Mystery Malfunctions: Mainboard killed by PSU?

Discussion in 'General Hardware' started by dms123, Feb 11, 2018.

  1. dms123

    dms123 n00bie

    Messages:
    8
    Joined:
    Feb 10, 2018
    First of all hello, and thanks for accepting me. :)
    The following is copied over, so it already includes some edits.

    Sorry in advance for a long post - I will try and give as much information as possible to hopefully get an informed response or two - I really am at a loss and quite desperate after a lot of struggle. Everyone who even bothers to read the whole thing is deeply appreciated. :)

    Story time first:
    So, about 2 months ago I built the following system:

    * MSi X370 Gaming Pro Carbon
    * AMD Ryzen 1700X
    * G.Skill FlareX 2x8GB DDR4 RAM (F4-3200C14D-16GFX)
    * Samsung 960 EVO 250 GB M.2 SSD
    * XFX Radeon RX 580 GTS 8GB Graphics card (for main monitor)
    * XFX Radeon R9 380P 4GB graphics card (previously used, additional monitors attached)
    * be quiet! Dark Power Pro 1000W PSU (old generation, previously used, for maybe too long)

    other probably not relevant stuff includes:
    * Corsair 110i Liquid CPU Cooler
    * 2.5'' SanDisk SSD used in previous setup
    * 2 3.5'' HDDs, also previously used
    * some case fans
    * all in a be quiet! case

    All of this is running on Windows 10 and it was going smoothly after the initial setup...

    The first problem I noticed (about a month in), was that my new (but cheap) USB headset attached to the front USB3 ports suddenly died, it started only producing distorted noises and then stopped being recognized by Windows. I shrugged it off as a probable case of you get what you pay for but put it aside for the moment to test it on another computer.

    After having a friends 2.5'' hard drive in the same port quite some time later and writing lots of data to it, everything seemed fine until I tried to read some of the data off of it. I "crashed", meaning the drive wasn't recognized any more, but everything worked fine on a small laptop. That's when I started to get concerned.

    I purchased a simple USB voltage and current tester and a PSU tester, as I was growing suspicious of the old PSU. The USB tester showed 5V flat when everything was fine, and dipped to about 4.7V when a problem occurred, like the headset making noise. The PSU tester showed the 5V SB voltage (not 100% sure what that does) as occasionally spiking to high - I'll try to include a picture.

    Further alarmed, but not panicking as the 12V rails seemed to be fine, I decided not to risk it and buy the newest generation of that PSU series, even if it was indeed failing, it served me well for a long time of heavy usage. I pushed the actual ordering back for a few days though, until I observed new issues: Occasionally I would hear a short "rattling" noise, which from experience with old, failing hardware I would accredit to some fan bearing making a noise when changing rpm. Which is odd, since the case fans are not only new, but also powered directly by the PSU and are controlled by a simple 3-step setting on the case. So the noise must have come from one of the 4 CPU fans or the radiator fans. (I am excluding the PSU fan because it was visible (mounted up) and running smoothly and silent as ever. Also I think I heard the noise for a moment today with the new PSU.) Unless I was wrong altogether and it came from one of the HDDs, or even a fan from the external jbod case or something. But I felt confident that the noise came from within the case.

    I finally bought the new PSU (Dark Power Pro 11 750W) and on the same day I get a freeze of the OS, after a few seconds of it "slowing down", meaning the mouse pointer actually gets sluggish before it freezes. No commands of any kind were accepted anymore. Later I would find out that when this happens, the RAM spikes to 100% usage lading up to the freeze, with no process in the task manager actually showing as using that much. I did not make it into the resource monitor in time so far to see if I can find out more. I decided to put the whole thing on ice 'til the PSU would arrive the next day.

    After I did arrive I went through the ordeal of taking the old one and all the cables out and redoing the cabling with the new one. Happy to have taken care of it, I put everything back together, booted it up and did a few updates for good measure. I decided to stresstest the system a bit and decided to run a round of PUBG. Game crashed instantly. No news there I thought, as there had been an update just prior and that's kind of what the dev does every time. So I started up Minecraft, mostly for fun, and that crashed with a nice message, which has never happened before. Shortly after, the RAM/freeze issue occurred again. On later attempts after making sure everything was up to date and tweaking some settings, I also noticed a weird "shrieking"/humming noise for a moment when trying to start a game. I could not place it at that time.

    Thinking the issue might not have been with the PCU (alone) after all, I ran a memory test. No results. I ran prime95 in different modes. No problems. Temperatures seemed to be withing reason. Looked into other options like Hot CPU Tester, but did not feel like purchasing a license to get stuff like the Chipset test, also all the tests take 6 hrs and I wasn't sure if I should do this kind of Burn-In test without narrowing down the issue first. I also attempted some narrowing-down and removed the secondary GPU. Same issues. Removed the main GPU and put the secondary one back. I got PUBG to start (with a lot of fan noise) but heard that alarming shriek again, this time continuous, 'til I killed the game (which was only in menu). I think it might have come from the lower area of the mainboard, possibly PCIe controller or audio stuff. Could be something entirely different also.

    Now... About my questions. My current theory (which I looking for confirmation for) is that the old PSU was bad and damaged the Mainboard. Which in itself is already pretty damn devastating. I am worried about further casualties as well though. For instance the fact that removing the main GPU lead to being able to start up a game that I couldn't otherwise. Makes me worry for it. Or the RAM suddenly filling up. Please, do correct me if I'm wrong, but would a fried Mainboard by itself be able to pull such a thing off? Or should I possibly start mourning my CPU too? Also, how likely is it that the RAM or the M.2 SSD were affected/damaged by this?
    I am interested to see if anyone could suggest methods of further narrowing down / confirming the issue without causing more damage. Also I'm really not looking forwarding to removing the mainboard and redoing all the stuff like thermal paste and cabling yet again. But I would be somewhat happier if I was sure it was the only broken part, as between the CPU, main GPU and RAM it is by far the cheapest. And even replacing just that I could barely afford right now after all the money I've already put into this. (PC hardware is not cheap in this country.) Not even to mention the time this will take...

    Anyway, to sum up: Would you say that my theory is likely? How can I confirm it or rule stuff out? Anyone here who has ever had similar experiences? Any advice on how to proceed in general? Or even on how not to loose my mind? :)
    Thanks in advance for any answers or hints, or even just making it all the way here :)

    EDIT: Just tried running Google Docs in Chrome for some urgent stuff I had to do. After the screen turned off and on a few times I was greeted with this (https://imgur.com/a/vDMSw) Windows 10 "bluescreen". A first for me...

    Pictures:
    Test of old PSU, HWMonitor stuff:
    https://imgur.com/a/5xtjp

    Logs:
    Full readout from HWMonitor:
    https://share.rtechsupport.org/HWMonitor.txt
     
  2. janas19

    janas19 [H]ard|Gawd

    Messages:
    1,570
    Joined:
    Jan 20, 2012
    First thing, you have serious hardware issues and you need to remove every component from your case. Isolating issue is difficult, though not impossible, when the components are cramped together in the case.

    My suggestion is to remove everything, spread the mobo and PSU apart, then determine exactly if the "shrieking" noise is coming from the PSU, mobo, or GPU. It sounds like your RAM is working fine if you passed a memtest.
     
    dms123, Spartacus and Speedeu4ia like this.
  3. Spartacus

    Spartacus [H]ard|Gawd

    Messages:
    1,829
    Joined:
    Apr 29, 2005
    No, it doesn't sound like the old PSU caused the problem.
    It sounds like the symptoms are the same with the second PSU.

    The CPU is fine, don't go RMAing that.

    As the last post said, you need to determine what the noise is.
    Turn off all audio while testing.

    Try testing only with the R9 until you figure out the problem.

    You could still have an issue with the mobo vs. the RAM even though it passes Memtest86, Prime95, whatever.
    Try dropping the clock speed on the RAM and then start the game and see if it still locks up.

    .
     
    dms123 likes this.
  4. MMitch

    MMitch Limp Gawd

    Messages:
    250
    Joined:
    Nov 29, 2016
    Have you tried a fresh Win install or Linux line USB key / DVD ? As mentioned, pull everything out of your case and install minimum components on a ESD safe bench + ensure nothing shorts... (Install riser on motherboard screws)
    Make sure all PSU cable are tight and reset optimal BIOS settings while ensuing latest BIOS is on.

    Inspect all PSU connections to device for contamination.
     
    dms123 likes this.
  5. Spartacus

    Spartacus [H]ard|Gawd

    Messages:
    1,829
    Joined:
    Apr 29, 2005
    I haven't worked with any Ryzen systems, are these voltages low or normal?

    OP, what do those voltages look like in the BIOS?
    What is the rated voltage spec for the RAM?

    Is the BIOS up to date?


    Hardware monitor Nuvoton NCT6795
    Voltage 8 0.94 Volts [0x76] (CPU VCORE)
    Voltage 9 1.34 Volts [0x54] (DRAM)

    .
     
    dms123 likes this.
  6. dms123

    dms123 n00bie

    Messages:
    8
    Joined:
    Feb 10, 2018
    First of all, thanks to you all for the response so far. :)

    I'll try to answer all questions:

    * I have not tried a new install so far, this one is fairly fresh. I might try that in the next couple of days when I have time, just to be safe.
    * Yes, the symptoms seem similar with the new PSU. Which is why I am worried the old one damaged the MoBo.
    * RAM timings/clock speed are on recommended. In my understanding the RAM was designed for Ryzen/X370 setups and is one of the few that was reported to work on full speed without issues at the time.
    manufacturer site gives these specs:
    Tested Speed 3200MHz
    Tested Latency 14-14-14-34
    Tested Voltage 1.35v
    Registered/Unbuffered Unbuffered
    Error Checking Non-ECC
    SPD Speed 2400MHz
    SPD Voltage 1.20v
    BIOS shows this: https://imgur.com/a/wfbJc
    * BIOS was updated when I built the machine. So less than 2 months ago. I'll check for a newer version tomorrow.
    * PSU cabling should be fine, as I did it just yesterday with new cables and a new PSU and the symptoms didn't change much. Not exactly cable porn, but this is what it looks like: https://imgur.com/a/53znE
    * I don't think the shrieking noise is coming from the GPU (as it is happening with both of them in either slot, just with different intensity), and I'm leaning towards excluding the PSU as well. Not sure what could cause the noise on the MoBo though? Some filter around the audio chip got fried maybe? But then the GPU-related bluescreen makes me consider the PCIe controller or something...

    Other questions I'm currently asking myself:
    Can the MoBo alone be responsible for the RAM suddenly filling up?
    Are there any affordable diagnostic tools that could make sure there is a problem with the board?
    After trying out a bit more, like the things suggested here, if I don't come up with anything, do you guys think it would be a safe bet to try and get the MoBo replaced?
     
  7. Spartacus

    Spartacus [H]ard|Gawd

    Messages:
    1,829
    Joined:
    Apr 29, 2005
    >>RAM timings/clock speed are on recommended. In my understanding the RAM was designed for Ryzen/X370 setups and is one of the few that was reported to work on full speed without issues at the time.

    I understand, but try clocking down the RAM anyway.

    >>SPD Speed 2400MHz

    That would be a great place to start to see if you can stabilize the system.

    .
     
    primetime likes this.
  8. dms123

    dms123 n00bie

    Messages:
    8
    Joined:
    Feb 10, 2018
    Thanks, I will try that today when I come back. Also a BIOS update if it's needed and I get around to it. Will report back then.
     
  9. RC-Heli-3D

    RC-Heli-3D Limp Gawd

    Messages:
    181
    Joined:
    Feb 12, 2018
    I think you found your issue. Mobo have caps/regulators that control voltage supply to USB. If your USB device is drawing more power than the board can supply. This will explain your problems. If your USB's on the board is crapping out, RMA it!
     
  10. Spartacus

    Spartacus [H]ard|Gawd

    Messages:
    1,829
    Joined:
    Apr 29, 2005
    Any new results from testing yet?

    The last post reminds me..... disconnect some of those drives while testing.
    You have a huge number of drives connected in that HWMonitor report.

    .
     
  11. dms123

    dms123 n00bie

    Messages:
    8
    Joined:
    Feb 10, 2018
    Apologies for being out of touch... have had very little time, but now there have been some developments:

    Got around to do that eventually, was a slightly newer version of the BIOS available, and updating to that also reset everything to standard settings, including RAM clock speed. It did not change anything.


    I finally did, and after finding the time to put everything back together and being all excited, it doesn't seem to have changed anything. Things are still crapping out when trying to start up a game. Haven't gotten the issue of the RAM randomly filling up just yet though.
    The only thing the swap seemed to effectively have accomplished is voiding my Win10 license really... And seeing as the MoBo apparently wasn't the problem, the manufacturer is likely to send my a bill after analysis, which is great fun on top of all this trouble :p


    Most of those are external, but I did unplug them. Together with the 2 internal HDDs and the 2.5" SSD. Only thing I kept was the M.2 SSD. I also unplugged everything else that I could really on my latest test. USB connector of the CPU cooling, front audio, front USB3, everything from the back short of the mouse, keyboard and LAN. Even swapped monitors. Did not seem to change anything.

    I have also used only one of the 2 RAM modules at a time, and swapped those around in the slots. I have alternated between using the RX 580 and the R9 380.

    Only thing left to do now that I can think of is using an entirely different GPU from my previous PC, which is a R9 270X if I recall. Don't expect it to do anything...


    All in all.. does anyone have any ideas here? EOS to the CPU from the previous PSU maybe? Really at a loss and frustrated (and out of money) at this point...

    Appreciate any help so far and any advice down the road. Cheers.
     
  12. BudgetBuilder

    BudgetBuilder n00bie

    Messages:
    12
    Joined:
    Dec 17, 2016
    I had a similar problem. It was a bad SSD. System would be running fine, slow down and then hang and reboot. I'm assuming it has something to do with the page file not responding and filling the ram? Not sure but after i swapped out the SSD i had no problems. At least it's something to try.
     
  13. dms123

    dms123 n00bie

    Messages:
    8
    Joined:
    Feb 10, 2018
    Yes that sounds somewhat plausible... I'll try anything at this point... Unfortunately unplugging all drives never changed anything, so if something in this scenario is the issue, it's the brand new M.2...
    I guess short of buying another one, the only thing I could do would be to take that out and install Windows on the old 2.5" SSD. Sounds like a plan actually now that I'm typing this... Will respond again when that yields results.

    On a sidenote, an issue that I mentioned earlier has been really persistent after the MoBo swap: The screen going black and then turning back to normal frequently, together with short freezes when browsing certain web pages. Trying to find the common denominator has me really confused though. Happens when streaming sometimes, like on YouTube. ALWAYS happens when going to Google Docs, BUT only in Chrome, never in Firefox. Bunch of other pages do this too, don't know what they have in common. Pretty sure Google Docs doesn't use Flash, right? So maybe JavaScript or something but then why does it only happen in Chrome... really puzzling.
    Other issues btw, have not occurred once since the replacement. Like the RAM filling up. Just the game crashes and browsing issues really. Which is why I usually would say GPU for sure. However I've tried 3 different ones... Also, video from files plays fine, on all levels of quality/bitrate I've tried.
     
  14. dms123

    dms123 n00bie

    Messages:
    8
    Joined:
    Feb 10, 2018
    So, I have another late update. Maybe this thread will help someone some day when their Google search finds it in some archive :D
    Turns out the old PSU must have not only killed the Mainboard, which was RMAd, but also both graphics cards, the RX 580 and the R9 380P. Talk about bad luck. I just tried a bunch of stuff including reinstalls, swapping memory around in the slots again, installing to the old 2.5"SSD instead of the M2 with no result until I replaced the card that was in there with my previous computers R9 270X. Suddenly it works. At first I was mainly confused because I could have sworn I tried that already. But it must have been before swapping out the Mainboard.
    Anyway now I'm just faced with RMAing the graphics card at the same place I did the mainboard. Also, it just had to be the most expensive part...
    Bit annoyed as well by the fact that this took so long to solve there is now the next generation of the processor I put in there out already... -_-

    Lastly, thanks for all the help!
     
    Meeho likes this.
  15. primetime

    primetime [H]ardness Supreme

    Messages:
    5,097
    Joined:
    Aug 17, 2005
    ahhh i think your jumping to conclusions for sure.....i personally seriously dont think both cards magically went bad without test them in a known good computer....it suck but for proper troubleshooting you need to be able to verify what hardware is good and bad. Most of us here are hardcore DIY people but most of us also have backup hardware to do testing....im kinda leaning on an incorrectly set up motherboard/cpu...or your board has issues. it be worth finding a way to test your cards even at a shop if it was the only way. you also need to be doing clean fresh installs to rules out software issues....just my 2 cents. Like was suggested to you try down clocking the cpu and ram.....cause for all know some of your memory could have issues.
     
  16. dms123

    dms123 n00bie

    Messages:
    8
    Joined:
    Feb 10, 2018
    I tried reassembling the older computer with one of the damaged cards in it to verify my problem, but it seems like that PSU is now proper dead, stuff only comes to live for less than a second then it switches off again. As for the mainboard issues, like I said, it was replaced. And I've been doing a fresh install with every part swap, did more than one today alone. I'm considering buying another PSU and CPU fan as those are the parts missing to construct another PC with what I have lying around. Would be good for when I need an optical drive or wanna set up a local server. Not sure if I want to do that before trying to get the graphics card replaced though.
    Don't get me wrong, I'd be doing ABA-tests all day long if I had all the hardware for it. Unfortunately this is my only AM4 PC.
     
  17. primetime

    primetime [H]ardness Supreme

    Messages:
    5,097
    Joined:
    Aug 17, 2005
    you cant try using the older setup with the newer power supply? i just dont think non generic supplys these days are destroying hardware often without a noticeable burnout taken place....I could replicate the same issue with freezing on mine by just changing some bios settings with voltages...you have tried with only one stick of ram idea? drop everything down to slowest speed....i have had my rig going on FOREVER and it just keeps on going (2 power supplys, 2 hard drives, 1 video card that went bad over 8 years lol0
     
  18. dms123

    dms123 n00bie

    Messages:
    8
    Joined:
    Feb 10, 2018
    I could in theory try it with the new power supply, but unfortunately I very thoroughly wired everything into the case after I replaced the mainboard, thinking it would finally resolve my issue. :/ I'd rather buy a decent but not too expensive spare PSU rather than taking all that stuff apart again, really don't have the time.
    To clarify about that bad PSU, it is definitively dead. I retested it with a PSU tester and it crapped out after a few seconds. Can't really blame be quiet! here as that thing surived several generations and configurations of my computer and has seen 10 years of heavy usage. It was just plain stupid of me to continue using it in the new build. Everytime I build or upgrade a PC for someone I give them the speech about using a good new PSU because it could kill other components otherwise. Should have listened to my own advise, especially knowing my luck... :D

    EDIT: Oh, and yes, I did try it with only one RAM module, tried either and tried different slots, not just the recommended that I had been using. Also tried everything on standard RAM clocks etc.