Question PC started crashing. No BSODs, it just plain crashes and Ihave to hard reset it via the button to restart it ?

androidk800

Distinguished
Aug 18, 2017
16
0
18,520
I have an old 2017 PC with Z270-GAMING K3 motherboard, recently the PC started crashing. No BSOD or anything it just plain crashes and you need to hard reset it via the button in order to restart it. This happens every couple of days or weeks, in general it happens rarely. There is no log in the system event viewer to point to anything critical, except when I hard reset the system.

I stress tested all the PC components and ran some diagnostics. The RAM is ok, the CPU is ok, the GPU is ok, the SSD is ok. So I am thinking that this is either a motherboard issue or a PSU issue, but I am drifting more towards the MB, because the PSU is a very good gold one and because the system runs ok most of the time.

Is there any configuration that I could change on the BIOS or something else that I could do on the MB to confirm it`s faulty before trying to search for a new or used one?
 
Last edited by a moderator:
When posting a thread of troubleshooting nature, it's customary to include your full system's specs. Please list the specs to your build like so:
CPU:
CPU cooler:
Motherboard:
Ram:
SSD/HDD:
GPU:
PSU:
Chassis:
OS:
Monitor:
include the age of the PSU apart from it's make and model. BIOS version for your motherboard at this moment of time.
 
When posting a thread of troubleshooting nature, it's customary to include your full system's specs. Please list the specs to your build like so:
CPU:
CPU cooler:
Motherboard:
Ram:
SSD/HDD:
GPU:
PSU:
Chassis:
OS:
Monitor:
include the age of the PSU apart from it's make and model. BIOS version for your motherboard at this moment of time.
Hi. Thanks for replying. The PSU is from around 2017 year. The other data is in the pic in short or the full below:
View: https://imgur.com/a/BaiiD6n
View: https://imgur.com/a/0VRDxon


System Information and Specs

Processor Information:


  • Model: Intel Core i5-6500 (Skylake)
  • Cores/Threads: 4 cores / 4 threads
  • Base Clock: 3.2 GHz
  • Turbo Clock: Up to 3.6 GHz
  • TDP: 65W
  • Socket: LGA 1151
  • Technology: 14nm
  • Cache:
    • L1: 4x 32KB (Data) + 4x 32KB (Instruction)
    • L2: 4x 256KB
    • L3: 6MB
  • Temperature: ~45°C to 55°C during usage
  • Supported Instructions: MMX, SSE, SSE2/3/4.1/4.2, AVX, AVX2, AES, FMA3

Motherboard and Chipset:

  • Northbridge: Intel Skylake
  • Southbridge: Intel Z270
  • PCIe: x16 (max)
  • Memory Support: DDR4, dual-channel

Memory Information:

  • Total Size: 28GB DDR4 (4 sticks, mixed brands and capacities)
  • Speeds: 2133 MHz to 2666 MHz (varies by module)
  • Timing and Voltage:
    • CAS Latency: 16
    • Voltage: 1.2V
Detailed DIMMs:

  1. 8GB TeamGroup DDR4-2666
    • Manufacturing Date: 2022 (Week 18)
    • JEDEC Timings: CL16 at 1333 MHz
  2. 4GB Crucial DDR4-2400
    • Manufacturing Date: 2016 (Week 43)
    • JEDEC Timings: CL16 at 1200 MHz
  3. 8GB TeamGroup DDR4-2666 (same as above)
  4. 8GB A-Data DDR4-2133
    • Manufacturing Date: 2016 (Week 47)
    • JEDEC Timings: CL15 at 1066 MHz

Graphics and Display:

  • Interface: PCIe x16

System Power and Voltage:

  • Package Power: ~30.93W
  • Core Voltage: ~1.16V

Software Environment:

  • OS: Windows 10 Professional 64-bit (Build 19045)
  • DirectX: Version 12

 
Last edited:
Because of symptoms, you described.

Oh. You changed your system description.
You have crazy memory configuration.

Try with 2x8GB Teamgroup ram modules in slots A2, B2.
Remove Crucial and A-Data modules for now.
Set DRAM voltage to 1.35V in BIOS.
Monitor system stability.
Okay, I have left only the two similar 8gb TG modules in the first and second slot. I saw in the BIOS that the default voltage for DRAM was 1.2 but set it to 1.35V as you said, no idea why, but I guess it would help.. For testing purposes, I also started one virtual machine and left it to run because the issue happened mostly when running virtualization in the recent days.
 
Last edited:
Thanks. I have replaced the modules on A1 and B1, but when I checked, CPU-Z is showing that the channel type is now: "Dual" in the memory tab, also the task manager shows 2666 MHz as should be the maximum speed. So I will probably not bother to open the box again.

Any suggestion what to do with the other 2 modules? My PC is better without them, or should I at least use the other 8gb module -8GB A-Data DDR4-2133? Although, it will default my entire memory setup to lower speeds..
 
Memory Information:

  • Total Size: 28GB DDR4 (4 sticks, mixed brands and capacities)
  • Speeds: 2133 MHz to 2666 MHz (varies by module)
This is hardly a recipe for stability. At the very least you should be fitting matched pairs only, with each DIMM in the pair being exactly the same part number, capacity, speed, etc. With three completely different types of DIMM, it's asking for trouble.

I saw in the BIOS that the default voltage for DRAM was 1.2 but set it to 1.35V as you said, no idea why,
When you run DDR4 at JEDEC default speed (typically 2133 or 2400MT/s) the BIOS sets the memory voltage to 1.20V. If you overclock the memory using XMP settings (held in the SPD chip on each DIMM), the BIOS increases the DDR4 memory voltage to, typically, 1.35V, to facilitate faster speeds.

In some rare instances, the default voltage at JEDEC speed is not sufficient. I have an old AMD system where I had to manually increase the DDR3 RAM voltage by 0.1V (100mV) at stock JEDEC speed, to make the system stable. The suggestion to (manually?) increase DDR4 RAM voltage from 1.20V to 1.35V might improve stability, especially if you stick to JEDEC 2133MT/s speed.

Since your system has been running OK for a number of years? with this strange mix of RAM, I suspect that RAM is possibly not the culprit and you may need to look elsewhere.

I stress tested all the PC components and run some diagnostics. The RAM is ok
General purpose stress tests do not always detect subtle faults. Did you boot the computer from a USB memory stick containing MemTest86 and wait until the tests reached the end (typically 2 to 4 hours)? If not, give it a go. If you get any errors in MemTest, retest each DIMM on its own and ditch any modules with even a single error.
https://www.memtest86.com/

the PSU is very good gold one and because the system is running ok for most of the days.
What is the part number of your PSU and who is the manufacturer? Is it mentioned in any of the Tier Lists? If your PSU is in the "Tier-F Replace immediately" category, get rid of it now. If it's above Tier-F but lower down the list than "Tier-C Low End", consider replacing it soon.
https://cultists.network/140/psu-tier-list/

The "Gold" rating is merely a measure of the PSU's claimed efficiency. Gold PSUs should be slightly more effiicient than Bronze PSUs, but that doesn't necessarily mean that Gold PSUs are better in all respects than Bronze PSUs. Some expensive Bronze PSUs may be of better quality than certain really cheap Gold PSUs. N.B. There exist levels of efficiency above Gold, including Platinum and Titanium.

Unless your Gold PSU comes with a 10-year warranty, in a circa 2017 computer, your PSU is probably at least 7-years old. If the PSU came with only a 3-year warranty, that means teh manufacturer wasn't confident enough in the design and construction to give it a longer warrant. An out-of-warranty PSU could well be reason for your computer crashing.

Electrolytic capacitors age with time and cheap PSUs tend to skimp on quality. Your PSU may contain dead or failing capacitors on the secondary rails, leading to high ripple voltages which are difficult to detect without an oscilloscope and bad news for your mobo. I do hope the main 420V DC bulk electrolytic on the primary side isn't failing. They can make a loud noise if they overheat and explode.

Do not be tempted to investigate any internal PSU faults yourself. Lethal voltages exist inside an ATX PSU. Stay safe.:)

Socket: LGA 1151

I can't open your Imgur links, so I'm not sure what motherboard you're using. I have loads of old LGA1151, 1150 and 1155 systems and for the most part the motherboards continue to work, but that doesn't mean your particular board is still OK. Parts do fail.

With any fault, I run a few tests, then consider swapping parts. With a fault like yours, I'd probably try changing things in this order:-

1). RAM - fit your pair of 8GB DIMMs, preferably in Dual Channel mode as described by other people.
2). PSU - find a suitable PSU from somewhere, e.g. another PC, a friend, etc. Consider buying new.
3). CPU - simply because it's easier to swap than the motherboard. It could be a lowly i3, just as a test.
4). Motherboard - I've bought numerous good working second-hand boards on eBay.
5). Think about a newer system.

If your LGA1151 mobo is sufficiently old to contain electrolytics llke the ones shown below, check for bulging tops and brown goo (that's the electrolyte leaking out).

capciture_failure-1.jpg
 
  • Like
Reactions: androidk800