup on my test bench. Not sure but im guessing, EK-Quantum Torque Extender Rotary MF 14 --> QDT3 EPDM --> QDT3-M10X13 ?
The kf runs hotter than ks so it would make more sense to use the kf. I'd recommend to put the liquid metal to the q-tip rather than directly to the cpu sometimes the liquid metal can go everywhere. I'd rather it go allover the q-tip rather than your motherboard!
notice non of the tech tubers have a propper benchmark and test? cause the product is defective, lots of EK DD already killed their CPU, its a shame, popular tech tubers just repeat the BS on this product
Dude my delider did the same exact thing! Now I am stuck with a have delided 13900k! I reached out to derbauer about trying to get a replacement because now my computer is down and ek won’t have more for another month! Please let them know about this issue so we can try and get replacements because I am so boned right now since I don’t have another 600$ for a new 13900k. Thank you soo much for showing this because I figured it was just mine that messed up but it seems the tool isn’t strong enough for the multiple passes required to remove the lid. Also I can’t even buy one from rockit cool because they are sold out as well. I don’t know what to do atm and am royally screwed. Please let them know about this problem
@@KitGuruJames yeah I tried that as well but I can’t get a good enough contact with the tool to pull it forward to move the lid. The threads on the tool literally came out like on his video and I can’t get anything to work. :(
You could order the normal Velocity² block and a direct die upgrade kit instead. There's a golden version too. EK also sells these metal parts separately, if you want to spend the money for that. You just have to contact customer support and tell them what you want. Since you already got the transparent version, that's probably the cheapest option for you.
Somthing failed on my motherboard after putting new cpu paste on and Iooked in and I must have caught one pin on the edge of the cpu and bent it completely around waiting on a new motherboard so I can use my pc again
I have to see what's going on. I'm still thermal throttling with my EK Direct Die kit on Cinebench. I'm not getting anywhere near these temps with my 13900KS. I have tried 2 different boards as well. I have an Aorus Z790 Master and an Asus Z790 Apex. I'm running two hardware labs Nemesis GTS 360's.
I think there can be chip to chip differences but it sounds like you are getting poor block to die contact if it's thermal throttling. What CPU frequency and vcore are you using?
I’m getting the same with my ks. Thermal throttling even with the kit, a delided k model I had with a copper ihs and the same block prior to the upgrade kit did not do this
@@KitGuruJames Contact should be good. Mounted it on 2 different boards as well. Running the cpu stock just have XMP enabled. Vcore is 1.25v under load.
@@StuartBoyer are you ruining the chip stock? Did you use a torque driver? I can't believe I'm thermal throttling. I have a pretty good example of silicon as well.
Have you ever looked at the direct die kit from Supercool Computer in Thailand? Very different approach with it's own upsides and downsides but it would be really cool to hear thoughts from an expert like yourself and maybe a performance comparison as well
@@KitGuruJames It's an interesting concept - instead of installing a traditional waterblock using the normal mounting holes, the cold plate itself is in the same form factor as the original IHS, so you merely replace the stock IHS with the cold plate and tension it in place with the original ILM, then you screw the rest of the block onto the cold plate. One the one hand it eliminates the worry of over tightening the block and cracking the die, but the design seems like it would put additional pressure on the ILM and you essentially can't remove the block without draining the loop first
It looks like it works great but it also cost as much if not more than the CPU it is cooling. I would like to do this kind of thing but I'm not willing to dump $500 + just for cooling.
Derbauer has direct die frames for lga1156, 2066 and am4. You replace your oem socket mount with them and reuse the cooler you already have. Since he was helping them design that i expect am5 and lga1200 versions to follow.
Those are some serious gains. but im a bit concerned with that tool that broke. is that confirmed to be an issue? I trust your knowledge james and while i know you are being diplomatic, thats a bit of a concern seeing those screws strip. Is that the best material to use for those screws?
The screw is high grade stainless steel. It was the thread of the delid tool that stripped which is aluminium. According to EK there has only been one other report of the slider thread stripping like this. EK is looking into it. There were some reports of screws breaking on early versions but the screw has been upgraded since then.
@@KitGuruJames I just did now because of this video. I was looking for alternatives but can’t find any and rockit cool doesn’t have delid tools in stock. It sucks having a half delided 13900k and no computer :(
Yes I set a stable oc/fixed frequency and vcore to ensure accuracy of testing the different coolers. P-cores at 5.7, e-cores at 4.5, vcore is 1.3v in bios. Load line calibration set to auto and all intel velocity boost features disabled.
@@KitGuruJames Thankyou I'll set mine to similar to see if mine is doing right. Did you get thermal throttling if left at default where it can hit 6ghz on two cores like mine?
@@KitGuruJames Thanks. I ended up taking the foam off as it wasn't a good fit to begin with and directions have you add the foam to the sticker and add that to the card unlike in the video and the sticker wasn't cut right it seems. cleaned and reapplied the liquid metal and temps dropped some more. Needs a better piece of foam and sticker, QC has been lacking a lot with EKWB lately. I had bad seals on a 4090 ABP set and a broken LED strip out of the box and QC on this kit seems to need some work. Good news is no thermal throttle now. Got 93 degrees on all core 5.9 with all e cores at 4.6 and got 42684 in cinebench, sure I can tune it better but that was only from changing the ratios. Definitely needed in my case to dump the foam and go slightly heavier on my liquid metal than I did with my 13900k and copper IHS.
To anyone looking to get an EK direct die kit... I would definitely wait another 6 months and let them sort things out. A lot of creators are hiding issues. I've had mine for a little under a week and just got stable temps with my 5th re paste/ re mount. They are worse than using the regular velocity 2 block with IHS. I wouldn't be posing this but it seems like this issue is common so definitely wait.
yup lots of people having same issue high temps if you get it to work and lots of failed contacts some killed thier procs so sad
Ive never messed with cpus, but have been considering it. Was planning on getting the contact frame and maybe lapping the IHS. That seemed much less stressful to me, compared to de lidding. Would something like that still worth doing? Using something like nzxt elite or maybe the nucleus, considering de lidding can be much more problematic, especially with what you've brought up here.
@@Zjombie if you've never messed with cpus and you're going to try to lap something? I would not do that. No wouldn't be worth it.
@@Deltatwo3 lapping the top of an IHS is pretty straightforward, I’m not worried that I wouldn’t be able to do that at all. Just wondering if it would be worth it. That, combined with a contact frame for a bit better temps.
Btw how else would I learn, if not by doing it? Pretty confident I can sand down the top of a flat piece of metal, especially with the tools available now. It’s the same when learning how to build your own pc for the first time. You just do it.
@@Zjombie then why are you asking me?
Given the direct collaboration between KitGuru and EKWB on this product, I expected a followup (from KitGuru) on EK's public admission that some direct die kits simply do not make sufficient contact and cause the chip to throttle. Instead of thermal improvements, many customers have received instead many wasted hours, disassemblies, and damage to their rigs. In other words, the exact opposite of the product's intended purpose!
You told us how great it is, and encouraged us to buy it, now do the right thing and post a followup.
thanks
Not on point. Installation is half the problem of any thermal solution.
What these videos often lack, is to tell how tricky this kind of installation actually is.
Even experienced builders can often find themselves re-applying the metal 3-5 times before getting the right coverage, correct contact frame pressure, and correct pressure for the block.
If you're not an experienced builder, I'd suggest that you lap the CPU, use a contact frame and Kryonaut Extreme. That will still net you an 8-10 degree drop and be more consistent.
I have delided many CPU's and done this for a while to say the least, and I as well as others, often don't get it right in the first try. It's often post issues, or if not that, redoing the metal so that you don't have 2, 3 or even 4 cores that sit higher than the rest.
The product is fine, it's just far from as easy to use as some would like you to believe.
I got a i7 14700k and delidded it as soon as I got it home and used direct die running it at 5.9 all core stable without a sweat. Direct die is definitely the way to go with 14th gen.. well that is if you're not scared to delid a 400 dollar CPU and I did it with a 8 dollar delid tool.
@KitGuru lots of people having poor performance on the EK DD, and i think we figured it out, the cold plate is made convex which i think is wrong, convex plate is good only for proc with IHS, on direct die it should be as flat as posible
I know the 7000 series does gold with direct die cooling, but I've never been able to see a direct comparison of stocking speeds with direct cooling between AMD and Intel. I know it's a niche market, but I'd still love to see an apples-to-apples comparison and see just how crazy the performance increase would be for both across the board.
I'd also like to see if EK would remake the sub ambient cooling mod they had for previous Intel CPUs. I'd love to see just how fast this thing can get with direct die sub ambient cooling.
see SkatterBencher
I remember the first time I delided my Intel core i7 4790k cpu in 2014, back then we didn't have these CPU "Delid tools", I had to use a metal vice, a block of wood & a rubber hammer to get the IHS off. 😮💨
this is so awesome brother. i am wiating for my direct die block from ek. i got everything else prepped i am very excited to see how this turns out.
I shall delid mine when the EK parts come into stock!
EK having issues with their Delliding tool was very surprising.
Sounds like you’re not familiar with the loose screw waterblock issues. Look it up. EK are expensive but definitely not perfect
It would probably be a good idea to put the CPU into the socket while removing the stock ILM just in case you accidently drop anything in the socket.
Yeah, good point.
Interesting idea. When I removed the Intel ILM from my motherboard to install acontact frame, the ILM jumped out like a spring and bounced back on the socket. I was sure I had damaged some pins but everything was ok (a miracle). I don't know why the ILM did a "jack in the box"; probably the spring on the arm.
So yep, next time I will protect the socket!
I think this is a good safeguard for the more clumsy or inexperienced of us but in my experience I have found that alot of these components are alot more durable and robust than we give them credit for.
@@Deltatwo3 i know but i'd rather avoid a very long night of trying to bend the pins back :P
@@blue-dragon agreed for sure lol
almost wants to make you embrace the horror of de-lidding and faffing about with liquid metal. Almost... (but 20C improvement is hard to argue with!) Bit of a disappointment with the de-lidding tool though. I think that's entirely Der Bauer's design from watching his channel where he was developing it. Another very informative in-depth video James.
Thanks Trey, appreciate you taking the time to comment mate
According to EK the delid tool fault is an isolated issue. It's going back to EK for investigation. Looks like it could be a manufacturing defect but not sure yet.
@@KitGuruTech you sure you are not running at 1.32? You kept saying 1.3…. Either or z790 tachyon is 1.32
@@KitGuruJames Good to know. Could be a bit embarrassing otherwise...
@@stefanradunovic8318 vcore was set at 1.3v in BIOS but in HWinfo under load read 1.284v I believe. Load line calibration was let set to auto in BIOS.
I delidded a 13th gen CPU and found it was easier with a heat gun. I wore some gloves and heated the thing for 15-30 seconds, tightend the bolt 1-3 threads, put it back on the heat for 15-20 seconds, checked the screw. Eventually the screw would move more freely, and going back and forth adding a little bit of heat and screwing the heat spreader I was able to get it off without over wrenching anything.
I’ve got a 13900ks and thermal grizzly 12/13th gen contact frame and a 420mm Corsair AIO. I’m definitely not advanced enough to be doing this, but it’s fun to watch!
Just got mine in the mail finally. Not the gold one but my first direct die project. Got a 13900ks as well with a really good imc
how did it go ?
Yes, how did it go?
Skipped straight to the delidding process!
Thank you for this great video!
I'm currently using an Arctic Liquid Freezer II AIO 420mm and wish I would have an EKWB Quantum Velocity² Direct Die Cooler.
The heat transition from Die to IHS to the AIO is that bad that it nearly doesn't matter how big the Radiator is.
The liquid doesn't get any warm and I can keep it very quiet.
But on the other side the thermals of my Core i9-13900KF quickly raise up to 100 degree Celsius,
no matter how fast the fans are spinning. The weak point is the IHS.
I can't get stable beyond 6.00 GHz or 5.6+ GHz All-Core and 300+ Watts result in 100+°C.
This is not as easy as it seems.
Do the following. Keep your 420, get the Thermal Grizzly lapping tool and a contact frame. Lap the CPU, use Conductonaut and the contact frame. Getting rid of the nickel layer makes a big difference.
One important thing if you go this route. Apply the metal. assemble and leave for a week. Disassemble again and wipe both the IHS and the plate on the block free of metal. Now the copper surfaces on both the IHS and block should have a layer of an alloy which is a mix the metal and the copper. Now reapply a new layer of metal to both the block and IHS. This way you get proper contact, and your application will last for as long as a quality paste. This way you should see 75-80% of the performance benefits of the DD solution.
I went direct die on my 9900k several years ago, and I dropped like 17c...I think I will do it again on my next system...It's just something fun to do, and it doesn't hurt that you can overclock more!
or you can just not and go AMD where you get more performance at half the energy consumption
It hurts if something breaks though.
@@toseltreps1101I saw a video of 7000 amd cpu on idle using 50w while 10w for Intel we need to stop only siding with 1 firm because if we all went to amd it will be same as Intel as you can see 14900k is literally the same as 13900ks lol meaning no improvement at all
I I I I
It's GOLD!... then they apparently cheap out on the delid tool? THAT is the bit you HAVE to get right; NOT the ruddy aesthetics! Not the kind of poncing about you want to see from a company in our current economic straights! If this kind of mod makes a noticeably difference to an energy bill while taming the somewhat ridiculous heat-load Intel have saddled themselves with in recent gens, we might be interested. But wasting time and budget on useless bling while your essential tool craps out? Not a good optic; I'd LIKE to give this direct die de-lid thing ago... But I think I'll look at what others might be up to, tah very much EK. Another fine video, James, that and the actual testing went well. Which is the important part as far as you and Kit Guru are concerned. looking forward to what you bring us next!
I just ordered this from ek glad to see how well it works on the 13900ks
how did it go for you?
@@AnnMeteo still waiting for the part to be made.
Another great video James, always learn stuff from your content. good addition to the team!
Thanks 👍
been following this since the start, I wondered why you were taking so long to finish it. Seems EKWB are battling their own quality control
reviewers hiding ther real temps and problems, now lots of complain from DD users and EK is still in denial, EK DD losses to competitors by around 15c to 20c if you ever manage to make it work and not kill your proc
I have the kit, and temps are kinda off. 77 on core 0 and 95 on 2 cores. Had to sand down the direct die frame to 4.h mm to get 85, i now sanded it down to 4mm and lapped the cold plate. Will probably assemble it tomorrow to see if i got it up to par now
@@MrWhizeGuy looking forward to your results
@@MrFlipcas it's better than i thought it would be, 72 degrees in CB23 after the 10 minute test. 4mm direct die frame + lapping cold plate really did it
@@MrWhizeGuy really appreciate the feedback. I just delided my cpu but yet to install the waterblock. I’ll be using PTM 7950 instead of Liquid Metal for better tolerance. I’m wondering, in case I need to sand down the contact frame, what # grit sandpaper did you use?
love EKWB, but I have to be 100% honest , sometimes their quality control seems to go astray. This seems like an easy fix mind you, but it might take some time. ill hold fire on ordering the kit. Glad you folllowed up - I liked derbaurers video, but as he is financially invested, I wonder if he will cover every single issue like you did. But he had a different kit is that right?
Der8auer has publicly stated (in one of his videos) that he doesn't make money on the product, which is the same deal he had with Lian Li for the O11 dynamic. I am not a lawyer or familiar with German/EU laws, but I'm relatively sure he'd be breaking the law stating he doesn't receive money for a product while receiving money for a product.
That being said he certainly is reputationally invested as he's name is prominently displayed on the product due his contributions with the design and validation of the product.
According to EK, his block was the same revision as our limited edition block.
@@fuglynugget I certainly never meant to be saying he was doing anything sinister I watch all his content and admire the guy immensely. I just think James seems to have put a lot more work into this kit than anyone else, finding all the issues and reporting them (still!). Seems like quite a massive issue if its not a one off. James never said it was a 'global issue' but that aluminum screw doesnt seem to be ideal for the job. Would stainless steel not prove more robust?
@@justindavis7600 The screw was steel, james just told me its the slider tool thats aluminum .
The delid tool stripping is DEFINITELY a durability issue.
@JayzTwoCents just posted his video with this kit, and he had a similar issue. The threads didn't get chewed up like yours, but the tool probably couldn't be used again after 1 CPU went through it.
Yeah mine also had the issue, and the IHS didn't want to come off even after 10 passes. I kept going and my die eventually just came off with the solder lol. I wish I just stopped and ordered a different tool, but I didn't expect the tool to be that bad honestly so I had misplaced faith with continuing to try.
youtube wil probably automod this but yeah even some of the PCB came up with it cdn.discordapp.com/attachments/660082151810662421/1083843178140029021/IMG_8950.jpg
@@C0manso wow that is bad luck.
@@KitGuruJames At least it's not all bad, EK accepted fault and refunded my 13900KS purchase yesterday. Unfortunately my chip was from Igorslab, when they tested 200 CPUs, and was 7th best, so I can't replace it with a prebinned one. Gotta hope for the lottery.
@@C0manso great news that EK sorted you out.
jayc after watching this video: dang it! time to rebuild my pc again.
I'd suggest inspecting the thread of the bolt for rough edges or damages and adding a bit of grease to it before putting into the delidding tool. Would be a good idea to look for a longer bolt with really smooth thread surfaces as a replacement. And press against the slider with your other hand while using the tool. It should survive a bit longer when having less resistance to overcome.
Finally, a video that shows the price of items tested. All I have to do is convert EUROS to Dollars. I'm sure I can find an app. for that. Shipping charges for items ordered from Europe are unbelievable. I ordered a torque screwdriver from EK recently. It cost $20.99, and DHL shipping was $25.90. The screwdriver weighs 90g. Does EK have US dealers?
Personally I'd mask the block leaving an area corresponding to the die (+ maybe 1mm) to be coated with LM, using the absolute minimum needed to wet it and allow the larger application on the die to flow over it readily - the tape allows you to vigorously work a tiny amount of LM across the area you want coated. Otherwise It's effectively two applications of LM.
I have found much better results when I roughly the same amount of LM on the die and the Block...But it really doesn't take that much! Most people tend to over do it with LM...
Hi algorithm I really engaged with this content.
can you review the supercool DD and iceman DD thier saying both direct die out performs the EK DD by 10c?
You're talking a 23-27 degree drop. Not possible all things being equal. Could you imagine a 14900K or 13900KS running 325 watts at 65-67 degrees? Nope.
Great video James just waiting to see if EK bring out a direct die water block for AMD 7000 series. Fantastic content love the channel.
Lol
i ran into the same problem with their delid tool. the glue on the sides of the ihs wouldnt break free. i sharpened a guitar pick with sandpaper to get in between the ihs to cut the glue and it finally popped off.
Same problem here. The glue just wouldn't budge. Had to pop it off with a very thin plastic "scalpel". Unfortunately I get really bad temps now. 75-80c in BIOS. 100C with any type of load. 😢 Contact pressure appears to be bad, even with the screws torqued to the max. I re-seated the block twice already and cannot figure out what's going wrong.
@@MrDuneTwo yep there was a lot of glue under the side tabs of the ihs and the tool doesn't push it far enough to break it free. I have yet to install the block but ive been reading everywhere that a lot of people are having those contact issues. did you hear anything back from ek?
Has anyone had any response from ek about the high temp issues? I'm having the same issues in getting up to 100 degrees Celsius whereas kitguru gets 75?
How many times did you try andedo your liquid metal?
Have you any prior experience with direct die and liquid metal cooling? I ask because it is not as easy as it seems. Even experienced builders can often find themselves reapplying metal several times before it's right.
@@AB-80X it was confirmed by manufacturer that they had imperfections which caused very poor contact
i'd personally apply like half the liquid metal.
seemed like a lot to my eye through the camera.
it really only needs a 0.01mm layer, and that stuff likes to squeeze out if there's excess
have 2 of these direct die coolers, cannot wait to test them. i was gonna use them on 13900k but now im definitely gonna use the 13900ks thank you for th vid :D
Considering the weight of some fittings and tubes, does that affect the contact if the tubes are pulling to one side? Thinking especially in a vertical mount rather than the test bed laying down
I don't think it will with velocity2 it's a very good mounting system
crazy the gains...**proceeds to show next to zero numbers** what about the cinebench r23 difference? how about pushing to 5.8-9-6.0 in cb to really see gains?
I went with the EK Magnitude block as it's supposed to be the best (?) non direct die cooling solution for 13th gen Intel
Isn’t that the small one? I can’t imagine that working better than the velocity 2 but who knows
I got a standard retail kit and had the exact same problem. My screw threaded as well and the IHS didn't come off. Send it back for RMA now.
EK is expensive stuff, and the quality is definitely not always there.
Hab ich mit einem 13900k gemacht. Hat viel zu viel Strom verbraucht :) Ist jetzt mein Zweitrechner. Der neue ist R9 7950X3D und ne 4070. Da passt das mit dem Strom. Aber der Kühler von EK ist der Hammer, bringt aber nichts wenn man normale Wärmeleitpaste verwendet.
I just bought one of the upgrade kits for my 13900KS because I wanted to keep my acetal block. Was excited to use the tool as it was gonna be my first time delidding and Debauer made it look so simple. Now I'm not so sure. Think I'll sit on it to see other reviews first and make sure it's not a manufacturing error...
yep tool didnt work for me either.
The Rockitcool tool worked perfectly for me.
Use a heat gun to warm up the solder, comes off pretty easy.
@@rock962000 it was always out of stock. Finally picked up a rockitcool kit this weekend. Just gonna use that. A bit nervous now cuz now I'm hearing the direct die kit is having contact issues for some people and making temps worse. Something about the plates being convex. Man I'm so nervous about this lol maybe I'll just stick to my regular block for a while.
@@Rashimotosan That's so weird, I must have been lucky cause it was in stock when I bought it and really soon after almost everything on their website was out of stock.
James, love your voice - accent is so unusual ! im french though. This kit looks lovely and nice to see EKWB appreciating all of kitgurus help in the last 2 months with the letters etc. I like the kit, but like others, im a bit concerned ordering it due to the stripped screw - that seems like the screw needs beefed up to deal with more torque pressure. Let us know if you get any word from EKWB on the fixes
It was the thread in the slider of the delid tool that stripped. The screw is made from a high grade stainless steel. The delid slider is aluminium.
@@KitGuruJames oh ok, I must have misheard you. thanks ! You reckon if it was all steel it would be stronger?
@@larwiemcdonnel5322 yes but technically it's not necessary since the forces being applied shouldn't be enough to damage the screw threads.
Do you need warm/heat up the processor before you start deliding process ?? (when deliding 13900KS processor) or is it only EK Delid Tool needed ?? many thanks Bartek BIG fan of you !
According to EK instructions, no heat is needed. I've delidded a few 12th gen CPUs that didn't require any heat but the 3900KS did.
@@KitGuruJames What about other 13th gen cpus?
I was wondering this too. My tool should be in next week and I'm nervous about all these snapped screw/stripped threads issues... I was planning to use my heat gun eapecially for the first few passes to loosen up a bit.
Hearing it up helps a lot.
Hello, is EK-Quantum Velocity² Direct Die D-RGB - 1700 Limited Edition compatible with Lga 1851 socket?
Have to be so brave to delid one of these CPUs! Not sure if I could stomach it.
There appears to be a batch of these directdie blocks that are milled incorrectly and have a convex raised plate.
Yes I have seen the reports of poor contact. Hopefully EK will sort it out.
Yep i have one of these. Ended up reseating it 8 times before I got it half working right and even then its not contacting one set of the ecores correctly and they thermal throttle. I have reached out to EK with no response so far. Also had to ditch the foam as it couldnt be reused after reseating so many times and was getting in the way.
Will this fit the upcoming Raptor Lake Refresh dies?
Is this a in combination with a water pump in the CPU water block or with still be connected to the water pump in other words is it liquid cool or water cooling.?????
I liked the tubing & fittings that you used. Are you able to provide a link for them to be purchased?
The tubing is EK ZMT tubing. The Qick disconnect fittings are from Koolance.
Hey KitGuruTech. I have done the same thing to my KS last week. I also stripped the threads out of my EK . I think they need to use a longer screw when engaging pressure on the threads. Its just partially in there when you start applying pressure. I hope that is all it is. I did go out and get a new m6 x 1.0 screw from the hardware store that was hardened. But my top slider threads were stripped as well. I was able to retap it but it and finish but still the beginning threads are gone. It took me over 12 times. I also noticed on my two (2) 13900KS there was a lot more solder than The 13900K i've seen on other delid videos and 12 series CPUS as well. Also. i tried to use the liquid metal to take off the old residue. I left it on there for 10+ mins each time but there was too much solder on mine. I had to finally razor blade it off. took 5 sharp blades.
I applied some heat on my desoldering station. Seemed to help. Did you contact EK about your delid tool?
@@KitGuruJames I did. They first were going to send me a screw but I had to show them that my delid slider screw hole was also stripped. So they told me to RMA it. Just started that today
@@3dprintingdentist599 ah.. mine just arrived today. and it's going to be used on a 13900ks. does this look like a one off manufacturing defect or you reckon it's a fault in design? should i just reach out to them directly or attempt to use it (would it cause irreparable damage to the ks)
@@Violentmercury I bought longer screws and had to retap the top. I think the longer screws will help with the engagement of more threads when a lot of torque is applied. The screw stripping should not cause issues to the actual cpu itself.
@@3dprintingdentist599 was EK not willing to send you a new tool that works?
Does it hit 100c when you hit it with a stress test? OCCT makes my 13900KS hit 100c almost immediately with the Delid kit from EK.
Mine did too, its an issue with the kits they need to fixed. Reach out to EKWB.
@@StuartBoyer today i received the new cpu bracket but it still does it. I think i need to take out the grinder.
@@StuartBoyer So you just took the foam off and the Temps were what you would expect? i also used a generous amount of liquid metal but apparently something is wrong.
@@TinyChrisTV Original piece of foam they gave me was mis cut and had gotten in the way. removing it helped some but its still overheating. EKWB has admitted that they have some issues with the die guards and I just now got the replacement one but they still are arguing hard that the miscut coldplate isnt an issue when people doing pressure fit test show clearly it is which appears to be my case too.
@@StuartBoyer Can you tell me how exactly you got into contact? Normal Support Ticket? I did remove the foam Piece and the temps are now somehow worse... is the contact frame the issue or their coldplate? Because i already had a Satin Velocity2 and just ordered the Upgrade Kit. But i should have assembled that correctly.
does touching the CPU pins actually hurt it after a few years?
I've seen finger oil get hard on mobo batteries and ram contacts.. not sure about CPU
Never noticed a problem with touching the pins.
14:15
Definitely a durability issue! KS or not, it is the same form factor. They are not using a better glue in the KS
LTT should get that thing. Even with a 5000W chiller, it couldn't remove heat quickly enough.
LTT don't have my experience 😁👍
@@KitGuruJames I agree. but he is the master of ninja suits and other nonsense !
linus? jesus - no thanks. its tech for kiddies.
Yeah I agree LTT should get it. preferably dropped on his head from a few meters.
@@boltsandbraces6261 😁
Great video! Didn't see in it the value of the package power, could you tell us how many watts it was using during the cinebench test? Cheers!
340w package power.
Nice video, been waiting for something like that for quite a long time. All this work and money spent for only 14 degrees. Unless there's a much sheaper way to do this, it's not worth it.
Der8auer was interviewing an Intel thermal engineer who suggested that under load the CPU should run up to 90C and mange it’s boost clocks around there. If the chip under load is sitting lower then it will be leaving some performance on the table. Have you pushed this setup for additional performance?
I think you have to look at this as improving the temperature headroom and being limited by some other factor. Running at 90C isn't going to unlock performance.
@@antoniog8276 yeah agree, this is the point I didn’t make clearly. If there’s now more temperature headroom then there’s likely more performance to unlock. I was interested if they were tuning for that headroom or just leaving stock.
There is some temp issue this block thikness is have some problem 0.15mm and socket guide Frame 0.1mm is thikness so there need to manufacture it
Great video james, love how you always cover all the points and are completely transparent. its invaluable in a world today based on percentages and sales figures.
Glad you enjoyed it. 👍
Those gains are immense, but it seems like right a ball ache atm to get everything working. might be worth leaving it a while until they find tune this
It's come a long way since the first prototype that I looked at. The installation process is very easy now. However the thermal improvement could depend on how good the stock CPU to IHS contact is too.
@@KitGuruJames How does it work with a 12900k ?
@@stickyfingers9710 I tested one of the late prototypes with a 12900K and didn't see as good improvement. however the contact frame design has had a tweak since then.
doesn't look like the screw goes all the way in before it starts pulling on the IHS, so at least one of the problems is the short screw they include. I'm sure it delids fine for most people, but still that is quite the fuckup. impressive result thermally though.
I wonder if POST and memory bus issues are sometimes caused by not dressing the coolant pipes carefully enough to ensure they're not exerting any force on the block.
I don't think this would be an issue with the velocity2 block.
No.
There's kind of a rule of thumb when using contact frames of a good quality.
Finger tighten to zero lash. Then tighten with a screwdriver 90 degrees.
What TIM did you use on the aio and other block? Maybe worth seeing what it's like non de lidded with liquid metal. I know you aren't meant to but I tried it once on a 7700k and dropped like 10 degrees.
I used Arctic MX6 with the other coolers.
@James Dawson ok cool. Thank you for answering. I suppose it does make sense to test it against a more typical set up that some one would use. Great video by the way. Think im still too scared to delid a cpu though lol.
Did they fix the gap issue from the prototype w/12900k? I’ve noticed a few people are having gap issues on some groups with the 13900k. Could be user error but also could just have been a tolerance off
I heard some reports of die contact issues too. I thought this was fixed but we will have to wait and see what the outcome is. I believe EK is looking at this now.
Thank you for showing the installation prosses epic video. Cant wait to get one myself. Are you guys going to do a video going for 6ghz all p cores?.
I don't think 6GHz all core will be possible with this CPU and direct die cooling. 5.8ghz was not easy to get stable, that's why I had to back off to 5.7 for reliability.
I had the same problem with the delid tool and had to switch over to the RockItCool.
Same here :/
Please send ek a message and let them know because it seems they think this isn’t a “widespread” problem
can you share the quick disconnect fittings that u used in the video.
They are Koolance quick disconnects.
It is expected new cpus will come next year much more efficient
The block seems to slide down at 18:50 which would have smeared the liquid metal on installation, which is then followed by an abrupt cut. Hmm
The block would not be touching the die by that point.
Thats crazy, would've never realized we would come to a point of time where people are now removing IHS for custom water cooling, doesn't this void the warranty of the CPU?
Yes, and its been going on for ages - I did my first deliding over 20 years ago
That said, it isn't without its risks - many people have lost their precious cpu's doing so over the years
Hello, very cool video, but what interests me the most is how did you get the CPU so clean, you could almost think your CPU has never seen any glue :D
I've already beheaded one or the other CPU, but I always had quite a problem cleaning the CPU after beheading it because the glue still sticks well and you don't want to scratch it with great force.
Maybe someone here has a tip on how to do it in one way or another.
use liquid metal - if you leave it on long enough, it will break-down the solder and eventually remove it all clean as a whistle
@@SmalltimRYes, okay, but also the circuit board itself, you can't see any adhesive residue and you can't See scrape a scratch from the adhesive anywhere
@@jensbiernot312
Liquid metal will also remove residue as well.
That said, there should be no scrapes or scratches on the die.
The assembly process itself, is none-destructive, and the die material is much harder than the solder.
I tried direct die cooling way back in the athlon days,,,, mind you having the coolant splash directly on the silicon of the CPU not just a regular setup minus heat spreader.
Besides back then CPUs did not have a heat spreader, well at least not AMD i cant remember if Intel had that.
Heat spreaders i dont really get with today's big chips, you was much more likely to crush a tiny athlon CPU back in those days.
It worked pretty well but it did entail quite a bit of DIY / improvisation.
any tests on 12th gen Intel? is it compatible at all? I need to rewatch all the videos again
It is compatible, but i had better gains on 13th gen. However I was using one of the last prototypes for 12 th gen testing and there have been some changes to the die guard since then.
yeah 12/13th gen can be delidded with same tool
Really impressive direct die cooling. I'm ordering one right now, wondering Koolance Quick Disconnect are those, really would like to copy your setup on my test bench. Not sure but im guessing, EK-Quantum Torque Extender Rotary MF 14 --> QDT3 EPDM --> QDT3-M10X13 ?
Yes they are Koolance QDC fittings, Torque double male extender, tubing is EK ZMT.
What is the score in CineBench R23?
The kf runs hotter than ks so it would make more sense to use the kf.
I'd recommend to put the liquid metal to the q-tip rather than directly to the cpu sometimes the liquid metal can go everywhere. I'd rather it go allover the q-tip rather than your motherboard!
You left us in a such cliffhanger could you pls tell what was the cinebench score in the end :)
notice non of the tech tubers have a propper benchmark and test? cause the product is defective, lots of EK DD already killed their CPU, its a shame, popular tech tubers just repeat the BS on this product
how bad is the results if i use normal thermal grizzly past that if i don't want to use Liquid metal. like 7c differents?
Not worth it.
Feel free to heat up the cpu a bit while trying to delid
Thanks...
Could EK also make the delta2 tec with a direct die version?
I don't see why not.
Delidding could go from 2-5 times or 30 times and more. So no worries.
yeah i heard its not energy efficient
Anyone knows all the cooler parts used in this video ?
if it was also available!
is it okay to use Thermal past instead liquid metal when direct die cooling?
Liquid metal will work much better than regular TIM.
Насколько громко работает такая систем охдаждения?
If the delid tool destroys your CPU will they give you anything towards a replacement or are you totally on your own?
I would think there is a disclaimer somewhere but that would be a good question to ask EK directly before you purchase.
Dude my delider did the same exact thing! Now I am stuck with a have delided 13900k! I reached out to derbauer about trying to get a replacement because now my computer is down and ek won’t have more for another month! Please let them know about this issue so we can try and get replacements because I am so boned right now since I don’t have another 600$ for a new 13900k. Thank you soo much for showing this because I figured it was just mine that messed up but it seems the tool isn’t strong enough for the multiple passes required to remove the lid. Also I can’t even buy one from rockit cool because they are sold out as well. I don’t know what to do atm and am royally screwed. Please let them know about this problem
Try applying some heat with a hot air gun. It might soften the indium solder enough for you to remove the IHS if the IHS has moved.
@@KitGuruJames yeah I tried that as well but I can’t get a good enough contact with the tool to pull it forward to move the lid. The threads on the tool literally came out like on his video and I can’t get anything to work. :(
Man let me buy that from you!!! I wanted to do a gold build and I had to buy the clear one!!
You could order the normal Velocity² block and a direct die upgrade kit instead. There's a golden version too.
EK also sells these metal parts separately, if you want to spend the money for that. You just have to contact customer support and tell them what you want. Since you already got the transparent version, that's probably the cheapest option for you.
@@AlexanderDinebier thanks man.. I knew the upgrade kit existed but my brain didnt even think of that as an option
Somthing failed on my motherboard after putting new cpu paste on and Iooked in and I must have caught one pin on the edge of the cpu and bent it completely around waiting on a new motherboard so I can use my pc again
I have to see what's going on. I'm still thermal throttling with my EK Direct Die kit on Cinebench. I'm not getting anywhere near these temps with my 13900KS. I have tried 2 different boards as well. I have an Aorus Z790 Master and an Asus Z790 Apex. I'm running two hardware labs Nemesis GTS 360's.
I think there can be chip to chip differences but it sounds like you are getting poor block to die contact if it's thermal throttling.
What CPU frequency and vcore are you using?
I’m getting the same with my ks. Thermal throttling even with the kit, a delided k model I had with a copper ihs and the same block prior to the upgrade kit did not do this
@@KitGuruJames Contact should be good. Mounted it on 2 different boards as well. Running the cpu stock just have XMP enabled. Vcore is 1.25v under load.
@@KitGuruJames I also used a torque driver to install everything.
@@StuartBoyer are you ruining the chip stock? Did you use a torque driver? I can't believe I'm thermal throttling. I have a pretty good example of silicon as well.
What cooling are u using
Have you ever looked at the direct die kit from Supercool Computer in Thailand? Very different approach with it's own upsides and downsides but it would be really cool to hear thoughts from an expert like yourself and maybe a performance comparison as well
I've not seen that but I'm interested to take a look at it.
@@KitGuruJames some say its around 10c better then ek DD thier getting like 65c
please review it
@@KitGuruJames It's an interesting concept - instead of installing a traditional waterblock using the normal mounting holes, the cold plate itself is in the same form factor as the original IHS, so you merely replace the stock IHS with the cold plate and tension it in place with the original ILM, then you screw the rest of the block onto the cold plate. One the one hand it eliminates the worry of over tightening the block and cracking the die, but the design seems like it would put additional pressure on the ILM and you essentially can't remove the block without draining the loop first
It looks like it works great but it also cost as much if not more than the CPU it is cooling. I would like to do this kind of thing but I'm not willing to dump $500 + just for cooling.
Derbauer has direct die frames for lga1156, 2066 and am4. You replace your oem socket mount with them and reuse the cooler you already have. Since he was helping them design that i expect am5 and lga1200 versions to follow.
There are ways to direct die without spending that much money...I think I spent about $200 all in for the Block, delider, and frame...
Those are some serious gains. but im a bit concerned with that tool that broke. is that confirmed to be an issue? I trust your knowledge james and while i know you are being diplomatic, thats a bit of a concern seeing those screws strip. Is that the best material to use for those screws?
The screw is high grade stainless steel. It was the thread of the delid tool that stripped which is aluminium. According to EK there has only been one other report of the slider thread stripping like this. EK is looking into it. There were some reports of screws breaking on early versions but the screw has been upgraded since then.
It happened to my tool as well!
@@KryptGamingTv interesting, did you report it to EK?
@@KitGuruJames I just did now because of this video. I was looking for alternatives but can’t find any and rockit cool doesn’t have delid tools in stock. It sucks having a half delided 13900k and no computer :(
@@KitGuruJames Could be mine. I made a comment above about it
14:26 pretty much the story of EK nowadays.
I got a good laugh out of that one. It's really sad when your premium product breaks before even getting a single use out of it.
Seems a bit unfair - but I dont follow them all that closely , have they had other issues?
that delid tool has the wrong style bearing in it for the way the force is applyed no wonder its hard
Did this to my ks with this kit and it’s hitting 90-100 at times. Did you adjust the settings in bios? I left mine at the intel defaults
Yes I set a stable oc/fixed frequency and vcore to ensure accuracy of testing the different coolers. P-cores at 5.7, e-cores at 4.5, vcore is 1.3v in bios. Load line calibration set to auto and all intel velocity boost features disabled.
@@KitGuruJames Thankyou I'll set mine to similar to see if mine is doing right. Did you get thermal throttling if left at default where it can hit 6ghz on two cores like mine?
@@StuartBoyer no. The default BIOS produces around 310W package power on the Z790 Aorus Master which is lower than with my own config.
@@KitGuruJames Thanks. I ended up taking the foam off as it wasn't a good fit to begin with and directions have you add the foam to the sticker and add that to the card unlike in the video and the sticker wasn't cut right it seems. cleaned and reapplied the liquid metal and temps dropped some more. Needs a better piece of foam and sticker, QC has been lacking a lot with EKWB lately. I had bad seals on a 4090 ABP set and a broken LED strip out of the box and QC on this kit seems to need some work.
Good news is no thermal throttle now. Got 93 degrees on all core 5.9 with all e cores at 4.6 and got 42684 in cinebench, sure I can tune it better but that was only from changing the ratios. Definitely needed in my case to dump the foam and go slightly heavier on my liquid metal than I did with my 13900k and copper IHS.
Can this be done with normal air coolers?
Potentially yes. You would have to modify the mounts at a minimum.
Nice video
I use it 2 min work easy to use