1. This site uses cookies. By continuing to use this site, you are agreeing to our use of cookies. Learn More.

Random Restarts - Need Some Advice Please

Discussion in 'General Hardware Issues' started by nabed4t, Feb 21, 2020.

  1. nabed4t

    nabed4t Established Techie7 Member

    Hey everyone,

    Over the last month I’ve started having issues with my new PC crashing. It’s only ever happens when I’m gaming, so far it’s happened in Fallout 4 and GTA 5, as they are the only games I'm playing. There doesn’t seem to be a pattern, sometimes I can play for 5 hours and it doesn’t happen, other times I will play for 10 minutes and it happens. The weird thing is it seems to correspond to when I do something in a game, like click on an object or open a door. If the game is running, but I’m away from the PC it doesn’t seem to happen. So, what happens is this, the computer hard resets every time, no BSOD, no error codes. GPU and CPU temps are fine, I’ve tried to catch the LCD code display on the MOBO when it happens but the code may change by the time I look at it, I think it displays a code “54” on the MOBO, but I may be wrong.

    What I’ve tried so far:
    • I’ve checked the event viewer: I’m not really sure what I’m looking at but there does seem to be a consistent code posted at the time of the crash. A Critical error, Source Kernel-Power with the event ID 41. I’ve read that this could be a faulty power supply. There are however quite a few error codes in the event viewer ranging from “The SecDrv service failed to start due to the following error: This driver has been blocked from loading”, to “The server {AB8902B4-09CA-4BB6-B78D-A8F59079A8D5} did not register with DCOM within the required timeout.” I’m unsure if these are related.
    • I’ve checked the reliability history which shows a critical error at the time of the crash that states “Windows was not properly shut down”.
    • I’ve checked all the connections on the PC, re-seated the RAM and scanned for virus/malware with no change.
    • I’ve changed the power connectors/power boards with no change.
    • I’ve run Memtest86 and this has come up with quite a few errors.
    • I've rerun Memtest86 on each individual stick of RAM and I'm only getting errors on one stick.
    I thought I'd isolated the issue to a stick of RAM, however the restarts are still happening even when using the "error-free" stick. I'm thinking it might be the power supply? My old Thermaltake Toughpower 850W W0131 doesn't work with the new computer so I'm unable to test it.

    Is there anything else I should complete?

    Thanks heaps.


    OS: Windows 10 Pro, OS Version 10.0.18363.657 (Win10 19H2 [1909] October 2019 Update)
    MOBO: MSI X470 Gaming M7 AC Motherboard
    CPU: AMD Ryzen 5 2600
    GPU: Gigabyte Aorus RX580 8G
    RAM: Corsair Vengeance LPX CMK16GX4M2B3000C15 16GB (2x8GB) DDR4
    Power Supply: EVGA SuperNOVA G3 Modular 650W Power Supply
    HDD: Samsung 860 EVO 2.5in SATA SSD 1TB, Seagate Barracuda 2TB ST2000DM008, Kingston SA1000 M.2 NVMe SSD 240GB
  2. Digerati

    Digerati Super Moderator Techie7 Moderator

    How new? Still under warranty? Did these problems start from day 1? Only happening when gaming suggests (but does not conclude it is) heat.
    What is "fine"? And how did you check?
    First, make sure it is PassMark’s MemTest86. There are several memory testers with very similar names but that is only one that remains under constant development. You should let it run for several passes and get no reported errors – not even one.

    However, no software based memory tester is 100% conclusive. If they report any errors, even one, the RAM is bad. But it is not uncommon for them to report no problems, yet the RAM still fails in use, and/or when paired with other RAM. So, swapping in all new RAM is often the best test - not always easy.

    What BIOS version is installed? I see on the webpage there was an update released last September that addressed memory compatibility.

    Why doesn't the TT PSU work?
  3. nabed4t

    nabed4t Established Techie7 Member

    Hey Digerati,

    Thanks for the reply, the answers are below.

    The computer was bought on the 5th of January 2019 and still under warranty. The problems only started in the last month. I initially thought it was the RAM as only one stick was giving errors, so I have put through a warranty claim to return the RAM. However, since the other stick is still giving me the restarts I'm not so sure now. But you did mention that they can report no errors but fail in use so this may be the case. The issues have only started in the last month. I haven't changed anything in the PC or installed any new software in that time.

    Using MSI Afterburner to monitor in-game. The CPU sits on a max of 55 degrees Celsius and and GPU is peaking at 65 degrees Celsius in Fallout 4.

    It is PassMark's MemTest86

    Unfortunately the free version only allows 4 passes. I have run the test 6 times for each stick of RAM and I've received the same results each time. One stick reports errors (the number of errors fluctuate between 1 and 20), the other is error free.

    The BIOS is version 1.3 E7B77AMS.130 18/07/2018. In the past I've read that updating a BIOS can be dangerous. This is the first system I've had with a UEFI, is the process of updating the BIOS simpler and less dangerous than it used to be? I've watched a few videos on YouTube on how to do it and is seems pretty simple.

    I'm not sure why it doesn't work. When it is plugged into the new system the system will not boot, although the lights on the MOBO light up. I know the PSU does work in my old system though.

    Thanks again for the reply. Would you suggest the next step being to update the BIOS? Also, should I go through with the return of the RAM? Even if they are not the cause of the restarts, there is something wrong with at least one stick.
    Last edited: Feb 22, 2020
  4. Digerati

    Digerati Super Moderator Techie7 Moderator

    Hmmm, really? I didn't realize that but 4 is plenty. Typically, if there are errors, they will be found in the first pass anyway.

    Yes, your temps are good.
    It still can be. But so is crossing the street. So does that mean you never cross the street?
    Yes, it is possible you can brick a motherboard should the firmware upgrade fail during the upgrade process. But chipset/motherboard makers have worked hard to minimize that possibility. I've done many 100s, perhaps over 1000 in the last 25 years or so since I had an upgrade brick a motherboard. And that happened because a co-worker decided to walk behind my workbench and tripped over the power cord in the middle of the flash! What are the odds of that?

    This could be your motherboard failing. If you call your motherboard tech support, they are going to tell you to make sure the BIOS is fully upgraded before they issue an RMA anyway. So you might as well get it over with. As with any major maintenance task, make sure you have backups of any data you don't want to lose. I always ensure my computers are running off a good UPS with AVR. If you don't have an UPS, make sure there is no severe weather in your area and you should have no problems.

    As for the TT PSU, assuming it has all the necessary power connectors, I don't see why it would not work. If me, I might try it again. For sure, whenever troubleshooting hardware, I want to make sure I am feeding my hardware good, clean, stable power. And while EVGA SuperNova PSUs are great PSUs (I have one in this computer), until Man can create perfection 100% of the time, there will always be units that fail prematurely. So if the TT won't work, and because EVERYTHING inside the computer case depends on good, clean, stable power, I would try to beg, borrow, or steal another supply to try, just to see what happens before assuming another component is bad.
  5. nabed4t

    nabed4t Established Techie7 Member

    There are quite a few updates after my version. Would they be cumulative or do I need to update to each version?
  6. Digerati

    Digerati Super Moderator Techie7 Moderator

    I don't know. Probably a mixture of both. Just let Windows decide. It will sort it out.
  7. nabed4t

    nabed4t Established Techie7 Member

    I've installed the Bios update, unfortunately the problem persists. I've also replaced the suspected sticks of RAM, and yet the problem persists.

    Try as I have, I've been unable to procure another PSU as the TT one I have refuses to work with my current system. I assume it's a power supply to the MOBO or CPU issue. I'm looking into getting a reputable new inexpensive PSU like this or this and testing it as I'm running out of ideas. Do you think a reinstall of windows would be worth a go or does it sound like a hardware, not software issue?

    Thanks again for your help.
    Last edited: Mar 20, 2020
  8. Digerati

    Digerati Super Moderator Techie7 Moderator

    Reinstalling Windows should always be a last resort option after trying everything else - especially when it sounds like a HW issue, which this. Reinstalling Windows not only wipes out your data and installed programs, it can put you months or even years behind in updates, including critical security updates. Plus it often does not resolve the problem and even when it does resolve the problem, it destroys any evidence of the problem so you don't learn anything to prevent recurrence.

    So again, if you don't have a working spare PSU, I would see if you could borrow one.
  9. nabed4t

    nabed4t Established Techie7 Member

    I've managed to get my hands on a Corsair 550w CV550. I will hopefully have it next week and I will test it out.

    Also, I've tested the new RAM as a matter of interest and fortunately no errors were detected. Hopefully it is a PSU issue. I will post again next week when I've tried the new PSU.

    Thank you for all of your advice, Digerati.
  10. Digerati

    Digerati Super Moderator Techie7 Moderator

    Well, understand the CV series is Corsair's entry-level budget line of supplies. I personally would not recommend it be used as a permanent solution in any build. However, it would be perfect to use in this scenario - to verify if your current PSU is the problem, or not.

    How did you test your RAM? If you used a software based program like MemTest86 (the only program I recommend), it is important to note while software-based RAM diagnostic tools are good, none are 100% conclusive. If they report any errors, even one, the RAM is bad. But it is not uncommon for them to report no problems, yet the RAM still fails in use, and/or when paired with other RAM. So, just like power supplies, swapping in all new RAM is often the best test.

    FTR, to conclusively test your RAM, you need to use sophisticated and very expensive test equipment, like this $2,495 Memory Tester (and that's for the cheap model)! So it is usually easier (and cheaper!) to swap in known good RAM and see what happens.
  11. nabed4t

    nabed4t Established Techie7 Member

    I definitely won't be using it as a permanent solution, simply as a diagnostic test.

    I tested the new RAM the same way I tested the old, using MemTest86.

    The good news is that with the new Corsair PSU I have experienced no restarts over the last 3 days in situations where I was experiencing restarts with the original EVGA PSU. I have sent the possibly defective PSU back to the store, I hope to hear back from them in the next week or so.

    I'm currently looking at purchasing this PSU as a permanent replacement.
  12. Digerati

    Digerati Super Moderator Techie7 Moderator

    Scorptec is having me jump through too many hoops just to see the product. :( But if you mean a Corsair RM750x Gold, it should serve you well. That said, I still prefer EVGA (despite your problems) or Seasonic.

    However, 750 watts is WAY more than you need. As seen here, your maximum power demand of all your computer components together is just 438W!

    Since no calculator ever wants to suggest an underpowered supply, they all pad the results with some adding a lot more, some a little more. The eXtreme OuterVision calculator is the most conservative (a good thing, IMO) but then adds an extra 50W - a nice, safe, conservative added buffer for a recommended a wattage of 488W. And that's after I also added extra headroom by setting CPU utilization to 100% (the default is 90%) and Computer utilization and gaming times to 16 hours/day. I even added a Blu-Ray drive and 3 x 140mm fans for good measure.

    So you could easily get by with a quality 500W supply though if me, I would go for a quality supply ~550W, I might go up to 650W if I was planning on some major, significantly more demanding hardware upgrades in the next year or two.

    550W will give your supply plenty of headroom for cool and quiet operation (assuming adequate case cooling) and future expansions/upgrades. Bigger than 650W - unless you find a fantastic discounted price - is just wasting your money.

    A couple things to remember. (1) Bigger does NOT mean better. (2) The maximum power demands assume all the components inside the case (the CPU, the graphics solution, all sticks of RAM, all drives - including the optical drive, and all fans - case, CPU and GPU) are demanding maximum power at the exact same time - and that rarely (if ever!) happens in any real-world scenario. But IF it somehow should happen, it typically is for a few seconds only.
  13. nabed4t

    nabed4t Established Techie7 Member

    Yes, that's the one I'm thinking about.

    It probably is overkill, but I am upgrading to a Sapphire RX 5700 XT Nitro+ which recommends a 650W PSU and I managed to snag the 750W Corsair for a price comparable to a 650W Corsair PSU.

    Thanks again for the help Digerati, support like you is why I keep coming back to good ol' D.A.L!
  14. Digerati

    Digerati Super Moderator Techie7 Moderator

    Thanks for the kind words. And yes, it's a good site. Too bad it is not as popular as it once was.