You convinced me. I'm buying an AM4 socket for around 20 usd (since I don't rust my fingers to not bend the pins) and I'm sanding that IHS on both my 3100 and 5900X.
sweet you did it, i told you man its because of how the 5000 series ccd is setup. and i forgot the easiest way to hold the cpu is to use the plastic protector thing the cpu came in from amd. if you open it up opposite ways it will get a tight snug fit without touching the pins. but yeah lapping it was def a good deal huh lol. good job. and i went 500,1000,2000,4000,5000. just followed debauers diminishing return of smoothness
I couldn't resist the idea lol had to do it 😁. It's def helped with temps. And yeah I know about the plastic case but for the sake of the video I did not use it. Just wanted to show worse case scenario 😊. Thanks for the idea man!
I am really considering Lapping my 5800X. I have a Custom loop and it hits 89c no OC just PBO. I tried Oc'ing but it wasn't worth it, but maybe if I lap it that might change. Also, I do autobody repair and whenever your sanding any surface you should go straight back and forth no circles and let the paper do the work, so your right not to add pressure. I did a video on bent pins. I was going to show how to fix broken pins but I felt ppl would just not understand dropping the broken pin in the socket array it belongs in and they might fry their stuff lol Thanks for the video though, i was not sure if this was good for ZEN 3 or not and the 5800X is a HOT Processor.
My r5-5600x clocked 4.96Ghz PBO at negative -30 and +138mV: 3 core boost goes 4.95Ghz and all cores goes 4.78Ghz, one core goes 1.46Vcore and all cores 1.288Vcore, 50C idle temp and all cores 78C temp on 10 yeqrs old water loop. Mixed disiled water with tap water. I can not run my r5-5600x clocked 4.85Ghz manually anymore becue of my ekwb water block is so bad now, when my ekwb water block was almos new (used for some years) it could keep my r5-5600x under 60C at 1.41Vcore 24/7 but liquid metal damage .y water block even if it was copper. Will never use liquid metal again from thermal Grizzly.
I bent the crap out of the pins on an LGA socket a few years ago. My ham hands couldn't stay still enough to fix it and I thought the mobo was toast. Pro tip: small tweezers, magnifying glass, and a patient wife with calm hands can fix it. She knows nothing about electronics and did fine. CPU pins are even easier to fix but man that LGA socket was a scary thing.
A word from the wise. Listen up people! Hehe you got yourself a techie one there. Mine is clumsy, she'd drive the twisers through the socket. Then she'd look back at me with that deer in the headlights look. But ask this woman to make you a sandwich and she'd make a killer tripple decker, toasted and all the juicy meat and veggies you can dream of. 🤣
Did a simple undervolt to my 5600 (non-x) which has a turbo speed of 4.4ghz. With stock heatsink and very gentle amount of thermal paste, I couldn't manually run 4.4ghz very well during intense benchmark testing. So I undervolted the thing over .1v (I think .137V) and running strong at 4.1ghz. Temps dropped from low 80s to low 50-60s under heavy load. Almost no hit to my 3d gaming score and barely a dent in CPU performance. Gonna try lapping and CPU shining. The trick I always used was find something purely flat, it was suggested a piece of glass would work... and fine grit sandpaper, at least 2000. Tape sandpaper down flat and to the glass and rub away.
So it went from 86c to 83c? CPU CCD1 reports max recorded temps to 83c. Am i missing something or were these the results? 3c reduction at same voltage and clocks?
@@TechnologyHive but if there's a future bios update that enabled multiplier overclocking on my board (I have an x570 meg unify) then I'll do it for that sweet extra thermal headroom. Shame I can't oc my 5800x3d atm. Great chip tho, has 7000 series gaming performance even at stock specs 😁
not a good idea with the 5800X3D due to the layout of the internal chips. it actually creates a heatspot since the clearances are extremely tight. lapping the 5800X was the solution to a lot of users that had this specific cpu had lids made from a different type of copper. like most liquid cooling radiators the best is called red copper which helps transfer heat a lot better. the lids on the 5800X however were made of a harder copper which ended up trapping heat instead of spreading it out. and when you delid the 5800X theres a lot more glue than the other cpus. 5800X3D just needs to be undervolted most of the time. =)
was the AIO plate also "lapped"? i've heard that only sanding the IHS is mostly useless as the point is to sand both the CPU and the cooler is being paired with for proper or noticeable gains.
@@shanksisnoteventhatstrongbruh from my understanding - it depends on how flat the cpu cooler’s cold plate is. It would be important to lap the cooler too if it has a concave/convex face. If the cooler is flat and mounting pressure is even you should be ok with just lapping cpu IHS
I'm literally in the same boat ATM 😂 I read a few posts on Reddit where people claimed a 5-6C delta which is actually some solid overhead so I think I'm going to risk it this Thanksgiving
@@TechnologyHive ...no i have the NH u-12s and lockt on 4,65 GHz all cores with 1,300V in cinebench R23 not higher then 87,8°C max. hotspot and gain 12029 Points I'm thinking to upgrade the cooler tho but to the NH D-15
Got a 5600x in my first ever pc build and the core with the lowest temp has 24°C while the hottest one is running at 73°C. Wtf is this, is it normal? This being at full load
Full load doing what? Gaming? Video editing? Some applications are single threaded. That is too much variation in temperatures between cores. Have you double checked that the cooler is installed correctly?
@@TechnologyHive Mining. I know it's stupid to mine with cpu, but I just build this pc with a 1070 and wanted to see how much it could mine with both cpu and gpu. The cooler is installed correctly, but I did put it the wrong way the first time and had to lift it and put it in again without reapplying thermal paste (used the one it came with).I will buy new thermal paste and install the cooler again because even in idle, the lowest temp is 23 and the highest 40, which is still too much I think. I was just curious what is considered a normal temperature difference between the cores.
@@LetMePickAUsernameAAAAAAAA No no. It's not stupid to mine with cpu. I do too with my 5600x. But I do it using the unMinable app for coins like XRP. I have a small farm of 8 cards doing dual mining using trex miner for the LHR cards and SRBMiner for my AMD cards. And still I do cpu ming. But, PBO is disabled. The cpu efficiency goes down while mining, though the profits are slightly higher with PBO turned on, it's just not worth it.
There's still a lot of visible scratches on the IHS, which means you didn't spend enough time on a certain grit. I'd recommend using 2-3 separate sheets for each grit because your sand paper is most likely getting clogged from all the debris coming off the CPU. Although it might not make a huge difference in temps, it just adds to the mystique and looks more professional.
luckily i never had a uneven heatspreader, thermals were always nice (or it was anyways a „on-die“ CPU without the casual heatspreaders) I really would love if AMD and Intel would sell 2 different types of the same CPUs... one with HS and one with the direct die solutions like seen in most notebooks Ofc mounting pressure would have to be exactly correct but only think about older CPUs... they mostly had anyways plastic spacers around the PCB at the height of the die to make sure there is not too much pressure on the die. I built already with about 13 years my first PC (Athlon 64, early 2000s) there wasnt even a CPU aviable with heatspreaders I think this heatspreaders are that common because you cant make much wrong. But instead i would prefer better cooling performance over easy/safe to build, its anyways barely a CPU which got swapped out and HS-less CPUs arent really harder to install, only with slighly more precision for perfect contact
i mean, do you remember as the i7 lineup got that hot so delidding became very popular? Instead of removing them, intel could have easily made any i7 lineup with direct die... cheaper, more users would have used these CPUs, especially the overclockers (which are pretty typical at the Intel CPUs...), also me Im now even thinking about delidding my good old 4790k... the PC doesnt have any usecase anymore (sold my loved GTX1070 after about 5-6 years of gaming and mining) and i want to get easier over 4,8GHz (for the sake of CPU bench-numbers and fun haha) But i dont really think this make sense with a 92mm small tower/designer aircooler (despite saying „UP TO 150W“)
Problem would be warranty. It's easier for AMD and Intel to sell you a product you're less likely to break than one with a metal cap on it to prevent noobs from cracking the die. But yes it could be a nice thing be able to choose IHS or no IHS even if at owners risk. Maybe in the future. I did however delid my R5 3600 and ran it direct on-die cooling with a corsair H150 AIO. There was only 2 degrees improvement. Hardly worth the risk.
Technology Hive if the IHS would be at ALL CPUs that great with only 2 degree difference, then i would even prefer an IHS for exactly the reasons you said But i think that result doesnt fit to every CPU because of ryzens internal design. But at least on nearly every intel there was a huge improvement in direct die vs IHS when we talk about the modern Core i CPUs. Probably stock not really any sense, but for (extreme) OC pretty needed
@@harrison00xXx There certainly are good scenarios like you pointed out where direct on die cooling would make sense. But it also depends on the circumstances, true 👍.
82 degrees at 1.25V and 4,5GHZ on an AIO? Wow Mine is running with removed current Limits at 1.25V with Auto overclock at 4,850GHZ with the STOCK Wraith Prism cooler at 62 degrees max. On Prime 95 small FFTs O.o Without lapping CTR2.1 Test says its a Platin Sample soo.. jea Could be that idk. It can Run with 1.375V at 5.350GHZ all core And the 4,850GHZ on an Auto Voltage of 0.9-1.1V But that Temperatur difference between yours and Mine is massive
At 2:00 you can see the difference in height on the IHS in the center and it takes quite some time to get that leveled out. Also the board does not respect the manual voltage, which actually ran at 1.3v instead of what my 1.25v. Anyway, sounds like you got real lucky there with the silicone! Congrats!
You convinced me. I'm buying an AM4 socket for around 20 usd (since I don't rust my fingers to not bend the pins) and I'm sanding that IHS on both my 3100 and 5900X.
Have fun 👍🏼 😊 and take your time.
sweet you did it, i told you man its because of how the 5000 series ccd is setup. and i forgot the easiest way to hold the cpu is to use the plastic protector thing the cpu came in from amd. if you open it up opposite ways it will get a tight snug fit without touching the pins. but yeah lapping it was def a good deal huh lol. good job. and i went 500,1000,2000,4000,5000. just followed debauers diminishing return of smoothness
I couldn't resist the idea lol had to do it 😁. It's def helped with temps. And yeah I know about the plastic case but for the sake of the video I did not use it. Just wanted to show worse case scenario 😊. Thanks for the idea man!
I am really considering Lapping my 5800X. I have a Custom loop and it hits 89c no OC just PBO. I tried Oc'ing but it wasn't worth it, but maybe if I lap it that might change. Also, I do autobody repair and whenever your sanding any surface you should go straight back and forth no circles and let the paper do the work, so your right not to add pressure. I did a video on bent pins. I was going to show how to fix broken pins but I felt ppl would just not understand dropping the broken pin in the socket array it belongs in and they might fry their stuff lol Thanks for the video though, i was not sure if this was good for ZEN 3 or not and the 5800X is a HOT Processor.
Did you do it? If yes what was the result?
My r5-5600x clocked 4.96Ghz PBO at negative -30 and +138mV: 3 core boost goes 4.95Ghz and all cores goes 4.78Ghz, one core goes 1.46Vcore and all cores 1.288Vcore, 50C idle temp and all cores 78C temp on 10 yeqrs old water loop. Mixed disiled water with tap water.
I can not run my r5-5600x clocked 4.85Ghz manually anymore becue of my ekwb water block is so bad now, when my ekwb water block was almos new (used for some years) it could keep my r5-5600x under 60C at 1.41Vcore 24/7 but liquid metal damage .y water block even if it was copper.
Will never use liquid metal again from thermal Grizzly.
I stopped using liquid metal all together. Too many horror stories online about it. I stick with a good thermal paste, like mx-6.
I bent the crap out of the pins on an LGA socket a few years ago. My ham hands couldn't stay still enough to fix it and I thought the mobo was toast. Pro tip: small tweezers, magnifying glass, and a patient wife with calm hands can fix it. She knows nothing about electronics and did fine. CPU pins are even easier to fix but man that LGA socket was a scary thing.
A word from the wise. Listen up people! Hehe you got yourself a techie one there. Mine is clumsy, she'd drive the twisers through the socket. Then she'd look back at me with that deer in the headlights look. But ask this woman to make you a sandwich and she'd make a killer tripple decker, toasted and all the juicy meat and veggies you can dream of. 🤣
Did a simple undervolt to my 5600 (non-x) which has a turbo speed of 4.4ghz. With stock heatsink and very gentle amount of thermal paste, I couldn't manually run 4.4ghz very well during intense benchmark testing. So I undervolted the thing over .1v (I think .137V) and running strong at 4.1ghz. Temps dropped from low 80s to low 50-60s under heavy load. Almost no hit to my 3d gaming score and barely a dent in CPU performance. Gonna try lapping and CPU shining. The trick I always used was find something purely flat, it was suggested a piece of glass would work... and fine grit sandpaper, at least 2000. Tape sandpaper down flat and to the glass and rub away.
So it went from 86c to 83c?
CPU CCD1 reports max recorded temps to 83c.
Am i missing something or were these the results? 3c reduction at same voltage and clocks?
I might do that with my 5800X3D but I don't think I need to do that in the future cos it already runs pretty cool in my rig
I don't think you need to lap it then.
@@TechnologyHive but if there's a future bios update that enabled multiplier overclocking on my board (I have an x570 meg unify) then I'll do it for that sweet extra thermal headroom. Shame I can't oc my 5800x3d atm. Great chip tho, has 7000 series gaming performance even at stock specs 😁
not a good idea with the 5800X3D due to the layout of the internal chips. it actually creates a heatspot since the clearances are extremely tight. lapping the 5800X was the solution to a lot of users that had this specific cpu had lids made from a different type of copper. like most liquid cooling radiators the best is called red copper which helps transfer heat a lot better. the lids on the 5800X however were made of a harder copper which ended up trapping heat instead of spreading it out. and when you delid the 5800X theres a lot more glue than the other cpus. 5800X3D just needs to be undervolted most of the time. =)
was the AIO plate also "lapped"? i've heard that only sanding the IHS is mostly useless as the point is to sand both the CPU and the cooler is being paired with for proper or noticeable gains.
Only the ihs was lapped.
@@shanksisnoteventhatstrongbruh from my understanding - it depends on how flat the cpu cooler’s cold plate is. It would be important to lap the cooler too if it has a concave/convex face.
If the cooler is flat and mounting pressure is even you should be ok with just lapping cpu IHS
My 5950x gave me that "oh you're not putting my $700 face on sand paper, no you're not".. Look just now.
I wouldn't either. It generally doesn't make much of a difference.
I'm literally in the same boat ATM 😂
I read a few posts on Reddit where people claimed a 5-6C delta which is actually some solid overhead so I think I'm going to risk it this Thanksgiving
126W … hmm my 5950x runs at 208W. I guess this would not help me too much?
It clocks down to 3800 after a minute or two from 4500 at the start.
It does run hot yes. An aio is a must for this cpu.
@@TechnologyHive this is with my 420 eisbaer AIO
@@TechnologyHive ...no i have the NH u-12s and lockt on 4,65 GHz all cores with 1,300V in cinebench R23 not higher then 87,8°C max. hotspot and gain 12029 Points
I'm thinking to upgrade the cooler tho but to the NH D-15
Got a 5600x in my first ever pc build and the core with the lowest temp has 24°C while the hottest one is running at 73°C. Wtf is this, is it normal? This being at full load
Full load doing what? Gaming? Video editing? Some applications are single threaded. That is too much variation in temperatures between cores. Have you double checked that the cooler is installed correctly?
@@TechnologyHive Mining. I know it's stupid to mine with cpu, but I just build this pc with a 1070 and wanted to see how much it could mine with both cpu and gpu. The cooler is installed correctly, but I did put it the wrong way the first time and had to lift it and put it in again without reapplying thermal paste (used the one it came with).I will buy new thermal paste and install the cooler again because even in idle, the lowest temp is 23 and the highest 40, which is still too much I think. I was just curious what is considered a normal temperature difference between the cores.
@@LetMePickAUsernameAAAAAAAA No no. It's not stupid to mine with cpu. I do too with my 5600x. But I do it using the unMinable app for coins like XRP.
I have a small farm of 8 cards doing dual mining using trex miner for the LHR cards and SRBMiner for my AMD cards. And still I do cpu ming. But, PBO is disabled. The cpu efficiency goes down while mining, though the profits are slightly higher with PBO turned on, it's just not worth it.
solid video, ty for the advice going to do the same with my 5800x
Glad to help 🙂👍🏼
There's still a lot of visible scratches on the IHS, which means you didn't spend enough time on a certain grit. I'd recommend using 2-3 separate sheets for each grit because your sand paper is most likely getting clogged from all the debris coming off the CPU.
Although it might not make a huge difference in temps, it just adds to the mystique and looks more professional.
Yea finish do pretty much nothing all is from flatness
1000 for final? should be 2500
Makes no difference. 1000 is plenty good.
@@TechnologyHive nah
luckily i never had a uneven heatspreader, thermals were always nice (or it was anyways a „on-die“ CPU without the casual heatspreaders)
I really would love if AMD and Intel would sell 2 different types of the same CPUs... one with HS and one with the direct die solutions like seen in most notebooks
Ofc mounting pressure would have to be exactly correct but only think about older CPUs... they mostly had anyways plastic spacers around the PCB at the height of the die to make sure there is not too much pressure on the die.
I built already with about 13 years my first PC (Athlon 64, early 2000s) there wasnt even a CPU aviable with heatspreaders
I think this heatspreaders are that common because you cant make much wrong. But instead i would prefer better cooling performance over easy/safe to build, its anyways barely a CPU which got swapped out and HS-less CPUs arent really harder to install, only with slighly more precision for perfect contact
i mean, do you remember as the i7 lineup got that hot so delidding became very popular?
Instead of removing them, intel
could have easily made any i7 lineup with direct die... cheaper, more users would have used these CPUs, especially the overclockers (which are pretty typical at the Intel CPUs...), also me
Im now even thinking about delidding my good old 4790k... the PC doesnt have any usecase anymore (sold my loved GTX1070 after about 5-6 years of gaming and mining) and i want to get easier over 4,8GHz (for the sake of CPU bench-numbers and fun haha)
But i dont really think this make sense with a 92mm small tower/designer aircooler (despite saying „UP TO 150W“)
Problem would be warranty. It's easier for AMD and Intel to sell you a product you're less likely to break than one with a metal cap on it to prevent noobs from cracking the die. But yes it could be a nice thing be able to choose IHS or no IHS even if at owners risk. Maybe in the future. I did however delid my R5 3600 and ran it direct on-die cooling with a corsair H150 AIO. There was only 2 degrees improvement. Hardly worth the risk.
Technology Hive if the IHS would be at ALL CPUs that great with only 2 degree difference, then i would even prefer an IHS for exactly the reasons you said
But i think that result doesnt fit to every CPU because of ryzens internal design.
But at least on nearly every intel there was a huge improvement in direct die vs IHS when we talk about the modern Core i CPUs. Probably stock not really any sense, but for (extreme) OC pretty needed
@@harrison00xXx There certainly are good scenarios like you pointed out where direct on die cooling would make sense. But it also depends on the circumstances, true 👍.
Привет из России! Я полировал свой 5800х3д и водоблок до состояния зеркала, так что они прилипали друг к другу. Результат минус 6 градусов
Очень красивый мужчина! И спасибо, что смотрите мое видео! Привет из Норвегии!
@@TechnologyHive :з
82 degrees at 1.25V and 4,5GHZ on an AIO?
Wow
Mine is running with removed current Limits at 1.25V with Auto overclock at 4,850GHZ with the STOCK Wraith Prism cooler at 62 degrees max. On Prime 95 small FFTs O.o
Without lapping
CTR2.1 Test says its a Platin Sample soo.. jea
Could be that idk.
It can Run with 1.375V at 5.350GHZ all core
And the 4,850GHZ on an Auto Voltage of 0.9-1.1V
But that Temperatur difference between yours and Mine is massive
At 2:00 you can see the difference in height on the IHS in the center and it takes quite some time to get that leveled out. Also the board does not respect the manual voltage, which actually ran at 1.3v instead of what my 1.25v. Anyway, sounds like you got real lucky there with the silicone! Congrats!
@@TechnologyHive jea true
Thanks Mate