Jump to content

Unknown Crashes


DarkSerge

Recommended Posts

Something is disturbing my system and I don't know what.

 

About a week ago, my system randomly and unexpectedly rebooted. I was sitting here and suddenly without errors or odd behaviour, the system just rebooted. Not a normal reboot with a shutdown screen - just BLINK and I'm looking at the boot screen.

 

Several days later, the system completely locked up when I attempted to open a browser. Complete freeze - even the mouse cursor wasn't moving.

 

Friends and coworkers suggested overheating issues. So all day today I've been monitoring the task manager and the system temperatures but I didn't see any red flags. Once every hour or so I jumped onto the computer just to check status. Eventually I came back and the system had restarted.

 

System logs have no errors before the reboot/crashes. There is only the error messages after the reboot telling me the system shut down improperly.

 

The only changes before these issuess were 2 Windows Updates the day before my issues started:

 

https://support.micr...pdate-kb4048955

https://support.micr...ovember-14-2017

 

Does anybody know of any issues with one of those?

 

There is a restore point before the update, I haven't restored it yet. If I do use a system restore point, should I disable Windows Update so it doesn't reinstall the update? I also have a bootable USB with Memtest86+ I will run later in the day when I have time.

Edited by DarkSerge
Link to comment
Share on other sites

You say your temps are not the problem but without knowing what your temps are, we cannot confirm that. So what are they? How are you monitoring them?

 

Note too not all critical components are monitored for temps. So is the interior free of heat trapping dust? Are all fans spinning?

 

And we know nothing about this system, not even if a notebook or PC, home made or store bought, or which version of Windows you are using.

 

Also, I always want to ensure I am supplying good, clean, stable power before getting too deep into troubleshooting hardware problems, especially odd, intermittent shutdown, reboot, or freeze problems. So I recommend swapping in a known good power supply and see what happens.

 

Last, development and bug fixes on MemTest86+ ceased long ago in 2013. This was before DDR4, the release of several generation of CPUs, and UEFI based motherboards became commonplace. Therefore, I recommend MemTest86 instead. Since PassMark took over development, it is regularly updated to remain current with the latest technologies.

 

While similar names and based on the same core code from may years ago, MemTest86 from PassMark is the better tool. That said, no software based memory tester is 100% conclusive. If it reports any errors, even just 1, the RAM is faulty. But in some cases, the RAM may test good but fail in real-world use or when paired with other RAM.

 

Be sure to let it run for several passes, or even overnight.

Link to comment
Share on other sites

I've downloaded MemTest86 from the link above. It's a Windows 10 64-bit desktop system, AMD A8-7600. I'm using a program called Speccy to monitor the temperatures. Sitting idle the CPU settles around 45-50 C and when there's activity it jump to around 55-60 C with a few spikes up to 65 C for a few seconds. The systems itself is less than 2 years old.

 

I've disabled the auto restart after a crash, though I've had BSOD issues before and so far the symptoms are different this time. No errors before the crash (checked the system log), no actual visible blue screen (I was present for one of the restarts) and last time after a BSOD on restart the system notified me of a crash.

 

The last time I had issues, it was the graphics driver. I removed everything and let Windows find it's own drivers and that fixed my previous issue. That was maybe 6 or more months ago.

 

I'm going to run the MemTest86 on it for a few hours today and see what happens.

 

Update: I let MemTest86 run for about 12 hours today and it found nothing wrong.

Edited by DarkSerge
Link to comment
Share on other sites

You could try booting with a rescue disk or a Linux system disk (or USB for that matter) and leave it on a while. That would rule out Windows software as a cause if you get a crash.

I have an A8-5600K AMD system that runs Linux all the time and it is stable.

I agree with Digerati's comments about power supply. That has been a cause of trouble for me in the past. Also a visual inspection for dust and fans running is a great idea.

Link to comment
Share on other sites

I do not have another power supply to test.

 

I wouldn't mind trying a Linux disk, but so far there is no pattern. The system lock-up was 5 days after the first random restart, then the last unwanted restart was a day and a half later. If anything I could try it next week on my nights off when I'll be able to leave the computer on all night and be home to keep and eye on it.

 

So nobody knows of any Windows Update issues since it was the last major system change before the issues started?

 

I'm taking today off from working on that computer, so I'll continue tomorrow after the holiday.

Link to comment
Share on other sites

V.T. Eric Layton

This sounds so like an intermittently failing power supply unit; a bad capacitor or overheating semi-conductor component therein. Unfortunately, intermittent problems are the very worse kind. You'd have to swap out the power supply and time test for a few days to make sure it was the culprit.

 

Regardless, I do believe this is some sort of hardware issue. Any failure (vid card, RAM, mobo, CMOS, CPU, etc.) can cause reboots. However, PSUs are the likeliest culprits.

  • Like 2
Link to comment
Share on other sites

but so far there is no pattern.
Which is a common symptom of power supply problems - that is, no rhyme or reason, or apparent cause for other/strange hardware problems is often caused by a power supply that is not outputting the necessary voltages within acceptable tolerances under all expected load conditions. You need to verify power before spending money on anything else.

 

Speccy is a great program but for some reason, it often does not report voltages correctly. For example, it is currently reporting the following for this computer:

+3.3V = 2.028 V

+5.0V = 3.367 V

+12V = 0.048 V

  •  

If any of those were true, this computer would not be running.

 

So I recommend you check HWiNFO64. The amount of information it provides can be overwhelming at first. Be sure to check the Sensors button at top. Then in the Sensor Status windows, scroll down to your motherboard to see temperature and voltages. But understand, no software based hardware monitor can measure ripple. And because they take a snapshot of sensor values, they cannot show if the power supply is regulating power properly at all loads and all times. According the ATX Form Factor standards, PSUs must maintain voltages within ±5% tolerances.

 

Acceptable Tolerances:

12VDC ±5% = 11.4 to 12.6VDC

5VDC ±5% = 4.75 to 5.25VDC

3.3VDC ±5% = 3.14 to 3.47VDC
  • Like 2
Link to comment
Share on other sites

I installed HWiNFO64. There's a lot there, I'm not sure what I'm looking for. There's info on power under sensors but a dozen values listed.

 

I'm annoyed that hardware issues is even a suggestion since I've had this system only a year and a half. I might have a friend who could have extra power supplies I can swap out. My power supply is a HuntKey HK280-25FP 180W supply. Do I need to find another 180W unit or would any supply work as long as it's 180 or above? I've only had power supply issues once before on my old system and it was already around 7 or 8 years old at the time.

 

When I have a could nights off again, I'll take the suggestion and run a Linux CD just to see if anything happens.

 

Here's a screenshot of the sensors window with the voltages. I don't know what's normal or what I'm looking for here:

 

https://drive.google.com/file/d/1CMA-DSbqjY389eQzYASbduj2YFZT4uxQ/view?usp=sharing

Edited by DarkSerge
Link to comment
Share on other sites

I have never heard of HuntKey which would suggest it certainly is not a name brand PSU maker. And 180W suggests this is not a typical desktop PC. Is this an all in one or notebook?

 

Those 122° and 114°C voltages are way too high but since they are show under the Auxiliary sensor, they may not mean anything. Your CPU and motherboard temp are fine.

Link to comment
Share on other sites

It's a desktop computer, Lenovo H50 90BG003LUS

 

I found a friend who has extra power supplies. What would be a recommended power supply wattage to replace or test with?

 

According to the Newegg power calculator, I need 434W:

 

https://drive.google...iew?usp=sharing

 

I did add a second hard drive and swapped the DVD-RW for a BluRay, but if I need that much why would it come with 180W? Is that power supply calculator accurate? When I calculate based on the computer's original setup, it says 419W

Edited by DarkSerge
Link to comment
Share on other sites

434 is way more than you need. While NewEgg is a reputable site, their calculator is notorious for overestimating needs by a large margin. All calculators pad the results because the last thing they want to do is underestimate your needs.

 

Use the eXtreme Power Supply Calculator to determine your minimum and recommended power supply unit (PSU) requirements. While it too adds a buffer, it is much more conservative and realistic. I recommend setting Computer Utilization to 16 hours per day and CPU Utilization to 100%. These steps adjust for capacitor aging and ensure the supply has adequate head room for stress free (and perhaps quieter) operation. These steps also add a little extra buffer for unplanned future upgrades or added hardware demands.

 

It is always okay to buy a PSU that is bigger than you need (except maybe for the budget) but always bad to buy too small. But buying way too big is not good either because power supplies tend to be the most efficient when operating at 50 to 60% load. So the eXtreme PSU calculator is, by far, the most conservative (a good thing) when it comes to padding results, and that is due, in part, because it has the most flexibility. It is also because the developers wish the eXtreme calculator to be the most accurate. And it is.

 

When I insert your specs, add 3 x 120 case fans, increase CPU utilization to 100% and Computer utilization to 16 hours per day, as seen here, I get a recommended 252W. And without a discrete graphics card, that makes more sense. So a decent 300W supply would suit your needs just fine.

 

That said, a 400 - 450W quality 80 PLUS supply will meet your needs and carry you into the future should you decide to add more drives or perhaps a graphics card too.

 

My problem, in terms of recommending something, is 180W is not typical for a "PC". That could indicate Lenovo has gone proprietary instead of with an ATX Form Factor standard power supply and motherboard. :(

 

Are you sure that supply you have now is a 180W supply? Or does it perhaps, have 180W capacity on the +12V rail only?

 

Friends and coworkers suggested overheating issues. So all day today I've been monitoring the task manager and the system temperatures
Where are you looking? As far as I know, Task Manager does not and never has monitored or displayed "temperature" values.
Link to comment
Share on other sites

Where are you looking? As far as I know, Task Manager does not and never has monitored or displayed "temperature" values.

 

I was using a program called Speccy to watch my temperatures.

 

I have a friend who has extra power supplies, he said he'd even lend me 2 of them to swap. I opened up my computer and took a picture of the label on the side of the power supply and in the tiny writing it says the output is 180W (which matches the specs when I Google search my PC brand and model and find specification list.) He also thinks it's my power supply when I explain the issues to him.

 

Here's the label of my current power supply:

https://drive.google.com/file/d/1J3h5YN_p6_ki5FwwlPCyHmBY8BTFCxbW/view?usp=sharing

 

One thing that kind of bugs me. Why would Lenovo not put in an adequate power supply? It sounds like there should have been a better PSU from the beginning. I did a search and didn't come up with any information about this model of Lenovo desktop having power issues, but if they all ship with inadequate power then why is there not more info? This isn't a custom build from someone who made a miscalculation, it's a name brand system.

Link to comment
Share on other sites

Looks like Lenovo has a 14 pin power connector instead of the standard 24 pin one. Go figure.

This means you must:

  • get a proprietary PSU
  • rewire the connector plug
  • get an adapter to go from 24 to 14 pins.

Annoying but what can you do?

Link to comment
Share on other sites

One thing that kind of bugs me. Why would Lenovo not put in an adequate power supply?
Money.

 

Note however, it is adequate, but just barely.

 

If you look at your image carefully, you will see it is the +12VDC output that is rated at 174W. That is enough for relatively short bursts (which is how most users use their computer). But I would worry about excessive heat if you tasked it maxed out for extended periods - say while "Folding" or searching for ET (SETI).

Link to comment
Share on other sites

Great.... I knew something was weird about it when the power for the drives was a connection to the motherboard and not the PSU itself... First desktop I buy in a decade and it has to have an abnormal configuration...

 

How about this:

https://www.newegg.com/Product/Product.aspx?Item=9SIA2RP5F16442&cm_re=24p_lenovo-_-9SIA2RP5F16442-_-Product

Edited by DarkSerge
Link to comment
Share on other sites

I ordered the cable, and a friend has a couple power supplies he can give me. So by the end of the week I should have both. Until then, I guess it's just wait and see what happens. I haven't had any restarts or issues in about 5 days.

 

Update: Spoke too soon. About an hour after this post the system restarted again. Same symptoms: no errors in the log leading up to it, no BSOD either. I disabled restart on BSOD so I know I'm not getting one.

 

Another Update: The Lenovo Solution Center software ran its monthly hardware scan and it came up with failures on my C: drive. Targeted Read Test, SMART Short Self Test, SMART Drive Self Test, all came up as failed. Could this be related?

Edited by DarkSerge
Link to comment
Share on other sites

Speccy reports "Good" on all fields on that hard drive, though the status is "Warning"

 

https://drive.google.com/file/d/1q3jCWbmNt4GkAP7fscoSpEJERZK-9q1W/view?usp=sharing

 

 

I also have HWiNFO64 if there's anything I can check there. Any recommended drive diagnostic software I could give it a scan with?

Edited by DarkSerge
Link to comment
Share on other sites

Again, what does HWiNFO show for your voltages?

 

Compared to the screenshot I shared last time, they're roughly the same, give or take a few small fractions of difference. The only major difference is the "VIN2" was about 1.2v higher when I first compared, but it's slowly dropped to close to what it was when I first shared a screenshot.

 

I'll run more scans from SeaTools on my hard drive today.

 

Update: "Short Drive Self Test" and "Long Generic" tests in SeaTools fail on the C: drive.

Edited by DarkSerge
Link to comment
Share on other sites

Sorry, you did indeed show a shot with your voltages. But sadly, that shot does not show everything needed. I suspect that is because it is really a notebook motherboard in a PC case. :(

Link to comment
Share on other sites

So are the failed hard drive tests not a problem?

 

Is another another set of values I should be showing? Sometime next week I should be able to put in a better power supply. I have a friend at work with spares and a cable adapter ordered.

 

On another note, I was off work last night and let my system sit on all night with a Linux Live CD and had no issues. May not mean anything though since I've had up to 5 days without an unexpected restart.

Edited by DarkSerge
Link to comment
Share on other sites

So are the failed hard drive tests not a problem?
I would never say that. The problem with errors reported in drive status reports (when no other read/write errors occur) is they merely suggest potential or "impending" errors. "Impending" does not automatically imply something will happen and that it will happen soon. The drive could fail tomorrow, or provide several years more of service.

 

But that does not mean you should start making backups now. Why? Because everybody should already have a via backup plan and current backups because catastrophic errors and data loss can happen at any time and without any warning.

 

IMO, everything at this point is just guessing and speculation - until you verify you are supplying your system (and the drives) good, clean, stable power.

Link to comment
Share on other sites

V.T. Eric Layton

You cannot troubleshoot hardware issues (particularly intermittent ones) with software testing exclusively. Software tests will consistently show PSUs to be producing voltages in the proper ranges, yet that one instant when a component overheats or breaks down for whatever reason will never be seen by software tests. Unfortunately, the only way to resolve this issue is to install another PSU and time test the system. If no more random failures occur, then you can pretty much assume the original PSU is ca-ca and can be disposed of properly.

  • Like 1
Link to comment
Share on other sites

Unfortunately, the only way to resolve this issue is to install another PSU and time test the system.
Right. Because short of swapping in a known good PSU, the only way to conclusively test a PSU is with the supply put under a full range of expected loads while monitoring over time with an oscilloscope, or with a dedicated power supply analyzer - sophisticated and expensive test equipment that requires training to use and expertise to properly analyze and understand the results.

 

The typical multimeter is not good enough. For one, there's those "full range of expected loads" and "over time" issues. And another reason is most multimeters are incapable of measuring ripple and other anomalies that affect computer stability.

 

If the monitoring software (or multimeter) shows a voltage is out of the allowed ±5% tolerance range, the PSU is most likely bad. But as Eric correctly noted, such monitoring looks at a specific instant in time. If the fault is not seen in that instant, a problem PSU may not be detected. So we are back again to where the best course of action for the normal user (because EVERYTHING inside the computer depends on good, clean, stable power) is to swap in a known good PSU and see what happens.

  • Like 1
Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
×
  • Create New...