Announcement

Collapse
No announcement yet.

Hardware issues - I'm just about hairless at this point. Some validation please.

Collapse
This topic is closed.
X
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

    [SOLVED] Hardware issues - I'm just about hairless at this point. Some validation please.

    Long story, started about a year and a half ago. I don't know for sure what's wrong and I hate just throwing money at parts hoping I guess correctly.

    Equipment: i7-6700k (1.5 yrs old), Corsair Vengeance RAM (1.5 yrs old), EVGA nVidia GTX 780 (3 yrs old), OCZ ZX 1000W PS (7-8 yrs old?), various new motherboards.

    Environment: Water cooled CPU, large tower case, plenty of fans - all spinning. Case has been open (no side panel) for months, so likely not a heat issue.

    Symptoms: All's fine and dandy day-to-day, then all of a sudden WHAM - shuts down instantly. No warning.

    History and behavior: This started with my first new mobo - an MSI. I had to RMA the first one right at the start (no post). Second one worked for awhile then the sudden shutdown started occurring. Happened most often when playing a game - either 0AD or War Thunder. Both are graphical, but neither really super taxing on the GPU IMO. Happened sometimes when not playing a game but very rarely. After each shutdown, if I pull the CMOS battery, unplug the computer, and let it sit for several hours, it will power back up. Eventually RMA'd the second mobo due to a BIOS issue. Third mobo had same issue. This happens like once every one or two weeks.

    MSI's RAM process is long and painful (weeks) and I assumed the mobo was the issue and ordered another mobo - Asus this time. Had good luck with my last Asus board - I upgraded with it still running fine just because I was like 4 generations behind on CPUs and put the MSI on the shelf. The Asus booted right up first time but booting was kinda slow. Checked with Asus and there was a BIOS update that resolved that issue so I went for it - and yes, the mobo bricked. There was a so-called "emergency BIOS recovery" but that didn't work. RMA'd the Asus and re-installed the MSI.

    SIDEBAR: Meantime, after I applied for the Asus RMA, I realized Amazon would just allow me to return the mobo no questions asked so I requested the return with Amazon, which I receive the OK for. The very next day we had a huge snow storm that stayed on the ground for weeks. Last time there was snow like that here was 1989 so obviously everything was shutdown for over a week, and I mean EVERYTHING. No mail, no UPS or Fedex, no trash pickup, nothing. So I emailed Amazon to inform them my return would be delayed. Amazingly, their reply was basically "Never mind, keep or discard the item, we refunded your money" along with something about me being a good customer, whatever (I live in a small town so Amazon is my primary shopping location!). I guess it wasn't worth the effort for them to deal with. Anyway, since they told me I could keep it, I went through with the RMA to Asus.

    The point of the sidebar is to explain this: Before I found out I was going to get to keep the Asus, I had already ordered a third mobo - an ASRock. It had already shipped and since the Asus mobo was free, I decided to let it play out.

    Enter mobo #3: I finally got around to installing the new ASRock. Just as before, all it fine and dandy for the first week, then today during a lunchtime game of War Thunder - down she goes.

    OK, clearly I'm slow, but I've pretty much decided it ain't the motherboard causing the issue. The numerous issues I've had that WERE mobo related - no post for #1 when new and a BIOS update bricking - has lulled me into believing that's where the problem lies. Apparently not.

    Today I checked the CPU heat after the shutdown - cool to the touch. I went through the RAM thing - swapped sockets, tried each stick solo - still no post. Cleared the CMOS - still no post.

    THEN I pulled the graphics card and weirdness began. With the GPU pulled, when attempting to turn on, the power supply cycles on and off every second. Apparently endlessly. With it re-installed, the PS powers on but still no post. This struck me as very odd. Then I pulled the PS and flipped it over so I could see it's fan and plugged in just the mobo (modular cable design) and attempt to turn it on - power cycles started again on-off-on... every second, but the fan on the PS did start to spin.

    THEN with the nVidia card out of the computer but plugged into the PS, the PS powered on.

    So right now, this is what I believe: The CMOS has nothing to do with it (just left-over in my brain from the earlier CMOS/BIOS troubles). The power supply is over heating or has some other problem. It's fan turns, but it's also old and is full of dust that I can't safely clean. The 3-4 hours delay that allows the restart is the PS cooling off. I need a new Power Supply.

    Unless someone has an argument with this conclusion that is backed with some experience, that's what I'm doing. I can't imagine anything causing the odd power supply behavior other than the power supply itself.

    Comments or thoughts?
    Last edited by oshunluvr; Feb 07, 2018, 01:36 PM.

    Please Read Me

    #2
    This is where it really, really helps to have a friendly PC repair nearby, one that isn't trying to purchase another company Mercedes from each job it does. I have one here, small privately owned. And one time, I even asked Best Buy if its Geeks could quickly check out a PS for a small fee--and they quickly did it for nothing, while I waited (the PS was confirmed faulty). Sorry I can't render any specific HW solutions (I probably would do as you and lean toward the PS).
    An intellectual says a simple thing in a hard way. An artist says a hard thing in a simple way. Charles Bukowski

    Comment


      #3
      I would be inclined to believe it's the power supply, especially with the "sudden shutdown" and, that it has repeatedly happened with replacement and different mobos. That you say the PS is old and contains dust you can't get at, and that have waiting a period of time, the PC again boots, points to a PS problem. PS's are relatively cheap. Get one that has the power to drive your system; don't scrimp on an underpowered PS. Get the biggest one (in terms of power output) that you can.
      Windows no longer obstructs my view.
      Using Kubuntu Linux since March 23, 2007.
      "It is a capital mistake to theorize before one has data." - Sherlock Holmes

      Comment


        #4
        Mike, good idea but no local PC shops around here. There is a Best Buy an hour away. I've never considered them for something like that.

        Paul, my thoughts exactly. I went back and found the order date - March 2011. so it's 7 years old. Warranty was 5 years so I guess I got my monies worth.

        Previously I had a 650W PS and during the time I owned it I gradually upgraded and added to my system. The CPU was a Q6600 (100W) overclocked to 3GHZ (+25W), a GT8800 nvidia card - eventually 2 of them, and at one point six hard drives. Then one summer I started having - wait for it - random power shutdowns. I did the math on all the stuff I had attached and I was at 638W on a 650W ps. The kicker was the heat - we didn't have A/C. As long as the room temp was 75f or below everything was fine. As soon as it climbed much above that the computer would turn off without warning. I did some reading and figured out the additional heat causes the power supply to perform less efficiently. Since I was already at the edge of it's max capacity, it couldn't take it. That's when I bought this one - a 1000W unit. Overkill even then, but I didn't know where I would go PC wise so I figured the extra capacity wouldn't hurt.

        External heat and capacity aren't the issue this time so something is seriously wrong.

        I just did the math on the current system and I'm using a little less than 600W - CPU @91W but only 1 video card and 2 hard drives, 2 SSDs, and the water pump. I'm getting a 850W Gold+ replacement.

        One of the hard drives is about to retire. It's really just a playground anyway. I keep second backups on it and a couple sacrificial partitions for testing. It's got like 90,000 power-on hours on it so it's ready for the bone yard. It's a pull from my server. My next upgrade will be a M.2 drive and the hard drive will exit, so my power use will actually go down instead of up.

        Please Read Me

        Comment


          #5
          Well, regardless, 'heat' is the most likely contributor to your issue. Exactly what is 'failing' due to heat is still 'up in the air'.

          This is a desktop rig, so consider leaving the case open so you have complete and unobstructed air flow.
          Windows no longer obstructs my view.
          Using Kubuntu Linux since March 23, 2007.
          "It is a capital mistake to theorize before one has data." - Sherlock Holmes

          Comment


            #6
            Since you are going to replace the PS anyway, why not take a high pressure air hose and blow it out? Or, disassemble it. I suspect that you have a thermal switch that is triggering the shutoff and because of poor air circulation through the PS it takes a long time for the iron core on the transformer to cool off. What eventually happens is that the excessive heat carbonizes the enamel paint insulating the copper windings and causes a short, which draws more and more current, heating up the iron core. When that spot fully carbonizes it will constitute a direct short circuit which will melt some of the copper windings.
            "A nation that is afraid to let its people judge the truth and falsehood in an open market is a nation that is afraid of its people.”
            – John F. Kennedy, February 26, 1962.

            Comment


              #7
              I monitor my temps in my panel. Yes, they go up when I play a game or transcode video but never anywhere near panic levels. I've seen 60s when transcoding multiple files at once on my CPU but panic isn't until 80. Also, it's never above 27 or so when this shutdown occurs. The GPU goes higher during game play but idles at mid-20s and peaks at high 30s. Well within operating temps.

              Here's what I think:

              The evidence points to graphics use - gaming activity is the usual precursor to the shutdown and the odd PSU behavior when the graphics card is unplugged. I suspect either a critical component on the GPU part of the PSU (it has specific ports for each purpose - GPU, motherboard, SATA, etc.) is worn out or it's what is overheating.

              The fact that a wait of 2-3 hours after shutdown allows restart points to heat. All the external parts that I can touch - heatsinks, CPU, GPU, MOSFETs, whatever - are cool within minutes of shutdown so I infer it's something I can't touch - like inside the PSU.

              I believe external heat is not the issue as;
              It's winter here and I keep the heat vents in my office closed because the computer put out enough heat to keep the room comfortable (mid-60s).
              The case has been open for months while this has been occurring because I keep having to get into it.

              Once I have the new PSU installed (this weekend), I'll run a stress test and see if I can "break" it and monitor the temps. Honestly, I use my computer 365 days so it's probably time for a new PSU.

              Please Read Me

              Comment


                #8
                Originally posted by GreyGeek View Post
                Since you are going to replace the PS anyway, why not take a high pressure air hose and blow it out? Or, disassemble it. I suspect that you have a thermal switch that is triggering the shutoff and because of poor air circulation through the PS it takes a long time for the iron core on the transformer to cool off. What eventually happens is that the excessive heat carbonizes the enamel paint insulating the copper windings and causes a short, which draws more and more current, heating up the iron core. When that spot fully carbonizes it will constitute a direct short circuit which will melt some of the copper windings.
                Good idea. I have an air compress so that'll be easy. The issue I have with this is how do I really test the PSU? I can check voltages, but unless I can stress it while monitoring I can't really achieve a valid test. I hate to keep forcing my expensive and needed PC parts to the stress of these shutdowns.

                Please Read Me

                Comment


                  #9
                  So GG, the odd switching on-and-off behavior when this happens: Any Ideas what would cause that?

                  Could heat expansion cause a temporary fault that causes this behavior? Seems odd that unplugging the GPU power connectors results is this happening when I would expect the opposite. Maybe the fault - a short or otherwise - is causing the shutdown, then it won't allow the proper voltage to be delivered to start up the CPU?


                  Occurs to me I could - once I have the new PSU in hand - run the old one until it powers off then immediately put it on a voltage meter and check all the voltages. That might shed some light. I could then also prove it's the PSU causing the problem by installing the new one and powering back up.

                  Please Read Me

                  Comment


                    #10
                    I guess the plus side is I may have a couple slightly used motherboards for sale...

                    Please Read Me

                    Comment


                      #11
                      One thing to remember is that power production, power conversion, and power control happen in several different places in a computer system. The motherboard being one of those locations. And there is power modulation/conversion circuitry on the MB. With (potentially) a lot of amperage passing through that part of the MB, there can be localized heat issues that can cause failure. So while the PSU can handle a lot of wattage, specific points within the rest of the system may be seeing amperage that on the high end of optimal. I know volts, watts, and amps live in a specific mathematical formula. But that doesn't keep problems away from places where power moves through circuit crossroads, if those crossroads aren't built for some conditions.

                      Dust is always a problem, and any case opening, intentional or not, will add to that blanket of insulation. In many respects, fans are useless - IF they are not allowed to funnel air efficiently. Actually, with a high end system, having a wide open case can be bad because the the fans are being pushed to move all the air in the room instead of just the air in the case. A properly sized fan with appropriately sized air openings in the case, can move air - and therefore heat - much more efficiently. A wide open case allows more dust to enter and just settle instead of being captured in the air stream. Cleaning out the PSU is critical to your general maintenance, and I'm not the best at this either, but it's also one of the hardest parts of a computer to keep clean. It has a relatively large fan, so good air movement, but it also has lots of places where dust gets trapped, and even a great fan won't move all the air out.

                      I've lost MBs and CPUs and every other component in more than one computer, and it always comes down to maintenance, the complexity of the components, air movement, and the passage of time.
                      The next brick house on the left
                      Intel i7 11th Gen | 16GB | 1TB | KDE Plasma 5.27.11​| Kubuntu 24.04 | 6.8.0-31-generic



                      Comment


                        #12
                        Humm ,,,I will go with the power supply as well , it's the oldest part , it acts wonky when the GPU is unplugged (sounds like a relay switch) and is full of dust,,,,,,the killer of most components heat wise .

                        if you decide to sell MB 's what will be the sale price ?

                        I've been thinking of maybe upgrading the TV box ,,,,,,a desktop that gives net to the TV .

                        VINNY
                        i7 4core HT 8MB L3 2.9GHz
                        16GB RAM
                        Nvidia GTX 860M 4GB RAM 1152 cuda cores

                        Comment


                          #13
                          Originally posted by vinnywright View Post
                          Humm ,,,I will go with the power supply as well , it's the oldest part , it acts wonky when the GPU is unplugged (sounds like a relay switch) and is full of dust,,,,,,the killer of most components heat wise .

                          if you decide to sell MB 's what will be the sale price ?

                          I've been thinking of maybe upgrading the TV box ,,,,,,a desktop that gives net to the TV .

                          VINNY
                          LOL, we'll talk!

                          Actually I have my old server mobo. With an Intel 945GC chipset, Single Core Intel Atom 230 CPU, 2x GbE LAN, 4x SATA II, VGA port

                          This: http://www.orbitmicro.com/global/ms-...--p-10191.html

                          No HDMI or DVI - VGA only. CPU and 2GB RAM included. Dual Ethernet tho...

                          You can have it for the cost of mailing, or drive down and buy me a beer!

                          EDIT: It does have a single PCI slot so you could put in a different video card.

                          Edit again: It has the 20-pin ATX power connector, not the 4-pin 12 volt
                          Last edited by oshunluvr; Feb 07, 2018, 04:00 PM.

                          Please Read Me

                          Comment


                            #14
                            Originally posted by oshunluvr View Post
                            LOL, we'll talk!

                            Actually I have my old server mobo. With an Intel 945GC chipset, Single Core Intel Atom 230 CPU, 2x GbE LAN, 4x SATA II, VGA port

                            This: http://www.orbitmicro.com/global/ms-...--p-10191.html

                            No HDMI or DVI - VGA only. CPU and 2GB RAM included. Dual Ethernet tho...

                            You can have it for the cost of mailing, or drive down and buy me a beer!

                            EDIT: It does have a single PCI slot so you could put in a different video card.
                            LOL the one I'm using now is (for the TV) is a bit more up to date than that ,,,dual core Pentium 3.20 GHz 6GB RAM intel graphics with HDMI 1TB HD

                            But one day I may chat ya up and drive down to by you that beer ,,,,just for that great dolphin service menu for snapshots

                            VINNY
                            i7 4core HT 8MB L3 2.9GHz
                            16GB RAM
                            Nvidia GTX 860M 4GB RAM 1152 cuda cores

                            Comment


                              #15
                              We'll definitely have to do that sooner rather than later!

                              The mobo's are;

                              MSI : https://us.msi.com/Motherboard/Z170A...M-EDITION.html
                              Asus : https://www.asus.com/us/Motherboards...-Z270G-GAMING/
                              ASRock : https://www.asrock.com/MB/Intel/Z270...eme4/index.asp

                              The MSI is a year and a half old and is a full ATX size. The other are a month or two old and mATX boards.

                              I haven't decided which I want to keep yet, but I'm physically downsizing my PC so I'm not using the MSI one for sure. The Asus has more stuff on it than the ASRock but I'm still checking them out.

                              They're all 6th and 7th gen socket 1151 boards and DDR4 RAM with overclocking capibilities. Probably overkill for a set-top box, but if you're interested in one PM me and I'll run tests on it before handing it over. I'd start at half-retail for one of the two newer boards, less for the MSI. Let me know.

                              Please Read Me

                              Comment

                              Working...
                              X