Honestly I think 90 euro is very fair. I was figuring you would say double that! Considering the work you put into bringing this to market that is more than fair!
@@der8auer-en A clean looking box looks better than edgy alien designs. Just keep doing your stuff, as it looks great even for poor people like me who couldn't ever buy it but at last can watch it.
No chance I'll be buying a 285K and delidding it, but 90 euro for a small volume production run heater product seems pretty good value to me (given the amount of engineering required to design and implement it).
Unless you're doing direct die I don't know why you would do it, replacing the TIM with conductonaut is a bad idea because conductonaut dries out in a year or so, turning into dry black flakes.
@celeriumlerium8266 conductonaut I haven't heard of gallium drying out like that unless it's formed a metal oxide by reacting to copper or eaten away at aluminum? My 4790K de-lidded with conductonaut on the die has been fine for 5 years
@@kekoraaaa He has a video on it, "I missed the obvious problem with my Threadripper". It converts into gallium oxide quickly due to unknown conditions. I put conductonaut on my 7700k and had the same problem, between die and IHS (no aluminium) it turned into a dry black "dust" after about 1 and a half years.
I recently delidded my 7700k with the Die Mate 2 and I was similarly shocked when I saw the temps. Stock with an NH-D15 it was around 65-70c during a run of R23, but afterwards it topped out at like 50-55c. Absolutely crazy. Thank you for developing such great tools for the community =)
It would be great if you could get Intel to disclose the temperature at which the die can deteriorate or get damaged. This is crucial information considering people using other methods might just cross the threshold and damage or degrade the processor.
Silicium is a strong material. The die itself will stay intact way before some of the tin balls on the PCB and the pads under the ceramic caps start to melt and make the chip unusable. Preheating plates for tin soldering can go above 250°C and a reflow process or a hot air station can peak at more than 350 so there is some comfortable margins with this indium solder with such a low fusion point under 200. Since those YTers started delidding soldered CPUs, I've always wonder why they didn't use such inexpensive tools.
I would be more worried about the SMD components on the CPU pcb board. If their solder melts, some may fall off, especially if there is violent shaking, like the one when you break the regular glue. Inert die should withstand solder melting temperatures used during manufacture.
Diffusion does exponentially increase with temperature, but it's such a short time that it would still be meaningless. Also, the CPU has to be soldered at the factory (with much higher temperatures), so I guess it's fine to heat it up to sub-200 degrees temperatures (while CPU is off ofc).
Because this is a new channel he made exclusively for English viewers. His original channel where he used to mix German and English content, has double the sub count.
this channel is new, roman used to upload a english version and then a german version to a channel, but decided to split the content and now has this channel and the original one
Mom: Son what are you doing with the Air Fryer? Son: I'm delidding a 285k, what does it look like? Mom: okay honey, just make sure to clean it out after you use it. 🤣🤣🤣
I can feel that. When I was 17 years old and did Dry Ice OC at home I had to dry my mainboard afterwards. I put it in the oven at 80°C. My mom was like "what the hell are you doing here"
I mean Caseking could sell pre-delidded CPU's but I guess it's not worth the RMA hassle. Normal users don't delid so you can pretty much be sure the delidded CPU's will be hammered with OC.
Probably too much risk. Selling proven delidders is much saver than providing a delidding service where besides killing the CPU during the delidding process there also have risks associated with a CPU which is not well protected and (partially) fool-proved with an IHS. Things like transport accidents, user error during installation and warranties (most countries including Germany have warranties by law that cannot be signed away) can become an financial burden for a company
I would have to give 2 years warranty according to German law if I offer a delidded CPU for sale. And there is no way I would do that with a product that I have no control over
You could analyze the die's convexity (with the water block mounted) with a high-powered CT Scan, but that would kill the CPU at that ridiculous level of radiation. You really are brilliant, Roman, and you deserve more subscriptions.
everything is much simpler, you can make a strip of plasticine around the chip, after pressing on the plasticine there will be an imprint. After that, all that remains is to measure the height of the plasticine around the perimeter of the chip. If the chip bends, then the plasticine in the center will be equal in height to the plasticine in the corners
9:52 you can still see the IHS right? I would loosen the bolts, apply a suction cup to the IHS and lift it simultaniously. That way its a single-step process. Can maybe even create some vacuum seal in another way, like having the top of the IHS delidder have a rubber seal, with a nipple on top, and than just attach one of those handbilges to create a vacuum seal to lift it together with the delidder top. EDIT: ah its already finalized, nevermind :)
It's always so interesting to hear about how this works and what people are doing with extreme stuff like this. It's super scary, but I love watching you all do it! lol
This was an incredibly entertaining and informative video. Insane decrees in temps there. Hopefully this will be their format for the coming years of this platform and we see some huge gains.
Thank you der8auer for your consistent time and effort into making these products. Your level of transparency and honestly are unparallelled and you're one of my favourite PC TH-camrs to watch!
So much work for this platform. Too bad the CPUs probably arent worth the trouble. But all of this will likely be useful in the future, too, so at least the knowledge will be helpful. I'm waiting to see how deliddable the 9800X3D will be. Would love to run it direct die. :)
@@jankees4037 a better CPU to water delta means I can run hotter water in my loop, increasing thermal efficiency on the radiator for less fan speed (noise). That's what delidding does.
So much respect for you to go through all the effort that you do. You personally test and tweak the settings of the CPU in order to test and verify everything to the best of your capabilities. On top of that, you go out of your way to separately sell the products in order to not force anyone into purchasing something that they may not want or need. You genuienly want the best product and best user experience for the buyer and it is conveyed to the audience so big props man! ❤️💯🔥👍
Quite amusing seeing the evolution of your delid tools. I've still got my delrin plastic one for the ivy bridge era chips. Things were a bit more simple back then lol
Looking at the bear die I kinda chuckled at Intel’s comment from a few years ago: “we don’t sell glued CPUs”. Looks like someone finally learned how to glued CPUs and sell them🤣
I not only have no plan to upgrade to the Arrow Lake CPUs, but wouldn't take the chance to de-lid one if I did. Still watched the whole video. Great work Roman 👍
Their transport would be a mess. If products that are designed to take a beating sometimes already come damaged, imagine the courier yeeting a CPU without heatshield
Too much risk for the reputation. Idiots end up buying it, crack the die, try to RMA under false pretense, manufacturer has to prove it wasn't DOA with a hairline fracture. All of that just to satisfy probably 10-50k people worldwide. You'd have people crying about 'why isn't my i3 released as bare-die and only the top of the line chip gets it' or complaining that 'I'm paying more for a binned CPU with NO metal? This doesn't make sense'
make it a Der8auer special edition that comes with a water cooling kit. pack it with IHS but unsoldered straight from intel or amd. limited series of course depending on water blocks available at the moment.
Why not design it so from the start you put cpu upside down? Then you would just cut open a witness window on the side to able to see if IHS has moved or not. Avoiding the heating of the IHS the second time and spilling the indium on the PCB from the first heating. Originally in factory when they put IHS and substrate together the IHS is facing down for that reason. Also scraping off the indium leftovers on tile design - could this part be as dangerous as delidding itself? Maybe it's safer to scrape diagonally to avoid catching the wedge between the tile's border?
It's tricky because the heater obviously has to be thermally bridged to the IHS, so the heater would have to be able to hold onto the IHS stronger than the force pulling the IHS off the CPU. Probably simpler and easier just to do it this way, if it doesn't cause any issues for the CPU to get this hot.
15:30 This is getting super fancy! If that's the route you're taking, Roman, why not create a gearbox to turn one screw in a single direction and make the delidder move back and forth when it reaches a stopper on each direction (similar to a differential in a car)? This way, you can keep the screwdriver in place all the time, turning it in only one direction while you watch the delidder and its gears do all the work.
Are we machining the gears? Also.. do we have the Swiss dwarves on staff to assemble the tiny gears, as they do for the watches? Also.. could we compromise on a custom Allen wrench that allows a full rotation based on the height of the heater so you don't have to do that thing, that I was seeing in the video, that we've all been through with an Allen wrench lol.
Роман, крышку лучше сдвигать не вдоль кристалла, а поперёк, как в скальпаторе для 1700 2:27. Для этого предварительно отпаять два мешающих резистора 3:30, а после снятия крышки вернуть их на место.
A friend and I used an oven in the chemistry lab to ease delidding of his 7950x3d. We got the CPU to 100°C (we were afraid of going any higher) and then popped it into the Delid Die Mate. It seemed to work well, but without a comparison CPU, we could not actually determine whether it made any difference.
use floss to cut glue, then position the cpu upside down between two cans, apply heat from underneath with hair dryer until IHS falls off... simple and 100% safe AM5 delid
You always find a way to delid any CPU even if the manufacturer has warned you. Greatness...Next on der8auer: deliding the roof of my house to improve cooling and overclocking my heating/cooling system 😆
Do you think it would be possible to cut fins directly into the ihs instead of delidding? Maybe we can finally make use of that thick heatspreader on AM5. Oh also will the indium solder stick to the surface mount components? I would rather not have to clean up tiny tiny solder bridges haha
Curious if the heating element can be placed onto the IHS itself so when you move the delidder to break the glue, take the PCB with the CPU and drop the IHS that quick. There would also be less heat getting to the PCB and related parts. Might need a different way of tightening the delidder to make this process a single-shot but is worth thinking about.
Quick comment for final removal, instead of second heating step, how about upside down process and upon click, some kind of suction applied, e.g. syringe with suction cup on the end.
The heater controller could have a progress bar, ETA and a piezo for acoustic indicator that its ready - like start clicking near target and beeping on target
What about "renting out" the delider with a kaution of the full price of the part + 2 way shipping, and then ship it back and refund the kaution minus the rental fee (maybe of 30-40 euro per week? maxing out at the full price if someone exceeds the kaution value).
guess you're a big StarGate movie fan .... your constellation marking tablet tattoos is a great giveaway ... excellent video thank you for all the knowledge your share... cheers 😎
it is a lot of fun for someone with extreme overclocking needs, and i hope you sell a-lot of kits as a previous buyer of your delidders, but i have to be honest and say that a max ~5 or 6% in best case gain in multicore for all the expense and work/risk should carefully be considered by everyone
I did something similar with the 12900K. Ran Prime95 for 30 minutes. Pulled the CPU and delid. Took one push to come part. The problem for Arrow Lake is the SMDs clearance of course. Still doable at home.
Can you do a video on proper overclocking of it please? (Bios settings in detail). Would be cool to see how to push the GSkill CUDIMMs to 10k as well (voltages, cas, etc.). Looks great! Can't wait to get it.
Shouldn't. Will just copy my reply from a different comment: "No that's fine. You have to keep in mind they also have to solder the CPU to make it. And according to the indum to silicon soldering patents, they ususally run this at about 180°C."
On the first heat cycle. Would it be possible to add a tool that pull the lid (magnet, suction) off and keep it lifter until cooled down? It would fit through the opening in the sled. Just a thought...
10:08 why not just cut a hole in the delidder so you can flip the whole thing over and the IHS falls away from the CPU while everything else stays in place?
I'm curious, will copper desoldering braid suck up liquid indium? Scraping it off doesn't seem that difficult, but being able to remove the bulk of it with copper braid might be handy as well if it actually works. I just don't have any idea if the properties are close enough to molten solder for capillary action to still work well or not.
First: I understand that due to clearance issues with SMD components, my question may not be valid. But if clearance issues were not a problem, would it theoretically be possible to create a delidder that applied rotational movement to the IHS? This would have the affect of creating very minimal stress to the solder near the center of rotation, and increasing stress farther away from the center. I think this might cause gradual metal fatigue in the solder that fails first at the perimeter, and gradually continues towards the center.
2 หลายเดือนก่อน
You should also have some logic in the control board if the CPU doesn't seem to heat up at an expected rate due to the temp sensor not making contact or if it's damaged. You also might want to do some rough temperature sensing by measuring the resistance of the heater or some backup way of stopping thermal runaways.
couldn't you just use high temp tape/glue to glue the heater to the IHS? instead heating the entire tool? You already have hole in the middle of the IHS slider. Also with this setup you could integrate the temp probe into the heater.
90 euro is a fair price if upcoming next gen cpu delidders could fit in that same heater case aswell, it could be one time buy, or atleast for multiple upcoming gens. Very nice job! keep it up.
The heater pricing makes sense, but its still pricey. Any chance that the outer dimensions of the delidders will carry forwards, to match the inner dimensions of the heater for future cpus?
I reckon this is perfect for boutique system builders rather than individual PC builder-owners. What amazing improvement the delidding makes. I wonder how much R&D Intel spend on the heat spreader & alternatives to it, given how much extra it costs their customers on cooling to get their CPUs below that thermal max.
Could you include something in the slider so that it grips the lid and lifts it about 1mm once it separates it from the die? If it holds it there while everything cools you might be able to just open it all up and lift the lid out.
Since heating is involved, you think it's possible to create a clamp design with a hinge that allows for some quick release that can lift the heated IHS as it opens like a waffle maker?
I don't understand why the second thermal cycle is needed. Can't you just machine a little extra space in the top piece so that when the glue breaks, you just invert the whole rig to let the lid fall into the extra space?
Been wondering this for a while, but is it possible to make a CPU riser and only place the CPU in liquid hydrogen/nitrogen to prevent condensation from forming elsewhere?
An extra cost for the chip heater portion, if it can be reused with other chips to be delidded, would take the initial sting off the cost. Not so much for Intel Core Ultra, but I have to wonder if that would help improve the odds of a successful delid with AM5? 🤔
Careful with chip height measurements, differences in thickness in mutli-die chips is a classic issue. Direct die cooling might be a very bad idea, especially with classic thermal paste which are relatively low thermal conductivity. The extra "distace" from the cooling interface (of whichever kind) can be an issue much sooner than you might think. Try to get temperatures of all the dies to make sure they also get adequate cooling. Maybe they did the finish very well tho, looking forward to that follow up video! Especially at the edge interfaces of the dies
Could you design this to hold the CPU inverted with space below it, allowing the IHS to fall once the solder is softened enough for the IHS to come free? That would make this a 1 step process.
I’d love to discuss simplified fatigue simulation and calculations with you. Should be able to estimate hypothetical fatigue for nominal materials and simplified or “equivalent” material models that approximate composites (like chiplets or pcbs)
why are the SMDs a problem for a contact frame? can't they just have the frame raised up right there? doesn't it only have to contact the IHS and can be any shape outside of that?
Oh man, I have the mycro direct die block for the 14900k sitting on a Z790 Apex and I need the silver top for the mycro direct die, it looks so good. Hopefully you offer them as a separate accessory so I won't have to buy a whole separate mycro direct die just for the silver top.
Honestly I think 90 euro is very fair. I was figuring you would say double that! Considering the work you put into bringing this to market that is more than fair!
I was quite surprised as well!
I don't think that price includes the actual delider part though just the heater, but either way it's really cool!
Yes I agree.
I agree for a low volume item with all that R&D the price is a lot lower than I was expecting.
@@martinrwolfeI had pretty much assumed it'd be more expensive than the CPU.
Designer: How should it look like?
Der8auer: A box.
xD I made it myself. I can make functional stuff but not good looking haha
I respect it
As Dan Gelbart says: If it is functional, it is beautiful.
@@der8auer-enI find beauty in functionally designed products.
@@der8auer-en A clean looking box looks better than edgy alien designs. Just keep doing your stuff, as it looks great even for poor people like me who couldn't ever buy it but at last can watch it.
Roman, you are a treasure to the community, just want you to know we appreciate you dude.
No chance I'll be buying a 285K and delidding it, but 90 euro for a small volume production run heater product seems pretty good value to me (given the amount of engineering required to design and implement it).
It’s worth it. I’d pay $200 for one
Not me. I'd get an air fryer and test how accurate it is with my fluke temperature meter.
Unless you're doing direct die I don't know why you would do it, replacing the TIM with conductonaut is a bad idea because conductonaut dries out in a year or so, turning into dry black flakes.
@celeriumlerium8266 conductonaut I haven't heard of gallium drying out like that unless it's formed a metal oxide by reacting to copper or eaten away at aluminum? My 4790K de-lidded with conductonaut on the die has been fine for 5 years
@@kekoraaaa He has a video on it, "I missed the obvious problem with my Threadripper". It converts into gallium oxide quickly due to unknown conditions. I put conductonaut on my 7700k and had the same problem, between die and IHS (no aluminium) it turned into a dry black "dust" after about 1 and a half years.
I recently delidded my 7700k with the Die Mate 2 and I was similarly shocked when I saw the temps. Stock with an NH-D15 it was around 65-70c during a run of R23, but afterwards it topped out at like 50-55c. Absolutely crazy. Thank you for developing such great tools for the community =)
Thank you for showing how much work goes into it and the development process
Fantastic Roman. Really good that you share your work with us. Thanks.
I'm thinking de-lidding stresses the Foveros interconnections on the tiles.
Great content. When do you ever see a manufacturer explaining there products so well and even showing revisions and the differences
The guy who made the CHRONOS high speed camera used by some yt channels have done the same on his TH-cam channel
It would be great if you could get Intel to disclose the temperature at which the die can deteriorate or get damaged. This is crucial information considering people using other methods might just cross the threshold and damage or degrade the processor.
Silicium is a strong material. The die itself will stay intact way before some of the tin balls on the PCB and the pads under the ceramic caps start to melt and make the chip unusable.
Preheating plates for tin soldering can go above 250°C and a reflow process or a hot air station can peak at more than 350 so there is some comfortable margins with this indium solder with such a low fusion point under 200. Since those YTers started delidding soldered CPUs, I've always wonder why they didn't use such inexpensive tools.
I would be more worried about the SMD components on the CPU pcb board. If their solder melts, some may fall off, especially if there is violent shaking, like the one when you break the regular glue.
Inert die should withstand solder melting temperatures used during manufacture.
Diffusion does exponentially increase with temperature, but it's such a short time that it would still be meaningless. Also, the CPU has to be soldered at the factory (with much higher temperatures), so I guess it's fine to heat it up to sub-200 degrees temperatures (while CPU is off ofc).
Aren't Intel cpu's degraded out of the box?
@christopherjames9843 very funny. Hahaha...
i don't understand why this channel doesn't have more subscribers. Every video is interesting.
Not enough hair
Because this is a new channel he made exclusively for English viewers. His original channel where he used to mix German and English content, has double the sub count.
this channel is new, roman used to upload a english version and then a german version to a channel, but decided to split the content and now has this channel and the original one
The answer is obvious : nobody needs this!
Also it's a very technical channel dedicated to a small niche market of the computer community.
Mom: Son what are you doing with the Air Fryer?
Son: I'm delidding a 285k, what does it look like?
Mom: okay honey, just make sure to clean it out after you use it.
🤣🤣🤣
I can feel that. When I was 17 years old and did Dry Ice OC at home I had to dry my mainboard afterwards. I put it in the oven at 80°C. My mom was like "what the hell are you doing here"
@@der8auer-enit seems you were doomed from the start to become an engineer 😂
Wouldn't it work faster in the microwave?
@@briankleinschmidt3664 too much metal.. you'll get fireworks.. 🤣
@@robertdelange2071 or you can get the new styropyro microwave and use the plasma :)
A CPU doesn’t ‘arrive’ until Roman de-lids it.
we de-lid as Romans.
Excellent video mate and incredible work! It's so cool seeing all the behind the scenes effort that goes into making these tools
we need cats 8auer.. get it together
Got an idea : why don't you offer delidding service directly, or even resell delidded chips ? It's becoming so complex and risky...
I mean Caseking could sell pre-delidded CPU's but I guess it's not worth the RMA hassle. Normal users don't delid so you can pretty much be sure the delidded CPU's will be hammered with OC.
"for aale my de-lidded Core Ultra, never over-clocked, one careful owner..."
Probably too much risk. Selling proven delidders is much saver than providing a delidding service where besides killing the CPU during the delidding process there also have risks associated with a CPU which is not well protected and (partially) fool-proved with an IHS. Things like transport accidents, user error during installation and warranties (most countries including Germany have warranties by law that cannot be signed away) can become an financial burden for a company
I would have to give 2 years warranty according to German law if I offer a delidded CPU for sale. And there is no way I would do that with a product that I have no control over
@@der8auer-en do it outside of Germany where EU laws don't apply. like Norway. ^^
amazing solution, might become new standard for delidding.
This engineering you've taken the time to share is impressive.
You could analyze the die's convexity (with the water block mounted) with a high-powered CT Scan, but that would kill the CPU at that ridiculous level of radiation.
You really are brilliant, Roman, and you deserve more subscriptions.
Whats about the mechanical sample he mentioned?
Feels like that would cost a fortune but would be cool to see
It definitely would be costly. Linus has one of those CT scanners, maybe a collab is in order?
everything is much simpler, you can make a strip of plasticine around the chip, after pressing on the plasticine there will be an imprint. After that, all that remains is to measure the height of the plasticine around the perimeter of the chip. If the chip bends, then the plasticine in the center will be equal in height to the plasticine in the corners
Beautiful job Roman!👍 World of the custom PC DIY modding wouldn’t be what it is without Thermal Grizzly!💪🤘Thank u for your contribution!🫡😎
Good job Roman! Love your work! Literally cant wait to get my hands on the new delidder.
Figuratively can't wait to get my hands on one, personally.
good job Roman, very proud of you and the team
9:52 you can still see the IHS right? I would loosen the bolts, apply a suction cup to the IHS and lift it simultaniously. That way its a single-step process.
Can maybe even create some vacuum seal in another way, like having the top of the IHS delidder have a rubber seal, with a nipple on top, and than just attach one of those handbilges to create a vacuum seal to lift it together with the delidder top.
EDIT: ah its already finalized, nevermind :)
The glue (or whatever it is) around the silicon got burned quite a bit and for a second I thought it was dead. Glad that this worked.
Really interesting upload, Roman. I really enjoyed it and there’s a ton of helpful information here. Well done on the next gen delidding tool.
I really enjoyed this short review of the whole process of delidding the new chip and designing the delidder. :)
Love your work Roman 😁 I think I speak for all of us, we appreciate all that you do for the enthusiast's community.
It's always so interesting to hear about how this works and what people are doing with extreme stuff like this. It's super scary, but I love watching you all do it! lol
This was an incredibly entertaining and informative video. Insane decrees in temps there. Hopefully this will be their format for the coming years of this platform and we see some huge gains.
Love the cooler setup! Awesome video, very impressive temps!
Oh man, I remember you from xtremesystems forum so long ago. Great channel. 🏆
Thank you der8auer for your consistent time and effort into making these products. Your level of transparency and honestly are unparallelled and you're one of my favourite PC TH-camrs to watch!
it is also stated as 5.7 gigahertz boost in the specifications, so they didn’t lie
Fantastic work! I'm really impressed by your end product.
Ty sir for taking the time out of your schedule to make this full tutorial ❤❤
I enjoyed this. Great work on that heater.
Epic work & video! Well Done, Roman!
So much work for this platform. Too bad the CPUs probably arent worth the trouble.
But all of this will likely be useful in the future, too, so at least the knowledge will be helpful.
I'm waiting to see how deliddable the 9800X3D will be. Would love to run it direct die. :)
Absolute no need to delid a 9800X3D, it won't reach high temps anyway.
@@jankees4037 a better CPU to water delta means I can run hotter water in my loop, increasing thermal efficiency on the radiator for less fan speed (noise).
That's what delidding does.
Thank you for the work and insight on this
So much respect for you to go through all the effort that you do. You personally test and tweak the settings of the CPU in order to test and verify everything to the best of your capabilities. On top of that, you go out of your way to separately sell the products in order to not force anyone into purchasing something that they may not want or need. You genuienly want the best product and best user experience for the buyer and it is conveyed to the audience so big props man! ❤️💯🔥👍
Quite amusing seeing the evolution of your delid tools. I've still got my delrin plastic one for the ivy bridge era chips. Things were a bit more simple back then lol
Looking at the bear die I kinda chuckled at Intel’s comment from a few years ago: “we don’t sell glued CPUs”. Looks like someone finally learned how to glued CPUs and sell them🤣
I not only have no plan to upgrade to the Arrow Lake CPUs, but wouldn't take the chance to de-lid one if I did.
Still watched the whole video.
Great work Roman 👍
What was the Hello Kitty canister for? lol
Great video btw, every der8auer part is always insanely high quality, keeps impressing me.
If CPU manufacturers sell naked CPUs as an option, then this whole "circus" wouldn't be needed.
Thank you for your service anyway Roman.
Their transport would be a mess. If products that are designed to take a beating sometimes already come damaged, imagine the courier yeeting a CPU without heatshield
They should, however Roman’s TG deliding tools would be out of business. Still both AMD and Intel should opt for naked CPUs besides classic IHS ones👍
Too much risk for the reputation. Idiots end up buying it, crack the die, try to RMA under false pretense, manufacturer has to prove it wasn't DOA with a hairline fracture.
All of that just to satisfy probably 10-50k people worldwide.
You'd have people crying about 'why isn't my i3 released as bare-die and only the top of the line chip gets it' or complaining that 'I'm paying more for a binned CPU with NO metal? This doesn't make sense'
@@bazooka712CPUs come inside a nice plastic blister that does NOT allow you to bend pins (AM4 models), so that shouldn’t be an issue
make it a Der8auer special edition that comes with a water cooling kit. pack it with IHS but unsoldered straight from intel or amd. limited series of course depending on water blocks available at the moment.
Insanely good temperatures, for such a powerful processor, and at 5700mhz there is almost no change in temperature, very good video bravo.
Why not design it so from the start you put cpu upside down? Then you would just cut open a witness window on the side to able to see if IHS has moved or not. Avoiding the heating of the IHS the second time and spilling the indium on the PCB from the first heating. Originally in factory when they put IHS and substrate together the IHS is facing down for that reason.
Also scraping off the indium leftovers on tile design - could this part be as dangerous as delidding itself? Maybe it's safer to scrape diagonally to avoid catching the wedge between the tile's border?
It's tricky because the heater obviously has to be thermally bridged to the IHS, so the heater would have to be able to hold onto the IHS stronger than the force pulling the IHS off the CPU. Probably simpler and easier just to do it this way, if it doesn't cause any issues for the CPU to get this hot.
Scores and temps look good! CB23 score is pretty much same as my 9950X delid and using one of your new WBs
Glad to see Asus HQ hooked u up with a 50th anniversary Hello Kitty Thermos for your OC'n needs!
15:30 This is getting super fancy! If that's the route you're taking, Roman, why not create a gearbox to turn one screw in a single direction and make the delidder move back and forth when it reaches a stopper on each direction (similar to a differential in a car)? This way, you can keep the screwdriver in place all the time, turning it in only one direction while you watch the delidder and its gears do all the work.
Are we machining the gears? Also.. do we have the Swiss dwarves on staff to assemble the tiny gears, as they do for the watches? Also.. could we compromise on a custom Allen wrench that allows a full rotation based on the height of the heater so you don't have to do that thing, that I was seeing in the video, that we've all been through with an Allen wrench lol.
Роман, крышку лучше сдвигать не вдоль кристалла, а поперёк, как в скальпаторе для 1700 2:27. Для этого предварительно отпаять два мешающих резистора 3:30, а после снятия крышки вернуть их на место.
A friend and I used an oven in the chemistry lab to ease delidding of his 7950x3d. We got the CPU to 100°C (we were afraid of going any higher) and then popped it into the Delid Die Mate. It seemed to work well, but without a comparison CPU, we could not actually determine whether it made any difference.
use floss to cut glue, then position the cpu upside down between two cans, apply heat from underneath with hair dryer until IHS falls off... simple and 100% safe AM5 delid
You always find a way to delid any CPU even if the manufacturer has warned you. Greatness...Next on der8auer: deliding the roof of my house to improve cooling and overclocking my heating/cooling system 😆
Do you think it would be possible to cut fins directly into the ihs instead of delidding? Maybe we can finally make use of that thick heatspreader on AM5.
Oh also will the indium solder stick to the surface mount components? I would rather not have to clean up tiny tiny solder bridges haha
14:25 What about partnering with one of those USB-C adapter makers and sell it separate at your shop?
Rowan is a beast, nicely done has always.
Curious if the heating element can be placed onto the IHS itself so when you move the delidder to break the glue, take the PCB with the CPU and drop the IHS that quick. There would also be less heat getting to the PCB and related parts. Might need a different way of tightening the delidder to make this process a single-shot but is worth thinking about.
Man that was quick...
the 3D printed case for the heater controller is absolutely fine!
Quick comment for final removal, instead of second heating step, how about upside down process and upon click, some kind of suction applied, e.g. syringe with suction cup on the end.
outstanding work. Thank you
Is there an option to order the silver 1851 mycro direct die pro block?
The heater controller could have a progress bar, ETA and a piezo for acoustic indicator that its ready - like start clicking near target and beeping on target
might be a good idea to send it with a link to this video on the box/manual of the delinter
Howdy,where be the felines? 😁
Not inside the IHS...
@@IIARROWSI hope not...
Yes where are Der8auer CEO Sheik and CTO Makita?
Delidding cat is necessary for success.
Please compare the vout of the delidded to the non delidded.
What about "renting out" the delider with a kaution of the full price of the part + 2 way shipping, and then ship it back and refund the kaution minus the rental fee (maybe of 30-40 euro per week? maxing out at the full price if someone exceeds the kaution value).
guess you're a big StarGate movie fan .... your constellation marking tablet tattoos is a great giveaway ... excellent video thank you for all the knowledge your share... cheers 😎
it is a lot of fun for someone with extreme overclocking needs, and i hope you sell a-lot of kits as a previous buyer of your delidders, but i have to be honest and say that a max ~5 or 6% in best case gain in multicore for all the expense and work/risk should carefully be considered by everyone
I did something similar with the 12900K. Ran Prime95 for 30 minutes. Pulled the CPU and delid. Took one push to come part. The problem for Arrow Lake is the SMDs clearance of course. Still doable at home.
Can you do a video on proper overclocking of it please? (Bios settings in detail). Would be cool to see how to push the GSkill CUDIMMs to 10k as well (voltages, cas, etc.).
Looks great! Can't wait to get it.
I wonder if the heat damages the Intel Tile 'Glue'?
Shouldn't. Will just copy my reply from a different comment: "No that's fine. You have to keep in mind they also have to solder the CPU to make it. And according to the indum to silicon soldering patents, they ususally run this at about 180°C."
the chiplets are not glued, thy just solder them
they could use resing to keep them in place but the heat used is not that high i think
On the first heat cycle. Would it be possible to add a tool that pull the lid (magnet, suction) off and keep it lifter until cooled down? It would fit through the opening in the sled. Just a thought...
10:08 why not just cut a hole in the delidder so you can flip the whole thing over and the IHS falls away from the CPU while everything else stays in place?
I'm curious, will copper desoldering braid suck up liquid indium? Scraping it off doesn't seem that difficult, but being able to remove the bulk of it with copper braid might be handy as well if it actually works. I just don't have any idea if the properties are close enough to molten solder for capillary action to still work well or not.
Der8auer delidding cat shirts when?
interesting idea :D
Schroedinger's de-lid tool
First: I understand that due to clearance issues with SMD components, my question may not be valid. But if clearance issues were not a problem, would it theoretically be possible to create a delidder that applied rotational movement to the IHS? This would have the affect of creating very minimal stress to the solder near the center of rotation, and increasing stress farther away from the center. I think this might cause gradual metal fatigue in the solder that fails first at the perimeter, and gradually continues towards the center.
You should also have some logic in the control board if the CPU doesn't seem to heat up at an expected rate due to the temp sensor not making contact or if it's damaged. You also might want to do some rough temperature sensing by measuring the resistance of the heater or some backup way of stopping thermal runaways.
seem to be a must-do for 285k deliding but this heating tool might be really helpful for universal deliding works
Thanks for providing all this data for me.
Genius tools for the enthusiast,great video.
couldn't you just use high temp tape/glue to glue the heater to the IHS? instead heating the entire tool? You already have hole in the middle of the IHS slider. Also with this setup you could integrate the temp probe into the heater.
90 euro is a fair price if upcoming next gen cpu delidders could fit in that same heater case aswell, it could be one time buy, or atleast for multiple upcoming gens. Very nice job! keep it up.
14:00 That’s a beautiful transparent print! I get nowhere close to that clear with PETG on my Bambu. Even at extremely slow print speeds. 🤔
The heater pricing makes sense, but its still pricey. Any chance that the outer dimensions of the delidders will carry forwards, to match the inner dimensions of the heater for future cpus?
I reckon this is perfect for boutique system builders rather than individual PC builder-owners. What amazing improvement the delidding makes. I wonder how much R&D Intel spend on the heat spreader & alternatives to it, given how much extra it costs their customers on cooling to get their CPUs below that thermal max.
Damn they killed the ring clocks :D
5:19 I think the reason this may be more risky is that i read this uses 3D stacking like Intel Foveros first used in.... lakefield?
Could you include something in the slider so that it grips the lid and lifts it about 1mm once it separates it from the die? If it holds it there while everything cools you might be able to just open it all up and lift the lid out.
What about wicking the indium while it's still hot with solder wick?
Since heating is involved, you think it's possible to create a clamp design with a hinge that allows for some quick release that can lift the heated IHS as it opens like a waffle maker?
I don't understand why the second thermal cycle is needed. Can't you just machine a little extra space in the top piece so that when the glue breaks, you just invert the whole rig to let the lid fall into the extra space?
Likely having two step would be the simplest way nothing goes wrong lol
yea it's about keeping it as safe as possible. 15min of your time is usually fine vs. losing a CPU :D
Okay how about the 245k? Thats more interesting. Does it have a different cpu tile?
should be slightly smaller die. I can probably do a follow up and check that :)
I'm not in the deliding scene, but watching this video is very very pleasing. A contact frame for me is suffice.
Your a mad lad to do all this.
Been wondering this for a while, but is it possible to make a CPU riser and only place the CPU in liquid hydrogen/nitrogen to prevent condensation from forming elsewhere?
An extra cost for the chip heater portion, if it can be reused with other chips to be delidded, would take the initial sting off the cost. Not so much for Intel Core Ultra, but I have to wonder if that would help improve the odds of a successful delid with AM5? 🤔
Careful with chip height measurements, differences in thickness in mutli-die chips is a classic issue. Direct die cooling might be a very bad idea, especially with classic thermal paste which are relatively low thermal conductivity. The extra "distace" from the cooling interface (of whichever kind) can be an issue much sooner than you might think. Try to get temperatures of all the dies to make sure they also get adequate cooling. Maybe they did the finish very well tho, looking forward to that follow up video! Especially at the edge interfaces of the dies
Could you design this to hold the CPU inverted with space below it, allowing the IHS to fall once the solder is softened enough for the IHS to come free? That would make this a 1 step process.
I’d love to discuss simplified fatigue simulation and calculations with you. Should be able to estimate hypothetical fatigue for nominal materials and simplified or “equivalent” material models that approximate composites (like chiplets or pcbs)
why are the SMDs a problem for a contact frame? can't they just have the frame raised up right there? doesn't it only have to contact the IHS and can be any shape outside of that?
Oh man, I have the mycro direct die block for the 14900k sitting on a Z790 Apex and I need the silver top for the mycro direct die, it looks so good. Hopefully you offer them as a separate accessory so I won't have to buy a whole separate mycro direct die just for the silver top.