5800x Causing My B450m Pro-M2 to BSOD (WHEA_UNCORRECTABLE_ERROR) Nonstop. BIOS Are up to Date. What Are My Chances of a Refund?

Recently bought a 5800x here upgrading from my 2600x, working fine but starting to show its age.

The motherboard's page says the CPU is supported with latest beta BIOS which I upgraded to.

At first, it worked OK, then I loaded a CPU intensive game and my PC froze (black screen), this happened a few more times, the same way each time, so I tried reseating everything, then I started getting BSOD (same error in title). Then I started getting that BSOD as soon as I logged in to Windows. I got fed up, so I changed back to my old CPU and everything is back to how it was before. I have tried a few things, like uninstalling GPU drivers, reflashing to the same BIOS, lowering RAM frequency, windows memory diagnostic, nothing helped.

A few other things, I tried a less CPU intensive game (maxed about 10% CPU usage), ran fine for over an hour. I also left it on overnight with just a few Chrome tabs open and it managed to stay on. Also, the temps were always fine, never went over 60. here is my full setup if it helps. (Yes this site says there's incompatibility issues, but the board manufacturer says otherwise)

My real problem is figuring out if I really have a faulty CPU or if the BIOS simply doesn't support it. Either way, I don't feel like this can possibly be my fault and I should be refunded.

So what should I be doing?

Comments

  • What RAM you using? PSU? GPU?

    • Just updated with full build. I did take them out, gave them a spray, and was going to try using 2nd and 4th slot, then remembered my board only has 2 slots. Again, windows memory diagnostic tool didn't find any errors. PSU is antec 620w 80 plus bronze and gpu is rtx 2060.

      • Just saw your PPP list. Have you overclocked CPU in BIOS?

        • Never over/underclocked anything.

          • @viology: Ok. Well I haven't read much review into the MB, but I don't think it is a CPU issues. Feels like heating problem or RAM compatibility with CPU.

            Is XMP on or off? (Unless it doesn't apply to amd?)

            • @avoidfullprice: It was on at first, but turning it off didn't help. I also don't get how it can be a heating issue if the temps never went over 60 and the ram is on the qvl and never caused problems with my old cpu.

  • I had stability issues using Ryzen 3000 + un-tweaked XMP profile RAM (3200mhz) on an older generation B350 board.

    The stability issues were resolved using 1usmus's Ryzen DRAM calculator and running the RAM at a slightly slower speed (2933mts). It's a Windows application that helps you dial in safe or OC timings for your RAM. Try this out before you return your CPU, as it's likely that your RAM just isn't playing nice with the new CPU.

    You can download it from Guru3D https://www.guru3d.com/files-details/download-ryzen-dram-cal… and watch the guide on Hardware Unboxed.

    Should only take 5~10 minutes to do.

    • I was running at 3200, turned down to 2400, made no difference.

      • Any physical damage to the CPU? Bent or missing pins, dust in the motherboard socket that might have fallen in during installation?

        Was the CPU package fully sealed when you unboxed it? (there should've been anti-tamper tape on the box) Just checking that you didn't get a returned item.

        • No physical damage and no dust.

          The box was very slightly battered and the tape was in tact, but I was able to remove it and replace it and it looks the same.

  • what did the seller say when you contacted them?

    • Haven't contacted them yet, was wondering if there was any point.

      • You wont know until you try eh?

  • -2

    Don't shoot me, but maybe your PSU doesn't cut it with the 5800x in.

    • Even though pcpartpicker estimates it uses 404W?

    • you know how little power a stock 5800x pulls?

      • +1

        When fully loaded, does it pull more than a 2600x?

        The PC works completely with a 2600x and works with a 5800x unless under load.

        What causes a PC to reset under load? Heat and lack of power. Voltages are related to power.

        A PSU isn't as simple as looking at the total output (620W) and calling it a day. The output is split amongst rails. The GPU consumes part of the output when gaming, the hard drives claim a bit and the CPU needs enough not to reset.

        The RAM settings are fine because the PC stayed up overnight. If they were a problem, the PC would have reset within minutes of booting, if it POST at all. The PC would have reset with the 2600x in.

        • 2600x is a 6 core CPU running 12nm node, 5800x is a 8 core cpu running 7 nm node, realistically speaking even if it cost more it would be somewhere between 20w.

          talking about PSU, it show that you definitely don't know what you are talking about.

          OP's PSU have +12V@48A = 576w, this is what CPU/GPU/MB/HDD all uses, let's do the math, under full load:

          5800x full PBO OC < 150w, this much power in 5800x is almost un-cool-able.
          rtx2060 TGP = 160w, say you raise power limit all the way = 215w (ventus is a low tier card so no where near that)
          HDD Operating Mode, Typical (W) = 3.7w say 5w
          SSD ~ 3w
          fan ~ 3w each, say op have 5 fan all spin pretty hard = 15w~20w
          ram ~ 5w, say he OC with 1.5v, that's less than 12w
          MB ~10w, say it uses 30w
          say he have few random sh*t plugged in ~50w.

          150+215+ 5+3+20+12+30+50 = 485w

          this is if you run hectic OC on EVERYTHING and run EVERYTHING ALL FULL LOAD(which is impossible) and it will still have 100w spare.

          FYI my 10850k OC + stock 6800xt uses around 450w from wall plug under heavy load, I measured it myself, this is before power conversion, so it uses less than that.

          • @OMGJL: Do you need to account for that 80+ efficiency?

            80% of +12V@48A = 80% x 576w = 460W

            So if the worst case is 485W, it has exceeded the rating and start pulling the 12V down?

            Just throwing out there.

            https://www.techpowerup.com/review/antec-hcg-620m/5.html

            It would appear the 12V line could drop below 12.0V with over 400W load?

            • @avoidfullprice:

              Do you need to account for that 80+ efficiency?

              12v=48A is AFTER the efficiency loss, say you have 80% efficiency, the PSU will draw 12*48/0.8 = 720w from your wall socket.

              Only newbies make this conversion mistake.

              It would appear the 12V line could drop below 12.0V with over 400W load?

              As per ATX PSU Spec, computer parts should have +-5% tolerance on 12v, meaning as long as it doesn't droop below 11.4v then all is good. This is called VDroop, voltage are meant to drop while supplying large Amps. It's very normal.

              You should research more, before keep getting paranoid about normal stuff.

  • I was going to say the same, maybe the PS is struggling with the new CPU? (ie power supply is on its last legs) Ideally you have another PSU or can borrow one to try? If you run Prime95 does it fall over immediately? Some of the prime test run in cache only from so hopefully eliminates GPU and mem.

    • Prime95, test using small fft's

    • -1

      I find it hard to believe it's the PSU if people are using much better GPUs with only slightly more powerful PSUs. Is the difference between 2600x and 5800x enough to go from no problems to making a pc unusable?

      • It is really hard to say, that is why you need to try to eliminate different components. In my experience PSU's can cause all kinds of weird issues. In my experience it would be more far more likely that a PSU is faulty than a CPU. Next I would say MB/Mem.

        If you don't want to work it out, try the seller. You have nothing to loose

  • My guess is overheating VRM on mobo … most B450's have crap VRMs :/

    The MSI B450 Pro-M2 is E Tier on the mobo list, meaning it's only good for about 75W CPUs, with active cooling it MIGHT cope with 95W CPUs …

    The 5800X is a 105W CPU, meaning your only chance is to use some cooling over the VRMs :/

    7

    • -2

      If this is the case, how should I get a refund?

      • Tier list is here (click on the hidden data links):
        https://linustechtips.com/topic/1137619-motherboard-vrm-tier…

        —-SNIP—-
        Tier E: 75A current draw on little ambient airflow, maxed out 1600, 3600 and 5600X, a bit more than maxed out 2500X, 95W TDP CPUs at stock with downdraft air cooler
        MSI: B450 Pro-M2 (& Max), B350 Gaming Plus, Tomahawk, PC Mate, Bazooka, Mortar, Pro-VDH, Pro-VH Plus, Pro-VD Plus

        Tier F: 50A current draw on little ambient airflow, maxed out 1200/2200G (CPU only)/3100/3300X or 65W TDP CPUs at stock with downdraft air cooler
        MSI: B450 Pro-M2 V2, B450M-A Pro, B350 Gaming Pro
        —-SNIP—-

        So depending on your actual mobo (ie is it a "v2"), it might even be F Tier :/

        I guess returns depend on the retailer purchased from … Amazon might be ok for returns, can't imagine places like NewEgg/PCCaseGear/MSY, etc being too fond of a return … the other option is buying a better VRM mobo ;)

        Similar issue here: https://www.reddit.com/r/buildapc/comments/h0it7h/vrm_heatsi…

        7

      • Probably contact the seller which you seem not to want to do for some reason.

  • +1

    What cooler are you using ? If you used the same cooler you had for the 2600x on a 5800x that will be your culprit, as there is no way that will provide enough cooling for the 5800x.

    Overheating will cause bsods and the problems you have described.

    • according to his PCPP list it's an Noctua NH-D15 so the cooling isn't the culprit.

      I'd say with about 95% certainty it's a motherboard compatibility problem. Looking at the BIOS revisions, the latest Beta was pushed out on 2022-01-26 with the last stable version on 2020-06-16. The 5800x was only supported when the Beta bios came out. VRMS also appear to be naked with no heatsinks at all

      According to the VRM tier list, the MSI B450M Pro-M2 also sits at the very bottom. The VRM's are not good at all for high wattage processors
      https://i.redd.it/7n48gewun0p21.png

      • -1

        If its not the cooler then its the power supply, looks to be at least 5 years old, the efficiency over time goes down, and it wasn't that efficient to begin with since its bronze probably doesn't provide enough power for the system.

        • He could try undervolting the CPU to reduce power consumption to see if that works, otherwise I think some hw upgrades are required.

    • I thought NH-D15S was one of the better choices. Also, temps never went over 60 degrees before crashing.

      • It's not CPU temps to worry about, it's VRM temps … your mobo is good for 75W CPUs …

      • As long as its well maintained and has very little dust that is true. The dust can build up and cause issues, you should clean it every 3 months to 6 months depending on how dusty it gets.

  • Hi OP,

    I think your issue is the motherboard, one of two things.

    1) the motherboard might not have enough wattage ti supply the CPU with or
    2) the BETA bios is not working properly, after all it's a beta and you might have to wait for a proper release

    • I think it's the motherboard, but more likely vrm overheating. Sadly, looks like I am stuck with a $450 cpu I can't use until I upgrade motherboard (and psu probably).

      • PSU should be alright but yeah VRMs have been a problem to me before too

      • If you have a small 80mm fan, you could hook that up to blow directly on the VRMs. Or do you have a thermometer to check the VRM surface temps?

        • I already have to use a splitter just to use a 2nd case fan, so not really an option. And no thermometer currently.

  • what chipset drivers…? Makes a difference, especially if they are fixed an issue. The latest chipset drivers are on amd.com. Then get ryzen master and choose the power saving mode in control panel -> power options. Maybe try running the fans on 100% + cpu stress tests before trying intensive gameplay. If that's working, then switch the power option to best performance and retry. I'd also check all the physical mountings, so motherboard screws, thermal paste applied evenly and how the cooler is mounted.

  • So far I see almost no productive answers apart from speculating VRM and PSU,

    Put it this way, if a PSU is running in voltages less than what they meant to, you are unlikely to see BSOD but rather power cut out.
    if a VRM is running hotter than what it is comfortable with, it will downclock the CPU until it's no longer overheats (that's how it should behave in stock settings)

    I do strongly speculate it has to do with your RAM kit, go download TM5 from here,
    run it -> select profile -> select [email protected] -> program will close after profile selected -> run the program in Admin mode again

    you are very likely see errors spit out. Keep in mind even 1 error means you have a bad kit of RAM/ bad memory controller / something bad in between RAM and Memory Controller (motherboard memory trace).

    In the case you do see errors, pull out your RAM, clean it's golden contact, put it back in
    still crashes then manually set SOC voltage to 1.15v and see if it helps.
    still crashes then borrow a known good kit and try again.

    if you are in SA I am happy to help in person.

    • Thanks, il try to test ram when I can use my pc, but given there are no problems with it on my old CPU, and windows memory diagnostic didn't find any errors, I have my doubts.

      • I don't think the RAM themselves are problematic. It is their interoperability with your new CPU.

        I would suggest running a ram test with new CPU using a USB boot drive

        https://www.memtest86.com/

        See if the ram data is getting corrupted when using your new CPU.

        • by posting memtest86 webpage it shows that you are definitely not into PC building in recent years (after DDR3 & DDR4).

          mt86 take days before detecting a single vague error where tm5 with good profile(extreme1@anta777) take ~1.5hr to run 3 loop and detect most of the issue if there are any.

          Memtest86 SUCKS. no one should ever use it anymore.

          • +1

            @OMGJL: Totally bro. I'll leave this thread to you entirely. You're the only one here who is qualified to fix OPs problem.

            Pretty sure I said RAM itself is not the problem but the interoperability with the CPU. The test is to put a load on the system. But whatever.

      • you rebuild your PC, means you touched it, though it's unlikely that electric static charges will kill component (everything have ESP now) but there certainly is a chance.

        5800x have different memory controller than zen+ , although it is more capable in most cases, but it IS DIFFERENT, so a memory chip that was barely over stable line may fail in a different system.

        also 5800x has much more performance, so it may slam the RAM way harder than a 2600x, so a vague error may not be that noticeable with your old CPU.

        • +1

          What is a vague error? Sorry I've been out for long in PC building. Is it when the ram can't decide if it's a 0 or a 1? Or are we talking about different RAM timing?

          • @avoidfullprice: say when I ask a kid to do 1+1, I should always receive 2 as result, if the kid did it 1000 times, and come up result of 3 one out of 1000 times, then there's a vague error in this kid. Is hard to detect but wrong is wrong

  • +1

    WHEA error isn't a faulty CPU, it's usually incorrect settings
    if I'm not mistaken it can be because of too little vcore on AMD

    in theory you can just drop in a new ryzen chip into motherboards and you are golden, but not always the case
    if you used the 2000 series chip to flash the bios to be able to use the 5000 series chip, did you keep the settings for the 2000 chip?
    if so that's probably your issue - youve probably got the vcore set for the 2000 series chip?

    download ryzen master and do some cpu stress tests with hwinfo64 running and see if you get WHEA errors
    ie try setting clock speed to 3500 then stress test and see if you get a whea error.

    also some motherboards can be fussy about using a M2 drive + all sata ports, it looks like you are using all of them?
    try unplugging all the sata drives just running it off the single m2 and see

    id reset the bios to stock and also clean install windows. i have had whea errors from the strangest things, on 12th gen intel iv had them from a pcie power setting in bios (i know comparing apples and oranges)

    btw just a word of advice, you are on the right track by troubleshooting & asking for help here, but you would be better off going to OCAU or overclockers.net and asking in the ryzen threads for some help IF THESE posts do not fix your issue, because there would be nothing worse than you complaining, getting a refund or an exchange
    dropping a brand new cpu into the socket and then getting the same issues confirming its 100% user error and not the chips' fault

    also googling your motherboard name + chip will produce a lot of results, its worth checking through them because you are prob not the first with this issue

    https://www.reddit.com/r/buildapc/comments/qw3n4l/issues_wit…

    https://www.reddit.com/r/buildapc/comments/lxsyx4/msi_b450_g…

    https://www.youtube.com/watch?v=RmHDVKnGvfo

    hope this helps, ciao

    • Thanks for the advice. My motherboard settings have been reset (multiple times). Not sure if it helps but here are some voltages https://imgur.com/a/FWhGdJz.
      I don't see unplugging all SATA drives as a solution as I wouldn't use my pc without all of them. I did order a 550M so hopefully it will work on that

      • hey mate, first up in that image your dram voltage should be like 1.2 - 1.35 ?
        manually set that to whatever your ram xmp says

        also, some motherboards share SATA & pcie (m2)
        so use of m2 top slot reduce power to SATA 3 and 4 etc
        all this information is in the user manual, it should be next to the QVL list on your motherboard website
        speaking of QVL its worth checking your ram & ssd against that list for compatibility

        i highly recommend lowering the vcore and the clocks by bios or ryzen master and seeing if the problem goes away
        because a bsod could be created by poor thermals. best of luck

      • 0.256v for dram seems weird. I don't know how that's possible when stock should be 1.2v or 1.35v.

        iirc i started getting WHEA errors when I changed my cpu cooler. Taking the cpu out of the socket and reseating it correctly fixed that.

        Bad bios settings for the memory/controller seem to be the most common culprit for these errors.

      • VDDP stock should be ~0.7v, and RAM stock should be 1.20v

        definitely has to do with that.

    • WHEA_UNCORRECTABLE BSOD code

      IS NOT

      WHEA correctable errors (ID 18, 19)

      • will keep that in mind if ever im back on ryzen, with constant crashes 😂

  • From memory, think that code is relating to storage. Might be wrong.

  • its not just your computer, my dell 11th gen 11700K + RX6600 was fine until I started playing games and it would reboot, BSOD, audio would often pre-crash before it shit itself

    turns out the mb has an older chipset (dell money pinching) and small PSU (500w im lucky many came with 320w)

    the answer is you need a new MB with a better power handling and id also get a better PSU too (pcpartpicker website)
    ASRock X570 Phantom Gaming 4 WiFi ax all-in-one, full atx with wifi/bt $199
    Gigabyte GP-P850GM 850W 80+ Gold Power Supply Modular $99 (the 750w version is $88)
    *** watercooling AIO from 80-120

    ~the good, this will give you a good AM4 base for your CPU ~ a good base for stable OC if you want too
    ~the bad, next gen PSU has a 16pin smart connector for 600w GPU power for upcomming AMDRX7xxx&nVidia4xxx (1200w+) been anounced but not in retail channel
    and the AMD AM4 might have one last round of CPUs released for it but the new AMD socket for the new chip is comming soon in sep/oct

  • +1

    @viology

    Please do me a favour and find & turn off C-STATES in the BIOS.

    Report back.

  • Had similar problems recently following a CPU upgrade from 2200 to 5600X CPU.

    After trying all manner of fixes, I eventually reloaded winblows onto a spare HDD in order to see if it was software related rather than hardware related and low and behold, it worked perfectly.

    As a result, I wiped and reloaded winblows onto the original hard drive and reinstalled games etc and the PC is running fine.

    Worth a try if you have a spare hdd sitting around somewhere.

  • tried reinstalling windows?

Login or Join to leave a comment