@der8auer. Will you follow up with a new video about EK Direct Die water block? Loads of people complain about uneven cold plate and the cooling results from EKs Direct Die kit doesn't match their expectations. Thanks And thanks for a very informative and good UA-cam channel. Continue as you do :)
Haha papusan? I'll catch up to you in enthusiast someday for real. Anyway, ya I got the direct die kit & performance was worse than LM on the IHS. I ended up setting the IHS back in place & keeping it there with the thermalrite frame & putting my heatkiller block back on. It was really disappointing ngl. I've been meaning to RMA but EK wants me to send them a bunch of screenshots with HWiNFO & I haven't gotten around to it yet. Sad
@@ggood003 Hi. Very sad the reviewers don't take a second look on the direct die kit from EK. The damage is already done. Good reviews showing ok results will push people to the sales sites and all they get back is a not good working direct die kit. Some more lucky than others. Pay a lot for minimale results you could get with only delidding is sad. And reviewers should go back and test the so called fixed Direct die kits from EK then compare with the the results they got from first revision. Oh' well. Maybe the interest for the product is gone.
"Niche tool" 3 hours after video release, all of them are sold out. Which is great! I hope that it will show EK that such tools are slightly more popular that they think, and maybe we will see more products like this in the future!
I've got mine on pre-order. I have run every system direct die up to the point of LGA 1700 for a long time, and I am so glad I am finally going to be able to do that again. Bare die cooling is so much better than a CPU with an IHS. I wish Intel and AMD would ship desktop CPUs with the IHS loose so that the added cost and inconvenience of delidding would not be necessary. The end user could then decide whether or not to use the IHS. For those that are scared of the idea, keep in mind that ALL laptops are bare die, so it is not a mysterious novetly idea. This is the right way to do it. The SuperCool Computer option is messy and makes it too easy to dump water on the motherboard the way it is designed.
If anyone ever finds themselves wanting to do a lot of indium solder removal, check the material safety datasheet (MSDS) for a bunch of different liquid metal products and look for ones with the highest gallium content by percentage. Liquid metal is a near-eutectic alloy of indium, gallium, and tin. Unlike a true eutectic, liquid metal typically has an excess of gallium to make it easier to spread and flow at room temperature. RoHS compliant indium solder is not pure indium, but instead an alloy of indium and 30-70% tin or bismuth, plus some trace elements (most typically a few percent of silver). When you put the liquid metal on the indium solder, the gallium leeches into the indium solder and forms a eutectic, leaving the tin/bismuth in suspension. This is why the liquid metal turns into more of a paste when used for cleaning indium solder. The further away from a eutectic you start with (i.e. the higher the proportion of gallium you have in your liquid metal) the more indium it will pull out before the suspended insoluble compounds saturate the mixture and prevent further leaching, and the less you need to use overall. For a one-off job it's such a small variance that it probably won't matter, but if you're setting up shop selling delidding services or similar then the costs and waste can add up.
@der8auer lots of people having poor performance on the EK DD, and i think we figured it out, your 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 would like to see this stuff be available for all high-end water cooling. For instance, I would like to see EK release a 7950X version of this and just continue to offer delidded water cooling kits for high end computing parts
Thank you for all the work you put in on this. I am bummed I did not get one of the kits, they are sold out. Will they make another full kit that is not limited edition? I seen the parts are available separately but the little foams and the liquid metal not included. I have a 13900ks on the way to me and would like to use his kit. I am a bit scared to screw it up but you make it look fairly fool proof if you are careful. I really wish I would have been one of the lucky 100 to get the Limited edition kit because it is quite special and you guys did an awesome job on the styling.
Great video. Love this explanation and instruction, along with what EKWB did with their instructional 2 part vids. By the way, your test performed better than 11C. it actually did 14.6C better. I wonder how this kit will work with the new Intel Ultra 9-14900KS CPU when that comes out here in a few months. Also, that's what Intel is naming it and getting rid of the "I" and calling it "Ultra." More unnecessary words that we don't need. If you won't be using the same deluding tool, will you be using a newer one for the 14th Gen CPU? Let me know what you can. Again, great vid!
With all the QC issues EK has been having the last 2 years I'm not sure if I trust them to make anything direct-die Edit: this comment aged beautifully, EK fucked up the blocks and has to replace a ton of them lol CALLED IT
Credit is due for their service - if there is an issue with their product and their product caused damage to your system (in case there is no negligence on your side) they will compensate you.
The ilm replacement kit works with any aio like 240 or 360 mm or it s just waterblock compatible only ??? Personally i have a contact frame from Thermalright and it did improve but im looking for more without delidding the cpu for now .Looking forward for you answer .Thank you
@@wololo10 If custom loop scares you, direct die should scare you more. And the pump and rad will always bottleneck an AIO. If an AIO came with a good rad/pump, it would be too expensive.
I tried the liquid metal to remove the intel residue, but intel put A LOT of material in there. I ended up using 3 sharp razor blades and slowly scraping/cutting it off.
Well @De8aur I sure hope your people did the D part of the R&D & the manufacturing for EK because some of their products have been Trash lately! I mean from my experience the waterbkocks have still retained decent quality so you might be ok putting your name on an EK product but I sure as hell would have made sure to test it personally b4 it went to market if I was you. Also why would they not pay you anythung in royalties for using your name? That sounds sketchy or rather hard to believe but if you say so I guess I believe you. EDIT: Oh just watched & I guess you were involved in the development lol, that was a coincidence. Glad to hear it though👍
der8auer, when will your delid tool be available? I'm getting my 13900KS next week, and am definitely going direct die. I have direct-die water block coming from SuperCool, which I don't think includes the delid tool.
I've been using your direct die cooling on my 9900k. My little bro has a 13900k. I might recommend this direct die cooling for his cpu. When is it available in the US? Great job.
@Jesus is Lord Thats what i was thinking , doesnt the "other" direct die get 8 - 15c difference ? Im wondering about how they bolted it to top piece , like is the fins in the bottom part that touches die ? I missed it if he said or showed it.
11 degrees off of a very decent cooler setup. That's pretty nuts, because you know the difference is even crazier compared to your typical 240 or 360mm AIO that many people run. The results here will probably start bringing back the days of delidding
@@rustler08 I don’t know why people stopped. I didn’t have 10th gen personally or work with 11th at all. It was still very beneficial for 10th gen and even more so for 12th in my opinion, especially for those with the i9 k skus. I can pull over 450w through my 12900k with the supercool direct die block and not thermal throttle. My daily OC of 5.4p, 4.3e running r23 stays in the low 70s while looping cinebench (about 350w). Gaming I normally stay in the lower 40c range. My CPU loop has a MORA3 360 with a manual knob to control the fans. I’m sure of the exact rpm they run at, but I put them at a level where I can’t hardly hear them in a quiet room about 8ft away. With any other noise you can’t hear them at all.
I think direct die should be compared to a delidded cpu + liquid metal. If you're going through the trouble of delidding it anyway, the difference that matters is only the extra layer of IHS and top paste/LM.
@@4gbmeans4gb61 I doubt it was a 50c drop, even in gaming. Maybe 30c. In this type of workload though, the tec cooler would be overpowered and the CPU would end up thermal throttling.
If only there was a real market for that an more people actually wanted it they probably would....expecting a company to spend time and money just to cater to a insanely small niche of people is a little absurd..if that is your think you can find people willing to do that for you for a price or...get this...if you really are interested in that sort of hobby spend time learning how to do it 😂
@@atticushexcel9567 He does have a point though - this makes no sense for the manufacturer. Especially as far as product reliability in the hands of the average consumer. There’s a reason they’ll void your warranty for doing this.
@@atticushexcel9567 Well but he has a point, hundreds of people would flood forums with cracked dies shouting that intel or AMD sucks because "I broke very fragile very expensive part", heatspreader mostly exist for safety reasons.
we had that in the early AMD athlons and that was very dangerous, many cpus burned including one of mine having watercooled or repasted gpus, the large size of the cooler and the gpu die helps but on smaller cpu dies and coolers it's very easy with uneven tightening force to break the chip or have part not touching, intel cpus have pretty good overheating protections but back then you pressed power on your cpu was smoking, it would be complicated to sell such products warranty wise today
this direct die from EK and Der8auer is a total failure lots of people now are having problems please explaine I have same issue. At water temp 20 degree celcius, my 13900ks with ek dd can only costs up to 400watt at 110 degree celcius. I reinstalled 7 times, repaste liquid metal, results are same. 400watt with 110 celcius. 330 watt with 80 celcius. This result is same as heatkiller 4 pro with Stock IHS. 400 watt with 110 degree. Actually, ek dd is worse than normal cpu block. Core temp difference on heatkiller is only 7 degree difference, however ek dd is more than 15 degree. Core4, 7, and 8 extremely high. Once I changed to iceman dd, i got cured. At 330 watt it is only 60 degree. 550 watt with 110 degree. Shame on you, ek What did de8auer do? Just rent reputation and have money? Than just decrease the price without name on it. I dont think ek will recall this garbage product. People who bought this product expect good cooling performance, not the same performance with stock ihs. And also, why does ek always delay their product? Once you said it will start deliver on mid Feb, it supposed to start at least 20 Feb. It is promise with customer, especially you received money first. If you guys think you cannot gurantee the time, than receive money later. I was big fan of ek, no more. Quality keeps decreasing, price keeps increasing, and most of product need to be rma. Im tired of this.
I found my way back here after search for an alternative de-lid tool to the EK one. The EK LGA1700 direct die product is total junk. After a three month wait the product turns up. Immediately the die guard is out of spec leading to poor contact, EK will replace the die guard part but you have to wait. The direct date cold plate face has been proven to be concave and makes poor contact, EK are ignoring this issue at the moment by claiming they are flat when they are not, so even when you get your new die guard it might not fix your poor die contact problem. The EK de-lid tool strips thread, EK will replace it, but you have to wait. But the big kicker...…despite knowing they had an issue brewing, EK did nothing except "business as usual" and so now EK support are not responsive, they don't answer phones, they are so back logged their response times have more than doubled, they are swamped with support tickets requesting new die guards .etc when they do reply to your ticket, they then ask for photo's which gives them another 5 days grace before responding again, so a further two weeks pass before you know whether they will send a new part out again. Meanwhile over on Facebook they are gaslighting customers by claiming the issues affect only a very small number of users. On Reddit, Der8auer basically has walked away from it all, I paraphrase., but it goes like this......"I got a sample, I tested it, I don't know what happened after that" (strange to be involved in the development of something only to then not take an interest in the final product, especially as it carries his brand logo) What a disappointing experience it has been for those of us who have been seduced by the prospect of running cooler CPU's and what a warning about mis-placing faith in UA-cam content creators who speak positively of, or even get involved in the development of products, have their logos appear emblazoned on them, only to then to go silent/walkaway and not follow up for their fans, when the going gets tough and issues arise.
I'd love to see you overclock this 13900k with the direct die cooler to see how high the clocks can go, and what kind of performance improvement that provides!
@@J.king87 He gained around 7-12% fps, for 15-30% more power depending on game. Seems good enough for me, i personally don't care about how much power i use lol. Still that is around 155watt max on the highest game and 68c i think. He was using 1.55v so his chip is NOT very good at OC lol. Iv seen some chips doing 6.1ghz at 1.4v on liquid cooling. I just ordered the 13900KS, hoping i get a good one.
@@4gbmeans4gb61 Just got my 1300KS in, but waiting on the other parts to do my new build. Looking forward to seeing what kind of OC I can get with this direct die mount.
I remember de-lidding my 4770k and being nervous doing it in my vice. Granted it was not soldered but after putting liquid metal on it I got at least a 20c drop in temperature!
Awesome stuff as always Roman. It is nice to see a collaboration like this. Could it be possible with EK's resources to release future products sooner?
Make a test unit out of pure SILVER. It has 11% higher Alpha, the product of conductivity and heat capacity, compared to pedestrian Copper. It MUST absorb transient spikes better. Have it plated with Rhodium, or perhaps even better, Graphene to prevent oxidation! The only other solid that could possibly be better would be a DIAMOND heat spreader. A company named "Diamond-Materials" might be willing to let you have or at least test with one. At ~1600 W/m·K @ 90°C, it would destroy even Silver with only 430 W/m·K. With Diamond's superior strength, you could get away with a very thin plate to maximize heat flow. To quote The Eagles, Take it to the Limit!
I don't know if I would call EK "mainstream". Their products are prohibitively expensive for most users making it unsurprising most users just opt for AIO watercooling over open loops. I used to buy plenty of EK products back when they were the little guys but nowadays they're just too expensive.
IHS needs to go away, and we need to return to direct die cooling. The whole reason IHS was invented was because CPU heatsink mounting was badly designed, it was done with spring clips and at a sharp angle and people where chipping the corner of the CPU die during heatsink installation. Now with CPU heatsink mounting mostly all using a straight down screw-in mounting style, there’s no reason to have an IHS anymore.
The best way I have found d to clean up liquid metal is the blue shop towels and 99% isopropyl alcohol. I just soak a corner of the towel with alcohol and then clean up the lm just like normal thermal paste. I was quite amazed how well it works. Edit: I just finished the video, went to order and it's already sold out... only 100 made is quite a small amount.
I hope to see the delid tool released soon. I suspect Intel accidentally replaced solder with toothpaste on my 13600K. Installing a contact frame made no difference. Contact between cooler and IHS is good, yet water temps (AIO) are not rising very much despite high CPU temps. A delid is sorely needed.
Any chance you can compare this to the super cool direct die waterblock, or just your overall opinion? I own one, but it is deeply installed in my Mini-ITX system and has allowed me to get very good thermal performance that would otherwise not be possible in my setup. Igors lab did a review. Cheers from the US!
Interesting, but I already get sub-80C temps on an i7-12700KF 250W+ at 5.2GHz all core with no throttling over a 10-min Cinebench R23 loop. If I delidded and used LM I'd maybe save 1-2C at considerable $$$ expense, and losing all warranty and resale value when I want to upgrade (probably with 14th gen). I also get that the 12700KF has slightly lower power consumption than the 13900K but all in all not that different. I use the TR contact frame (TG was 6x more expensive, had similar performance in reviews, and was constantly out of stock) with EK 1700 block. Also use Noctua NT-H2 with a 3x360mm, single CPU (12700KF), single GPU (RTX 3090), single D5 pump.
Absolutly the same here but on my 13700K. All temps at the mid 80 degrees mark (just double checked and one core hits 90 degrees) with a 256w load and a consitent 5.4Ghz withthe Thermalright contact frame (so so cheap and works a treat!) and a 360 AIO. Scores 31222 on Cinebench R23 and I know I can get better...Though it would be great to drop temps another 10 degrees to push even harder...
You can still resell a delided CPU. I watched a 12900KS go for *far* more than I thought it should have: $377, Essentially the same price as a _brand_ _new_ 12900KF!
Curious to see how it compare with the Supercool direct die block. I was hoping better cooling perf from this block. It seems Supercool still perform better.
I think it would have been nice to see cpu power usage in these charts, just so we can see similar load between runs (even though direct die will use less power because its colder).
I'm a bit disappointed in this video, it's such a wonderful walkthrough of the process to direct die cooling, but is missing so much critical information that would help people choose if they wanted to take this step or not. What's PLL1 and PLL2 at? What was your max wattage, what was your vcore ( looks like 1.35 but it was never stated). what were the actual result differences in r20 or r23? Why didn't you do the mounting kit with the corsair block? There's so much information in this video, and so much more that's missing.
The TEC is more experimental, it require Intel software and seems to turn itself on/off based on load, it will not run subzero on a full load from what I've seen. The EK DD is also having issues. See the recent posts, I wouldn't do it if I were you. Last option waiting for der8auer to release a new DD coldplate...
Great job on the de-liding - nice block too. But one question here is, once you run this setup a while and LM has dried-up somehow,. and your block is stuck to your die, how do you separate them? Supercool solution is way better for that = you simply use the same delid tool you used to remove the IHS. But what will you do when this block is stuck to your die?
I wonder if something like that will be possible for am5 3d cache variants considering that 1 ccd would have cache on top, 1 wouldn't and there is also IO die that could have separate height as well
leave the cpu in the socket for protection...yea I shoulda done that when I tried...and failed... to delid my haswell :P tried ye olde razor...which didn't quite get to the glue and made some marks on the CPU but the CPU was fine...the dumb thing I did was take the CPU out (to try and cut the IHS off) and then place the plastic cover back on the CPU retention mechanism...but I thought it went on the inside of the mechanism instead of the outside.....derp thankfully only board I ever killed >.>
With the '14th gen' launch I'm waiting for the 14900KS after Intel kept back all the 'good' 13900K chips for the KS release. Hopefully January 2024. This tool is supposedly compatible as the die hasn't changed dimensions. Would it be possible for you to do a direct die test of your new 14900K using LM, Kryonaut and graphene sheet? I'm sure a lot of people would be interested in the results and it would potentially remove the need to re-apply liquid metal every year or two. Excuse me posting this on your 14900K video as well. Also, any chance of you offering binned delidded 14900K chips with likely KS level of SP and IMC?
I've got my i9-13900k overclocked 100mhz on all cores. That is, 5.6ghz all core, rising to 5.9ghz with 2 active cores. I use an EK AIO 360 with 3x Noctua iPPC 3000rpm. Under all-core stress test Cinebench r23, I stay around 92 degrees C, with no thermal throttling. (CPU drawing 263W, 1.15 volts.) Cinebench r23 multi-core score of about 41'100. (Stock is about 38'500, or so, according to reviews.) By the way, I picked that AIO cooler thanks to your video, since it performed so well. When looking at people running this CPU in stock configurations, I see them throttling on all-core loads, even with AIO coolers. So, I guess my setup is pretty good? My quick overclock is simply that I set SVID to "best case scenario", and offset the voltage by +0.1 Volt. This gives me stability and thermal headroom to boost everything by 100mhz without getting too hot.
I first dipped my toe into the water with my 2080ti and EK water blocks a few years ago - never again and NEVER with EK the support is non existent , they dont care about people in any way shape or form
Thanks for the shoutout. We were happy to help start this. Great work Roman as always ❤
@der8auer. Will you follow up with a new video about EK Direct Die water block? Loads of people complain about uneven cold plate and the cooling results from EKs Direct Die kit doesn't match their expectations. Thanks
And thanks for a very informative and good UA-cam channel. Continue as you do :)
Haha papusan? I'll catch up to you in enthusiast someday for real. Anyway, ya I got the direct die kit & performance was worse than LM on the IHS. I ended up setting the IHS back in place & keeping it there with the thermalrite frame & putting my heatkiller block back on. It was really disappointing ngl. I've been meaning to RMA but EK wants me to send them a bunch of screenshots with HWiNFO & I haven't gotten around to it yet. Sad
@@ggood003 Hi. Very sad the reviewers don't take a second look on the direct die kit from EK. The damage is already done. Good reviews showing ok results will push people to the sales sites and all they get back is a not good working direct die kit. Some more lucky than others. Pay a lot for minimale results you could get with only delidding is sad. And reviewers should go back and test the so called fixed Direct die kits from EK then compare with the the results they got from first revision. Oh' well. Maybe the interest for the product is gone.
"Niche tool" 3 hours after video release, all of them are sold out.
Which is great! I hope that it will show EK that such tools are slightly more popular that they think, and maybe we will see more products like this in the future!
Yeah, it really pisses me off that they create some thing people want and then make 100 for the entire fucking planet. Stupid fucking people.
@@guristoski2625 They also have pre-order for nickel version of it. Will be available mid-february.
It's also 150euro cheaper
Apparently it isn't niche enough!!! 🥰🥰🥰🥰🥰
Maybe they didn't make that many expecting low demand so they were easily sold out
He said they only made 100 if that version at the end
I've got mine on pre-order. I have run every system direct die up to the point of LGA 1700 for a long time, and I am so glad I am finally going to be able to do that again. Bare die cooling is so much better than a CPU with an IHS. I wish Intel and AMD would ship desktop CPUs with the IHS loose so that the added cost and inconvenience of delidding would not be necessary. The end user could then decide whether or not to use the IHS. For those that are scared of the idea, keep in mind that ALL laptops are bare die, so it is not a mysterious novetly idea. This is the right way to do it. The SuperCool Computer option is messy and makes it too easy to dump water on the motherboard the way it is designed.
I'm really excited to see a day when CPU manufacturers make the water block apart of the CPU it self.
If anyone ever finds themselves wanting to do a lot of indium solder removal, check the material safety datasheet (MSDS) for a bunch of different liquid metal products and look for ones with the highest gallium content by percentage.
Liquid metal is a near-eutectic alloy of indium, gallium, and tin. Unlike a true eutectic, liquid metal typically has an excess of gallium to make it easier to spread and flow at room temperature. RoHS compliant indium solder is not pure indium, but instead an alloy of indium and 30-70% tin or bismuth, plus some trace elements (most typically a few percent of silver). When you put the liquid metal on the indium solder, the gallium leeches into the indium solder and forms a eutectic, leaving the tin/bismuth in suspension. This is why the liquid metal turns into more of a paste when used for cleaning indium solder.
The further away from a eutectic you start with (i.e. the higher the proportion of gallium you have in your liquid metal) the more indium it will pull out before the suspended insoluble compounds saturate the mixture and prevent further leaching, and the less you need to use overall.
For a one-off job it's such a small variance that it probably won't matter, but if you're setting up shop selling delidding services or similar then the costs and waste can add up.
@der8auer lots of people having poor performance on the EK DD, and i think we figured it out, your 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
Thank you Roman and EK! Awesome job!
I've been really impressed with James at KitGuru, he's ultra professional and his builds usually make me really envious.
Love the transparency bit. Like for it.
Well done and thanks.
I love EK honestly, their products are my go to for all my water cooling needs.
You used PCB of another CPU to remove the glue on other CPU, wasn't expecting that 😆 Btw, good work as always 😁
Not just that, but a dead 12900K. Bro just used the corpse of its older brother to clean off the glue.
I would like to see this stuff be available for all high-end water cooling. For instance, I would like to see EK release a 7950X version of this and just continue to offer delidded water cooling kits for high end computing parts
Great news for all your hard work. Congrats!!!
Dude if I didn't have a baby on the way, I'd be purchasing a kit right now. LOVE IT!!!
Great video, as always!!
I love the Vanilla Ice (rapper) style haircut! It also looks like a giant heatsink! No wonder you're such a cool UA-camr! It's in the hair! 🤣
Using a delidded CPU to remove the glue on another delidded CPU is something only der8auer would do.
8:28 Nice scraper 😆
Thank you for all the work you put in on this. I am bummed I did not get one of the kits, they are sold out. Will they make another full kit that is not limited edition? I seen the parts are available separately but the little foams and the liquid metal not included. I have a 13900ks on the way to me and would like to use his kit. I am a bit scared to screw it up but you make it look fairly fool proof if you are careful. I really wish I would have been one of the lucky 100 to get the Limited edition kit because it is quite special and you guys did an awesome job on the styling.
Liquid metal and foam does come in the upgrade kit
Congrats, I still remember when you had a patent issue with the ek but I guess you have a reminder on your wall behind you :)
its always a good idea to tape the area around the cpu to not spill liquid metal on the other components.
Der8auer, the direct die is great, but we need direct die monoblocks.
Oh that voice crack at 3:48 lol
Gigachad using a 12900k to scrape glue off a 13900k (I know it's a dead chip but what a flex lol)
I bet mine would be great if it actually made contact with the die
Fingers crossed for a 7000X3D version!
Love it!!! Sold!
80 degrees with exotic cooling. holy how..
My man just cleaned his 13900k with his 12900k 😂
Great video. Love this explanation and instruction, along with what EKWB did with their instructional 2 part vids. By the way, your test performed better than 11C. it actually did 14.6C better. I wonder how this kit will work with the new Intel Ultra 9-14900KS CPU when that comes out here in a few months. Also, that's what Intel is naming it and getting rid of the "I" and calling it "Ultra." More unnecessary words that we don't need. If you won't be using the same deluding tool, will you be using a newer one for the 14th Gen CPU? Let me know what you can. Again, great vid!
Great news! But you didn't mention if existing AM5 velocity 2 can also do direct die cooling.
With all the QC issues EK has been having the last 2 years I'm not sure if I trust them to make anything direct-die
Edit: this comment aged beautifully, EK fucked up the blocks and has to replace a ton of them lol
CALLED IT
Credit is due for their service - if there is an issue with their product and their product caused damage to your system (in case there is no negligence on your side) they will compensate you.
@@eliadbu they've tried blaming manufacturer defects on user error before. Lots of poeple in watercooling forums have had issues
would love to see a delta tec version of this direct die block
only thing is when the 14th gen comes out id regret doing this while the 14900k runs almost as cool and faster with a standard 240mm AIO.
I've never seen someone cleaning a cpu with another cpu
Amazing.
Also, is there a reason why you're using R20 instead of the current R23 for Cinebench?
The ilm replacement kit works with any aio like 240 or 360 mm or it s just waterblock compatible only ??? Personally i have a contact frame from Thermalright and it did improve but im looking for more without delidding the cpu for now .Looking forward for you answer .Thank you
Congrats bro, please convince EK to do an AIO with direct die
Ew.
@@rustler08 whats the problem? 10 free degrees to non maniac builders.
@@wololo10 If custom loop scares you, direct die should scare you more. And the pump and rad will always bottleneck an AIO. If an AIO came with a good rad/pump, it would be too expensive.
I tried the liquid metal to remove the intel residue, but intel put A LOT of material in there. I ended up using 3 sharp razor blades and slowly scraping/cutting it off.
Where can I buy the direct die frame for 13th gen? I cant see it on the thermal grizzly website
11:20 jumpscare
Well @De8aur I sure hope your people did the D part of the R&D & the manufacturing for EK because some of their products have been Trash lately!
I mean from my experience the waterbkocks have still retained decent quality so you might be ok putting your name on an EK product but I sure as hell would have made sure to test it personally b4 it went to market if I was you.
Also why would they not pay you anythung in royalties for using your name? That sounds sketchy or rather hard to believe but if you say so I guess I believe you.
EDIT: Oh just watched & I guess you were involved in the development lol, that was a coincidence. Glad to hear it though👍
Since Ryzen 7000 shoots straight to 95 degrees C, it would be interesting to see direct die cooling on those chips as well.
i would like so much to test direct die with cry cooler .
3:49 :D:D:D:D voicecrackkkk
der8auer, when will your delid tool be available? I'm getting my 13900KS next week, and am definitely going direct die. I have direct-die water block coming from SuperCool, which I don't think includes the delid tool.
I've been using your direct die cooling on my 9900k. My little bro has a 13900k. I might recommend this direct die cooling for his cpu. When is it available in the US? Great job.
EK has worldwide shipping so unless they're already sold out, check their site.
@@jorkusmalorkus Fine! I'll make my own! With blackjack! And hookers!
In fact, forget about the waterblock!
Ah, screw the whole thing...
@Jesus is Lord Nobody that does this does it for practical reasons. Just run stock and you'll be fine.
@Jesus is Lord Thats what i was thinking , doesnt the "other" direct die get 8 - 15c difference ? Im wondering about how they bolted it to top piece , like is the fins in the bottom part that touches die ? I missed it if he said or showed it.
9900K gang
11 degrees off is nothing to scoff at. Big improvement. Long live der8aur!
11 degrees off of a very decent cooler setup. That's pretty nuts, because you know the difference is even crazier compared to your typical 240 or 360mm AIO that many people run.
The results here will probably start bringing back the days of delidding
@@rustler08 I don’t know why people stopped. I didn’t have 10th gen personally or work with 11th at all. It was still very beneficial for 10th gen and even more so for 12th in my opinion, especially for those with the i9 k skus. I can pull over 450w through my 12900k with the supercool direct die block and not thermal throttle. My daily OC of 5.4p, 4.3e running r23 stays in the low 70s while looping cinebench (about 350w). Gaming I normally stay in the lower 40c range. My CPU loop has a MORA3 360 with a manual knob to control the fans. I’m sure of the exact rpm they run at, but I put them at a level where I can’t hardly hear them in a quiet room about 8ft away. With any other noise you can’t hear them at all.
I think direct die should be compared to a delidded cpu + liquid metal. If you're going through the trouble of delidding it anyway, the difference that matters is only the extra layer of IHS and top paste/LM.
I thought it would be a lot more. The tech cooler was
Like 50c drop.
@@4gbmeans4gb61 I doubt it was a 50c drop, even in gaming. Maybe 30c. In this type of workload though, the tec cooler would be overpowered and the CPU would end up thermal throttling.
don't be scared to answer the public about problems with ek products. it is becoming more damaging not saying anything.
Looking forward to trying this out with the new 13900KS @ 6Ghz. Will look killer on my K|NGP|N motherboard along with more gold accents in the build.
If only intel sold top of the line versions Delidded all this would be much easier for "average" people
If only there was a real market for that an more people actually wanted it they probably would....expecting a company to spend time and money just to cater to a insanely small niche of people is a little absurd..if that is your think you can find people willing to do that for you for a price or...get this...if you really are interested in that sort of hobby spend time learning how to do it 😂
@@samgoff5289 who is buying these performance cpus? Only enthusiasts.
@@atticushexcel9567
He does have a point though - this makes no sense for the manufacturer. Especially as far as product reliability in the hands of the average consumer. There’s a reason they’ll void your warranty for doing this.
@@atticushexcel9567 Well but he has a point, hundreds of people would flood forums with cracked dies shouting that intel or AMD sucks because "I broke very fragile very expensive part", heatspreader mostly exist for safety reasons.
we had that in the early AMD athlons and that was very dangerous, many cpus burned including one of mine having watercooled or repasted gpus, the large size of the cooler and the gpu die helps but on smaller cpu dies and coolers it's very easy with uneven tightening force to break the chip or have part not touching, intel cpus have pretty good overheating protections but back then you pressed power on your cpu was smoking, it would be complicated to sell such products warranty wise today
That's so cool. Would be nice to have it for AM5 also.
These things help more with intel because they get stuck on a socket every two iterations. AM5 will just offer you an upgrade.
@@quintrapnell3605 what does this possibly have to do with socket longetivity?
I'd buy one
AMD is dead. lol
@@bornagain2641 in your dreams
Great work as always, huge improvement! 10-15 degrees is huge
Now we just need a direct die kit for AM5 so we can see how well X3D can run.
true!
"WE" ? I don't need it. Fine with my 13600K as it is.
run what? Games? lol
@@DrNoBrazil Didn't knew the world revolved around you
This! Would love to go direct-die on my AM5 CPU.
this direct die from EK and Der8auer is a total failure lots of people now are having problems please explaine
I have same issue.
At water temp 20 degree celcius, my 13900ks with ek dd can only costs up to 400watt at 110 degree celcius.
I reinstalled 7 times, repaste liquid metal, results are same. 400watt with 110 celcius. 330 watt with 80 celcius.
This result is same as heatkiller 4 pro with Stock IHS. 400 watt with 110 degree.
Actually, ek dd is worse than normal cpu block. Core temp difference on heatkiller is only 7 degree difference, however ek dd is more than 15 degree. Core4, 7, and 8 extremely high.
Once I changed to iceman dd, i got cured. At 330 watt it is only 60 degree.
550 watt with 110 degree.
Shame on you, ek
What did de8auer do?
Just rent reputation and have money?
Than just decrease the price without name on it.
I dont think ek will recall this garbage product.
People who bought this product expect good cooling performance, not the same performance with stock ihs.
And also, why does ek always delay their product? Once you said it will start deliver on mid Feb, it supposed to start at least 20 Feb. It is promise with customer, especially you received money first.
If you guys think you cannot gurantee the time, than receive money later.
I was big fan of ek, no more.
Quality keeps decreasing, price keeps increasing, and most of product need to be rma. Im tired of this.
I found my way back here after search for an alternative de-lid tool to the EK one.
The EK LGA1700 direct die product is total junk. After a three month wait the product turns up. Immediately the die guard is out of spec leading to poor contact, EK will replace the die guard part but you have to wait. The direct date cold plate face has been proven to be concave and makes poor contact, EK are ignoring this issue at the moment by claiming they are flat when they are not, so even when you get your new die guard it might not fix your poor die contact problem.
The EK de-lid tool strips thread, EK will replace it, but you have to wait.
But the big kicker...…despite knowing they had an issue brewing, EK did nothing except "business as usual" and so now EK support are not responsive, they don't answer phones, they are so back logged their response times have more than doubled, they are swamped with support tickets requesting new die guards .etc when they do reply to your ticket, they then ask for photo's which gives them another 5 days grace before responding again, so a further two weeks pass before you know whether they will send a new part out again. Meanwhile over on Facebook they are gaslighting customers by claiming the issues affect only a very small number of users.
On Reddit, Der8auer basically has walked away from it all, I paraphrase., but it goes like this......"I got a sample, I tested it, I don't know what happened after that" (strange to be involved in the development of something only to then not take an interest in the final product, especially as it carries his brand logo)
What a disappointing experience it has been for those of us who have been seduced by the prospect of running cooler CPU's and what a warning about mis-placing faith in UA-cam content creators who speak positively of, or even get involved in the development of products, have their logos appear emblazoned on them, only to then to go silent/walkaway and not follow up for their fans, when the going gets tough and issues arise.
I'd love to see you overclock this 13900k with the direct die cooler to see how high the clocks can go, and what kind of performance improvement that provides!
He got 6.2 all core with that tech cooler. And stayed below 65c, I really want one of those.
@@4gbmeans4gb61 how much fps more then if you compare 5.7ghz vs 6.2 ghz
@@J.king87 He gained around 7-12% fps, for 15-30% more power depending on game. Seems good enough for me, i personally don't care about how much power i use lol. Still that is around 155watt max on the highest game and 68c i think. He was using 1.55v so his chip is NOT very good at OC lol. Iv seen some chips doing 6.1ghz at 1.4v on liquid cooling. I just ordered the 13900KS, hoping i get a good one.
@@4gbmeans4gb61 Just got my 1300KS in, but waiting on the other parts to do my new build. Looking forward to seeing what kind of OC I can get with this direct die mount.
@@4gbmeans4gb61 do you know the cinebench score aswell when overclocked ?
I remember de-lidding my 4770k and being nervous doing it in my vice. Granted it was not soldered but after putting liquid metal on it I got at least a 20c drop in temperature!
I did a 3770k and ruined it lol, I tired a clamp with 2 2x4s rofl, it tore part the pcb. 4790k I bought a delis kit haha.
8:28 *Removing glue from CPU with another CPU*
Yeah that's definitely derBauer
Awesome stuff as always Roman. It is nice to see a collaboration like this. Could it be possible with EK's resources to release future products sooner?
Does anyone know yet if existing Ryzen 7000 delidding and direct-die cooling tools will be compatible with Ryzen 7000X3D SKUs?
No
It should be the same
Still 7950X tops R20 chart in this video)
Give us a 7950X block PLEASE!!! Take my money
Make a test unit out of pure SILVER. It has 11% higher Alpha, the product of conductivity and heat capacity, compared to pedestrian Copper. It MUST absorb transient spikes better. Have it plated with Rhodium, or perhaps even better, Graphene to prevent oxidation!
The only other solid that could possibly be better would be a DIAMOND heat spreader. A company named "Diamond-Materials" might be willing to let you have or at least test with one. At ~1600 W/m·K @ 90°C, it would destroy even Silver with only 430 W/m·K. With Diamond's superior strength, you could get away with a very thin plate to maximize heat flow.
To quote The Eagles, Take it to the Limit!
I don't know if I would call EK "mainstream". Their products are prohibitively expensive for most users making it unsurprising most users just opt for AIO watercooling over open loops. I used to buy plenty of EK products back when they were the little guys but nowadays they're just too expensive.
I want a Direct-die cooler for AMD Threadripper though mainly Ryzen AM5, hopefully some company makes one. Does anyone know if they already make on?
IHS needs to go away, and we need to return to direct die cooling. The whole reason IHS was invented was because CPU heatsink mounting was badly designed, it was done with spring clips and at a sharp angle and people where chipping the corner of the CPU die during heatsink installation. Now with CPU heatsink mounting mostly all using a straight down screw-in mounting style, there’s no reason to have an IHS anymore.
I see that 94ºc there...😅
8:28 I love Roman's glue removal scraper choice = another CPU haha!
Shouldn't direct die be compared to a proper delidded cpu with liquid metal? It seems unfair to compare with stock IHS+goo
Time to pre order one for the 13900KS!
The best way I have found d to clean up liquid metal is the blue shop towels and 99% isopropyl alcohol. I just soak a corner of the towel with alcohol and then clean up the lm just like normal thermal paste. I was quite amazed how well it works.
Edit: I just finished the video, went to order and it's already sold out... only 100 made is quite a small amount.
That's for the gold one. I think the regular clear one should not be limited. At least I hope not. I don't know if that one is even released yet.
Don't wipe the liquid metal off in the direction of SMD's he says, while doing exactly that
I've done a delid already, so having the direct die and contact frame will be awesome.
Love kitguru. Leo is my favorite ❤️Nice collaboration with ek. You deserve it. Keep up the great work bro and stay safe.
I hope to see the delid tool released soon. I suspect Intel accidentally replaced solder with toothpaste on my 13600K. Installing a contact frame made no difference. Contact between cooler and IHS is good, yet water temps (AIO) are not rising very much despite high CPU temps. A delid is sorely needed.
Does it at least smell nice?
Direct-Die Water Block hits 85c :D sounds like 13th gen to me
Any chance you can compare this to the super cool direct die waterblock, or just your overall opinion?
I own one, but it is deeply installed in my Mini-ITX system and has allowed me to get very good thermal performance that would otherwise not be possible in my setup. Igors lab did a review. Cheers from the US!
How’s the am5 delider coming along? 🤔🤔🤔
I'm disappointed to hear that you don't get any royalty on this or the Lian Li items.
11c not very good considering a relid will easily get that.
Super cool a company like EK got involved with this type of project! Awesome Video
Interesting, but I already get sub-80C temps on an i7-12700KF 250W+ at 5.2GHz all core with no throttling over a 10-min Cinebench R23 loop. If I delidded and used LM I'd maybe save 1-2C at considerable $$$ expense, and losing all warranty and resale value when I want to upgrade (probably with 14th gen). I also get that the 12700KF has slightly lower power consumption than the 13900K but all in all not that different.
I use the TR contact frame (TG was 6x more expensive, had similar performance in reviews, and was constantly out of stock) with EK 1700 block. Also use Noctua NT-H2 with a 3x360mm, single CPU (12700KF), single GPU (RTX 3090), single D5 pump.
Absolutly the same here but on my 13700K. All temps at the mid 80 degrees mark (just double checked and one core hits 90 degrees) with a 256w load and a consitent 5.4Ghz withthe Thermalright contact frame (so so cheap and works a treat!) and a 360 AIO. Scores 31222 on Cinebench R23 and I know I can get better...Though it would be great to drop temps another 10 degrees to push even harder...
You can still resell a delided CPU. I watched a 12900KS go for *far* more than I thought it should have: $377, Essentially the same price as a _brand_ _new_ 12900KF!
People pay so that way the risk of damaging isn't on them.
Are you planning to release a delid tool for Ryzen 7000?
Please answer :)
So how much torque should you apply? Feels like one of the biggest steps.?
I want to know this as well.
Curious to see how it compare with the Supercool direct die block. I was hoping better cooling perf from this block. It seems Supercool still perform better.
I think during your direct die cooling test, the CPU was run with a higher voltage of 1.35V compared to the 1.27V from the non-delided result. 11:55
I think it would have been nice to see cpu power usage in these charts, just so we can see similar load between runs (even though direct die will use less power because its colder).
I'm a bit disappointed in this video, it's such a wonderful walkthrough of the process to direct die cooling, but is missing so much critical information that would help people choose if they wanted to take this step or not. What's PLL1 and PLL2 at? What was your max wattage, what was your vcore ( looks like 1.35 but it was never stated). what were the actual result differences in r20 or r23? Why didn't you do the mounting kit with the corsair block? There's so much information in this video, and so much more that's missing.
which one is better? The intel tec2 cooler from ek or direct die cpu cooler from ek. Which one is better for heaving gaming and consistency?
@der8auer
The TEC is more experimental, it require Intel software and seems to turn itself on/off based on load, it will not run subzero on a full load from what I've seen.
The EK DD is also having issues. See the recent posts, I wouldn't do it if I were you.
Last option waiting for der8auer to release a new DD coldplate...
Great job on the de-liding - nice block too. But one question here is, once you run this setup a while and LM has dried-up somehow,. and your block is stuck to your die, how do you separate them? Supercool solution is way better for that = you simply use the same delid tool you used to remove the IHS. But what will you do when this block is stuck to your die?
LM wont dry ever
The limited gold edition is sold out. Great job. I'd love to see an am5 version.
I wonder if something like that will be possible for am5 3d cache variants considering that 1 ccd would have cache on top, 1 wouldn't and there is also IO die that could have separate height as well
Tell them to make this for AM5!
leave the cpu in the socket for protection...yea I shoulda done that when I tried...and failed... to delid my haswell :P tried ye olde razor...which didn't quite get to the glue and made some marks on the CPU but the CPU was fine...the dumb thing I did was take the CPU out (to try and cut the IHS off) and then place the plastic cover back on the CPU retention mechanism...but I thought it went on the inside of the mechanism instead of the outside.....derp thankfully only board I ever killed >.>
What’s the release date on this tool?
With the '14th gen' launch I'm waiting for the 14900KS after Intel kept back all the 'good' 13900K chips for the KS release. Hopefully January 2024. This tool is supposedly compatible as the die hasn't changed dimensions. Would it be possible for you to do a direct die test of your new 14900K using LM, Kryonaut and graphene sheet? I'm sure a lot of people would be interested in the results and it would potentially remove the need to re-apply liquid metal every year or two. Excuse me posting this on your 14900K video as well. Also, any chance of you offering binned delidded 14900K chips with likely KS level of SP and IMC?
I've got my i9-13900k overclocked 100mhz on all cores. That is, 5.6ghz all core, rising to 5.9ghz with 2 active cores.
I use an EK AIO 360 with 3x Noctua iPPC 3000rpm.
Under all-core stress test Cinebench r23, I stay around 92 degrees C, with no thermal throttling. (CPU drawing 263W, 1.15 volts.)
Cinebench r23 multi-core score of about 41'100. (Stock is about 38'500, or so, according to reviews.)
By the way, I picked that AIO cooler thanks to your video, since it performed so well.
When looking at people running this CPU in stock configurations, I see them throttling on all-core loads, even with AIO coolers. So, I guess my setup is pretty good?
My quick overclock is simply that I set SVID to "best case scenario", and offset the voltage by +0.1 Volt.
This gives me stability and thermal headroom to boost everything by 100mhz without getting too hot.
I first dipped my toe into the water with my 2080ti and EK water blocks a few years ago - never again and NEVER with EK the support is non existent , they dont care about people in any way shape or form