Have you ever damaged a CPU due to overclocking?

Discussion in 'Intel Processors' started by Bandalo, Sep 7, 2017.

Have you ever damaged a CPU due to overclocking?

  1. Yes, once.

    16.2%
  2. Yes, more than one.

    12.2%
  3. No.

    71.6%
  1. arestavo

    arestavo [H]ard|Gawd

    Messages:
    1,344
    Joined:
    Mar 25, 2013
    That would probably work on Link equipped corsair AIOs, but my H80s and H100 were the originals that didn't have any interfacing options. The H115i did, but I'm done with corsair for lack of longevity.
     
  2. Frito11

    Frito11 Limp Gawd

    Messages:
    272
    Joined:
    Dec 13, 2016
    degraded an IMC on a 6700k trying to max mem performance out a bit too much. cpu still overlocked just like before and actually still did just fine with even really crazy DDR4 speeds it just required much higher than normal vccsa and io voltages and couldn't run anything other than 2133 default on stock voltages anymore.
     
  3. Nenu

    Nenu Pick your own.....you deserve it.

    Messages:
    17,532
    Joined:
    Apr 28, 2007
    What speed ram + overclock attempt and voltages caused it?
    What voltages did you end up having to use?
    For my curiosity.
     
  4. Frito11

    Frito11 Limp Gawd

    Messages:
    272
    Joined:
    Dec 13, 2016
    was when i first got my Z270 Apex. i had been leaving VCCSA and IO on auto and running DDR4 4000 C12 timings at 1.9v. the cause was either the apex being overly aggressive on auto with VCCIO and SA with crazy ram settings (it was pushing 1.3/1.4v on them which is usually more than needed and getting dangerous on ambient cooling) or it was when i attempted to bring timings even tighter down to C11 and tried for 1.95v (failed) then for kicks tried 2V also failed but post codes changed so was a sign that i was doing something the cpu was very unhappy with. ever sense that point if ran at auto the CPU v-core would run SUPER high on auto (1.4v at 4.2 ghz!) and only run 2133 ddr4 on auto. I had to manually set all voltages to get it to act normal but at that point it worked just fine and still was even able to do DDR4 4000 C12 at 1.9v ram as well as xmp at 1.4v. vccsa and io needed to be raised to around 1.2v to make anything beyond 2133 work however.
     
  5. Nenu

    Nenu Pick your own.....you deserve it.

    Messages:
    17,532
    Joined:
    Apr 28, 2007
    I think the ram voltage harmed the CPUs IMC. Ram can take around 2V, the CPU cant.
    The highest voltage ram is 1.5V DDR4600, I get the feeling that is pushing the limit.
    Tread carefully, your CPUs IMC will probably degrade further with 1.9V VRam.
     
  6. McDork

    McDork Limp Gawd

    Messages:
    463
    Joined:
    Nov 13, 2007
    No, I go for middling overclocks though.
     
  7. Vader1975

    Vader1975 Limp Gawd

    Messages:
    418
    Joined:
    May 11, 2016
    Well, the answer is a touch complicated. I have destroyed 3 CPU's in the quest for overclocking and one motherboard.
    #1. I used a Peltier unit and it cooled so efficiently that it caused condensation to drip onto the motherboard. (see magic smoke)
    #2. I installed a cooler with just a very slight tilt on a Pentium 3 900 this chipped the corner of the die and made it inoperable. (This mammoth cooler was for overclocking)
    #3. I incorrectly installed a new cooler on an AMD Thunderbird 1200 which left a razor-thin gap between the CPU and the heatsink. So much so that the compound made it look like everything was all good to go. Then heat was applied and the compound didn't have enough pressure in-between and more magic smoke.
    #4. I had an Intel i7 970 that stopped working because it melted a pin on the CPU socket of the Rampage III extreme. I was able to clean that CPU and add alternate interface material on that 1 spot and get it working on another board. However, the board was destroyed and the CPU damaged but recoverable. It was likely due to the rather high voltage being used over time. So I would consider it overclocking related.

    Now I have had overclocked systems where components failed years down the road like motherboards and fans and pumps and video cards and PSU's but that is just wear and tear over time.
     
  8. Frito11

    Frito11 Limp Gawd

    Messages:
    272
    Joined:
    Dec 13, 2016
    yeah it probably was the trying for 2v on the ram. i before and sense then ran 1.9v plenty of times on multiple cpus for bench runs without any damage but i also now manually set in vccsa and io voltages because so far at least i've not ran into any 1151 cpus that need more than maybe 1.2v to do crazy ram settings for benching.
     
  9. dgingeri

    dgingeri 2[H]4U

    Messages:
    2,398
    Joined:
    Dec 5, 2004
    I had a Cyrix 5X86 100MHz chip that I decided to run at 120MHz. It worked well for the first 6 months, but then became unstable after a time, and trying to clock it back down didn't fix it. It wouldn't even run at 66 after that.

    I did the same with a P3-450 at 533MHz, an Athlon XP 2700+ at 2400MHz, an Athlon 64 3200+, an Athlon X2 5200+, a Core 2 Duo E8300, a Core i7 920, and a Core i7 3930k. All of them lasted months to even one up to a year and a half, but eventually became unstable. It's not a quick thing. It's a slow burn.
     
  10. Frito11

    Frito11 Limp Gawd

    Messages:
    272
    Joined:
    Dec 13, 2016
    yep that is what is known as degrading a chip via overclocking.
     
  11. SticKx911

    SticKx911 [H]ard|Gawd

    Messages:
    1,756
    Joined:
    Mar 14, 2004
    E6400 @ something crazy like just below 4ghz for over a year. Out of warranty. No crazy overvolt or heat (was on water). It just stopped working one day. In all fairness it could have been the board, but symptoms indicated cpu so I just used it as a reason to upgrade.
     
  12. d50man

    d50man W[H]iskey Lover

    Messages:
    3,033
    Joined:
    Jan 22, 2002
    damn how? can you elaborate?
     
  13. =Gunslinger=

    =Gunslinger= Limp Gawd

    Messages:
    491
    Joined:
    Jan 25, 2009
    Almost always from pushing them to the max on LN2

    Z77, Z87 chips would degrade from high volts and heavy load. Todays chips are more durable and don't degrade as easy.

    The 7900X and 7740X both died because of moisture on the board at high volts, I did a poor job of prepping the board.
     
    Hakaba and Frito11 like this.
  14. d50man

    d50man W[H]iskey Lover

    Messages:
    3,033
    Joined:
    Jan 22, 2002
    LN2 damn did you do any water runs before they died?
     
  15. Sarra

    Sarra [H]Lite

    Messages:
    73
    Joined:
    Sep 2, 2017
    I popped MANY socket A AMD CPU's back in the day. I popped a Palomino on accident, when I swapped it in for testing and forgot to reset the CMOS, and it tried running at 2.5 ghz instead of 1.8.

    I haven't done any OCing since Socket 939, though, so I can't say I'm in the 'modern CPU popper club'. Nothing sucks more than letting that magic blue smoke out of a brand new expensive CPU.
     
  16. bigblueshock

    bigblueshock [H]Lite

    Messages:
    70
    Joined:
    Oct 27, 2017
    Never damaged a CPU during overclocking. All my chips, I found I'd hit my max overclock (temperature wise) within or below intel's max voltage spec. And IMO if you're under inte's max voltage spec, you're pretty safe...
     
  17. greyboxer

    greyboxer Limp Gawd

    Messages:
    186
    Joined:
    Oct 23, 2017
    Considering I am testing my 3770K at 1.45volts, I think I should sub to this thread :D

    I had a 2500K degrade over time. Dropped from 4.8ghz to 4.6ghz after about 6 years of overclocked goodness at 1.46 volts. Was the catalyst for my upgrade to 3770K actually (yes, in 2017)
     
  18. acascianelli

    acascianelli [H]ardness Supreme

    Messages:
    6,555
    Joined:
    Feb 25, 2004
    I had a Pentium 4 2.4C running at 3.4 for over 2 years. It started becoming unstable at 3.4, then 3.2, then 3.0. Eventually it wouldn't even run stable at 2.4 with extra voltage.
     
  19. Dan_D

    Dan_D [H]ardOCP Motherboard Editor

    Messages:
    52,304
    Joined:
    Feb 9, 2002
  20. jfreund

    jfreund Gawd

    Messages:
    761
    Joined:
    Sep 3, 2006
    I've overclocked every CPU I've had since I bought my first PC, a Pentium MMX 150 that I ran at 166. I don't go for the bleeding edge, and I've had really good luck. I do remember chipping a naked-core AMD or 2, maybe a Duron and a Thunderbird, but obviously that was from cooler mounting. I killed my first 6700K with a botched delid. I've always aimed for longevity. Each system I've had since the 920 I ran at 3.4GHz is still in use in my house, and there are older systems still used by friends and family.
     
  21. magoo

    magoo [H]ardForum Junkie

    Messages:
    14,410
    Joined:
    Oct 21, 2004
    Never borked a cpu from overclocking.

    I did kill a C2Duo one time because I didn't seat the cooler correctly, hit the power and heard a "snap".
    The cooler was elevated off the CPU IHS about a millimeter....DOA.
     
  22. Susquehannock

    Susquehannock 2[H]4U

    Messages:
    2,609
    Joined:
    Jul 26, 2005
    Have never ruined a CPU from overclocking.

    Been out of the game for a long time. Back in the day I did OC the hell out of some AMD Thoroughbreds. 1700 and 1800 'low volt' DLT3C chips.

    100% stable at over 1ghz above stock clock. 266 bus speed running at 434. All on air cooling. Still have them all and fired up a few at stock settings in a Soltek nForce2 board about a year ago. No problems.
     
  23. MikeRotch

    MikeRotch Limp Gawd

    Messages:
    351
    Joined:
    Aug 23, 2013
    I killed a 300A once trying to push it past 500 Mhz.

    I managed to return it to my local computer shop and then got the 2nd one running at a more conservative 450 Mhz.
     
  24. Dan_D

    Dan_D [H]ardOCP Motherboard Editor

    Messages:
    52,304
    Joined:
    Feb 9, 2002
    I have mixed luck with CPU's. I've had CPU's that would barely do anything over stock. I'm the only one I know who's first Celeron 300A would only do 416MHz. I've worked with a 6950X that degraded to a point where it would barely run at stock speeds. I've owned a Celeron 1.8GHz chip that would go to 2.8GHz or more with ease. I've got lots of examples that cover both ends of the spectrum. We had particularly good luck with Core i5 2500K's. We had at least three or four of them that would do 5.0GHz or better. I had a C0 stepping Core i7 920 that would do something like 4.0GHz or more without doing anything special and a D0 that couldn't. That's the reverse of what normally happened.
     
  25. Kodin

    Kodin n00bie

    Messages:
    10
    Joined:
    Sep 18, 2017
    Yeh e7200 chip running at 3.8 ghz, too much voltage degraded the chip after 2 years. Could only do 3.4 stable, I cant remember what the volts were. Out of interest how much money did some of this Old cpu fuck ups cost you because they were very expensive back in the 90's, first Cpu i had in my first gaming rig was an AM2 3800 x2 OC'd at 2.4 ghz. ahh nostalgia
     
  26. spine

    spine 2[H]4U

    Messages:
    2,232
    Joined:
    Feb 4, 2003
    Yeah, pretty sure I've done a number on the IMC of my current Haswell 4790K now that you mention it also! :p

    My system has now become completely 'burnt' into its current overclock (see sig). If I change anything, it fails. I mean like, downclock mem and change nothing else, or change bclk down and adjust everything else back inline, fail. Stock, fail! It won't detect my SSDs at stock either! Then again, I thought this mobo blew a ram slot a year ago (stopped working). Then 6 months later I re-tested it and it was fine. Very odd. Also, one of the CPU pins has lost its head (top 0.5mm), so I bent it up a bit to make contact. That affected nothing though.

    So yeah, my system is most definitely slowly but surely dying from overclocking. Honestly, I'm absolutely mindblown that this 22nm Haswell with on-die FIVR has taken 2.2v(external) 1.52v (internal) for just over 2 years straight!

    And I've only lost 33mhz OC! :ROFLMAO:
     
    Last edited: Jan 5, 2018
    Nenu likes this.
  27. Dan_D

    Dan_D [H]ardOCP Motherboard Editor

    Messages:
    52,304
    Joined:
    Feb 9, 2002
    People always say things like this but it isn't true. CPU pricing has changed very little over the last 20 years or so the grand scheme of things. Many people keep suggesting that Intel would completely lose their minds and prices would sky rocket without competition. Yet, we haven't seen anything competitive from AMD until recently. We've had about 10 years without any real threat to Intel's performance dominance. You had some price points where Intel allowed AMD to hold a greater price / performance ratio, but it almost seems like charity, or even mercy. If this was intentional, it was to allow AMD to limp along on life support to avoid anti-trust lawsuits. In the past, Intel had far more competition, but the price range (accounting for inflation) of it's products was roughly the same. The real difference between the market then and now is that Intel had far more competitors in the arena than just AMD. Competition (price only, not performance) used to come from AMD, IBM, Cyrix, IDT, Centaur, NextGen Systems, and VIA depending on the year in question. Intel had no competition in any performance category most of the time and certainly nothing on the higher end of the spectrum. Everyone else was relegated to bargain basement knock off chips that only worked with software "most of the time" rather than all of the time. In the 1990's,

    Let's look at the numbers.

    Pentium (P54c) Pricing in 1996
    • Pentium 166 MHz =$498 ($783.93 today)
    • Pentium 150 MHz =$364 ($572.99 today)
    • Pentium 133 MHz =$257 ($404.56 today)
    • Pentium 120 MHz =$188 ($295.94 today)
    • Pentium 100 MHz =$134 ($210.94 today)
    • Pentium 90 MHz =$134 ($210.94 today)
    • Pentium 75 MHz =$106 ($166.86 today)
    Pentium Pro Pricing in 1996
    • The Pentium Pro Processor:
    • Pentium Pro 200 MHz 512k cache =$1,217 ($1,915.74 today)
    • Pentium Pro 166 MHz 512k cache =$809 ($1,273.49 today)
    • Pentium Pro 200 MHz 256k cache =$707 ($1,112.92 today)
    • Pentium Pro 180 MHz 256k cache =$600 ($944.49 today)
    • Pentium Pro 150 MHz 256k cache =$534 ($840.60 today)
    Source.

    Today's Mainstream Processor Pricing:
    • Intel Core i7 8700K =$359.99
    • Intel Core i7 7700K =$279.99
    • Intel Core i5 8600K -$249.99
    • Intel Core i7 7600K =$189.99
    • Intel Core i5 7500 -$179.99
    • Intel Core i3 7100 =$109.99
    Today's HEDT CPU Pricing:
    • Intel Core i9 7980XE =$1,799.99
    • Intel Core i9 7960X =$1,399.99
    • Intel Core i9 7920X =$999.99
    • Intel Core i9 7900X =$799.99
    • Intel Core i7 7820X =$499.99
    • Intel Core i7 7800X =$329.99
    Source.

    As you can see, the prices on the highest end of the spectrum are roughly the same when accounting for inflation. Keep in mind that I've used Microcenter pricing for today's CPUs. Microcenter is going to be just about the lowest you can go in terms of retail pricing. The Intel processor pricing from 1996 was from an old CNET article which showed pricing from CPUs in 1,000 unit quantities. There is some difference with Intel CPU's usually costing more than I've shown today, and procably slightly more if you were to look at retail prices in 1996. Overall, I think we have CPU's that are somewhat cheaper today, but not by much.

    Even PC pricing isn't really any different than it was in the late 1990s regarding high end hardware. These days you don't have to pay for separate drive controllers, network controllers, MODEMs, or I/O controllers. These are all integrated into the motherboard and though we pay more for motherboards, I think we come out slightly ahead. However, we are paying more for cases, power supplies, cooling hardware and motherboards than we ever have. The video card or GPU has gotten to be ridiculously expensive over the years as well. This is why a top end PC from the late 1990's was around $3,000-$5,000 and a top end PC today is still basically the same price as it was in the 1990's. Of course we also have a greater range of options in the market than we did in the 1990's. Back then, all PC's were more than $1,000 and today you an get a piece of shit Dell with a monitor, keyboard, printer, mouse, etc. for $299.99 or so. Of course those costs are absorbed by companies like Symantec who basically bloat the machine with adware before it leaves the factory, but I'm getting off topic here.
     
  28. ShuttleLuv

    ShuttleLuv [H]ardness Supreme

    Messages:
    6,890
    Joined:
    Apr 12, 2003
  29. loafer87gt

    loafer87gt Limp Gawd

    Messages:
    384
    Joined:
    Feb 2, 2005
    Recently killed a 3930K due to my second Corsair H100 AIO failing. I feel so horrible as the chip was a winner, hitting 4.9Ghz with little voltage. I was running at a mild overclock of 4.6 Ghz until the pump failed on my second Corsair unit. I did not have the Corsair link software installed, and I had my workstation rendering out an animation and it took me a while to realize that the reason my poor computer was running so sluggish and giving me BSOD's was that my 3930K was throttling horribly due to my temperatures being upwards of 95C for an extended period. I had to pick up another used 3930K, but this one doesn't seem to overclock nearly as well as my old unit only hitting 4.2 Ghz. I have since made the decision that I am never taking a chance with a AIO or water cooling setup on a processor again. Once my third repalcement H100 arrives it is going up for sale. I understand mechanical bits fail, but two pumps failing within a year and a half time time period each doesn't give me great faith in water cooling technology. I have since installed an Evo 212 and couldn't be happier.
     
  30. Batboy88

    Batboy88 Limp Gawd

    Messages:
    175
    Joined:
    Dec 25, 2017
    Back with AM3+ think Dogged out that 8320 at 5ghz and like 1.6v...that chip slowly degraded. then got a 9590 was a lot better, alittle over 5 and like 1.57v. still awful cpu's lol, Killed a old 920 and w3520 with too much PLL voltz
     
  31. thenjduke

    thenjduke Limp Gawd

    Messages:
    344
    Joined:
    Jul 24, 2007
    I have never damage one from overclock but have with static electricity which was not fun! This was long time ago on Athlon