Solved Recent rash of BSODs

I've recently been suffering a rash of BSODs. I think I have calmed things down by doing 2 things:

First, sfc /scannow identified a corrupt module it couldn't repair. It eventually turned out that my Component Store contained some corruption. Repairing that allowed sfc to correct the corrupt module. Things pretty much calmed down.

But the second part of the story involves File Explorer. It was crashing often. A web search turned up the advice to clear its history. Doing that seemed to bring things back to what I had hoped was complete stability.

For the past few days while these BSODs have been plaguing me, I have not been able to identify any particular usage pattern on my part that was triggering them. They seemed to just be random & maximally inconvenient. Given what sfc discovered, I suppose this wasn't directly related to anything I was doing. If you have a corrupt system component, your symptoms will seem totally illogical. Completing the above 2 tasks did appear to restore what looked like complete stability.

However, after a promising start, I just got a BSOD. A kind user over in this thread offered the advice that has led me to post here:


That's the first of a short sequence of posts that give some background.

I think it will be extraordinarily kind of anybody who looks at this collection of dumps:


Following various instructions, I have copy/pasted the information from Settings -> System -> About into the field below. More information on my system is included in my elevenforum profile. I have not overclocked anything on this system. I don't know enough about that subject. I have looked at a lot of things in my BIOS, & I just shook my head in uncomprehending wonder & left everything at default. The date quoted below for when this was installed appears to be the date of installation of the most recent major system maintenance from Microsoft. The true inauguration date of this system is December 3 or 4, 2023. This is a brand new system, new hardware that I put together myself, on which I installed W11. This was not an upgrade or migration. I ditched my old W7 computer, built this new one, & did a clean install of W11 onto this virgin machine.

The dump collector appears to have reported on all the BSODs I have had in the past few days. I don't know how to even begin interpreting the stuff that's in there. I did unzip my report & tried to look through it for anything that would have given me the error code on my most recent BSOD from within the past little while. But it's such an overwhelming volume of stuff that I didn't learn anything. I imagine it takes some specialized skill gained by some intensive study of . . . I don't even know what . . . to be able to make any sense of the file I've posted on Google Drive. Which makes me even more appreciative of whoever looks at it.

Fortunately, my system seems to have returned very close to its former stability so there isn't much urgency here for me to get an answer. Frankly, my expectations are low that anybody will give me an answer. There's so many BSOD reports here. So if I ever get an answer, I will be pleased beyond expectations. I thank my unknown benefactor now, whoever you turn out to be.

Please perform the following steps:


1) Run Tuneup plus > post a share link





2) The computer has 4 RAM modules, 32 GB each = 128 GB

Remove all RAM modules and label them.

Test one RAM at a time in the same DIMM for 8 or more passes > take pictures > post images or share links

Just 1 error is a test fail and testing can be aborted.

Thank you for paying attention to my problem. I wish I hadn't had to create yet another membership on yet another forum. I had to join tenforum in order to download the .bat file. Oh well. Add that to sevenforums & here. I didn't fill out my profile because that information is in my profile here. Plus I'm not running W10. Never have.

I've run Tuneup. The log file is small so I'm attaching it here. I read the .bat file before I ran it. It's about 2 levels of complexity above my skill level but I got the gist of it. It runs sfc /scannow & then it runs 2 dism commands. As it turns out, I did all of that yesterday. On my own initiative, I ran sfc & it did find one corrupt DLL that it was unable to correct. So I followed the instructions in the tutorial on here for correcting my Component Store. Apparently, the DLL in question was corrupt in the Component Store. So Tuneup just repeated what I already did yesterday. But I get it. You're being thorough. I appreciate that. The only steps I had not done were a chkdsk & an analysis of my pagefile. The only error chkdsk found was in a Windows.old directory. I had tried several repair installs before I came to the conclusion I needed to try sfc. I kind of expected to have several Windows.old directories because I did multiple repair installs. Why multiple? Because several of them failed & 2 or 3 of them succeeded. But none of them corrected my BSOD problem. So I thought I would have as many Windows.old directories as successful repair installs. But it turns out I have only 1. I'm not sure the thing that got corrected by chkdsk was of any significance. I'm going to delete Windows.old at some point in the next week or two. As far as I can tell, there were no issues with my pagefile.

It looks like you changed a key in my registry. It was AutoReboot=TRUE but you changed it to FALSE, as far as I can tell. What effect does that have? When I have experienced my BSODs in the past few days, they have all automatically done reboots after doing the data collection they say they are doing. I assume those were the dumps I posted earlier. I hope you didn't change that behavior. I would request that you give me complete instructions for what I should change back in my registry if you did change this behavior. Otherwise, I would like to know what that key does & why you changed it.

For the memory test, that is going to take a while longer. It's not easy for me to get into my system case with the geography of where my computer is. In order to get into the case, I have to disconnect all the cables (monitor, network, keyboard, etc.) so I can move the box into the middle of the room where I can open it & get at the RAM. So I'll have to disconnect, reconfigure the DIMMs, reconnect, rinse/repeat, at least 4 times. Pulling out all the DIMMs & booting up with just one DIMM at a time is probably going to be an operation taking several hours. Oh well. I understand its necessity. I will do it. Just not right away.

I assume that is what you want me to do. You want me to boot memtest with just one of my DIMMs, each one in turn. Then run memtest on each one. Is that correct? Is memtest not capable of detecting whatever it's going to detect with all 4 DIMMs installed? Also, my motherboard manual says that with a single DIMM, it needs to go in slot A2. So I will not be able to check the other 3 DIMM slots. If there's a fault in those other slots, we're not going to detect it like this. That's more of a comment to see if I understand things correctly. I'm sure this is something you already know well.

Well, I finally read some more of the instructions. I'm going to need to run 16 tests, 4 DIMMs in each of 4 slots. Despite what it says in the motherboard manual, I can run a single DIMM in any slot, right? Each test is something that is best left to run overnight. I will not have results for this very quickly. I guess I'll post results for each DIMM/slot combination as I get each one. But it's going to take a while. At least I won't need to connect anything more than the monitor & keyboard while I run one of these test.

It occurs to me that I can start by just testing with all 4 DIMMs in place. If that passes 8 cycles of testing, then I should be good. It's only if memtest detects any errors that I'll have to go to testing each DIMM in each slot one at a time. Does that sound acceptable?

And you say you want pictures. I guess you're saying I need to take a picture of my monitor showing the memtest results. There's no screenshot function, is there? I'm good at screenshots. Photography is not my strong suit. I'll do what I have to.

Tuneup plus results were good.



The older versions of Memtest86+ were slower and the downtime was ok with 8 and 16 GB RAM.

With 128 GB RAM downtime may be several days to run 8 or more passes.

The more the passes the better the testing conditions > it reduces the false negatives (a test pass with malfunctioning RAM)


During the troubleshooting BSOD can be missed using automatic restart.

Turning off automatic restart reduces the missed BSODs.

After the troubleshooting has completed you can apply any preferences.



For Memtest86+ results please take pictures with a smartphone camera or digital camera since it run outside Windows.


If you choose to test a four RAM simultaneously and there are sufficient passes with no errors then that will conclude the use of the software.

If there is just 1 error you can abort testing. Then the RAM modules can be tested in pairs and then one at a time to identify the malfunctioning module.

In general testing one at a time becomes more efficient.

Several days? You will not be surprised when I tell you that isn't going to happen. This isn't a spare computer that I can leave in a corner running a RAM test while everything else in my life goes on as normal. This is my only computer. (My phone doesn't count.) I cannot be without it for a week. OK. It will have to be the 4x4 tests over a span of I have no idea how long. I assume overnight should be good enough to test one 32G DIMM in one slot.

Was there something specific in my dumps that made you suspect my RAM? Or is this a standard procedure to cover all the bases? Given the dramatic improvement in the state of affairs as a result of getting sfc to fix that one corrupt DLL, I'd say this has been a software problem all along. I would be surprised if it turns out there's anything wrong with my RAM.

OK. So my BSODs (may they be few & far between now) will not automatically reboot. The BSOD will sit there staring at me until I . . . do what? Hit the power button? I don't have a reset button on my system case. Or is there some magic key sequence that clears a blue screen? Ctrl+Alt+Del? Escape? Enter? Something else?

Once this is done, I would very much like to revert the setting so BSODs just automatically reboot after the data collection. I don't entirely trust my expertise to have gotten the registry key right from the tuneup log. Would you be so kind as to explicitly specify what key I'm supposed to revert to what value? I'm comfortable using regedit. I have a feeling this isn't something that is available in the regular GUI Settings dialog. I searched Settings with BSOD, errors, blue screen errors, it didn't find anything. I also looked in a couple of the sections of Settings for anything that might be related but I didn't find anything.

Speaking of Settings, have you noticed that every once in a while, not a lot, just maybe 2 or 3 times a week, Settings will spontaneously close while you're looking at something? Reopening it & trying the same thing again pretty much always works fine.

And I have scheduled a chkdsk /f on my boot partition for my next reboot. Since Tuneup did detect that one error, I may as well be safe & do it. I've done it before & it runs way faster than my old rotating boot drive on my old systems (W7 & its ancestors). Now I'm booting from an NVMe SSD & it runs in almost no time.

See this tutorial to make preferences with automatic restart:




All software RAM testing may be able to be skipped if automatic restart is kept off by using only one RAM module at a time.

Later software RAM testing can be used to double check any prior findings.


How long had the 4 RAM modules been installed where there was computer stability?


The BSOD and live kernel bugchecks were:

I truly appreciate your help. But I don't feel like I'm moving towards an answer. I am convinced this is a software error. My File Explorer keeps crashing. My Firefox keeps crashing. Those seem to be the only pieces of software that are overtly having trouble. In the case of File Explorer, it was occasionally packing it in before these BSODs started. My online searches turn up quite a lot of similar problem reports. In the case of Firefox, that is new behavior this week. I suppose I could reinstall it. I don't think that's going to help but I might do it anyway. However, I have had BSODs when I didn't even have Firefox running. I had neither File Explorer nor Firefox open earlier today, as I'm about to explain.

Thank you for the pointer to how to restore the automatic reboot from a BSOD. Here's an example of why I need it. Earlier today, I went to a theater to see a satellite transmission of today's live opera from the Metropolitan Opera. This is a regular practice for me. In case you don't know, the operas are on the radio every Saturday but only a few go into theaters each season. When I do attend one of these shows, I like to record the radio broadcast on my computer. The radio commentators & intermission features are not entirely the same on the radio & in the theater, so I like to catch those bits from the radio after the fact. I record the radio broadcast livestream off my local NPR affiliate using ffmpeg. I have concocted a script for doing this. Despite my expectation that it would fail, I did launch my recording script before I left for the theater. When I got home nearly 5 hours later, the fans in my system were blowing up a hurricane. This was because the system was sitting on a blue screen. As far as I can tell, the only way out of a blue screen is to power off & power back on. I have had a bad experience with just powering off. It caused me to do a repair install on W7 a few years ago, an experience that was horrible. Repair install on W11 appears to be hugely improved. (Plus I have a faster machine now.) Still, I really don't like just hitting the power button. In addition, once I did get back to my desktop, it appears the system was sitting here doing its best to dissipate heat for about 3 & 1/2 hours. This is not acceptable. It's not good for the hardware. So I'm back to automatic rebooting from BSOD. Nothing will be lost. A dump happens with each one. So I've reverted that setting.

And just by the way, my recording of the opera was useless. It claimed to have recorded for about 45 minutes but there was really only about a minute there. The abrupt interrupt apparently kept ffmpeg from correctly writing what little it had tried to do. Seems a bit surprising, but that's what happened.

I also looked up those bugcheck codes:


I must say some of them are just gibberish to me. Some of them sound like they may be related to programming errors. I did chase down this:


It told me to run the Get Help app, which I did. I answered the questions the app posed, & it ended by apologizing for not helping me. Thank you so very not at all.

So I am extremely appreciative for your help. But I was hoping you could tell me which application or applications is or are causing at least some of these errors. I accept that I do need to eliminate my RAM as a source of problems. This RAM went into service as brand new hardware at the beginning of this past December, not quite 5 months ago. I would be astonished if this RAM is faulty, but I do accept that I need to prove it's not faulty.

But I remain convinced this is a software issue. It should not be necessary to uninstall & reinstall applications one at a time. That's just guessing. These dumps, at least some of them, should be telling us that a certain application or a few is/are doing something wrong.

We've verified twice that I have no corrupt system modules. We've run chkdsk twice on my boot partition. I just now ran another chkdsk scan & it said there was nothing wrong. With 132G of RAM, I should expect my pagefile doesn't get much activity, but we've verified it's OK. I have scrupulously kept current with Microsoft maintenance. After the repair installs I've done in the past few days, I have spent the not insignificant time it takes to apply all the available maintenance . . . again, more than once within the past week. These are all things I've read you need to make sure you do.

I repeat, I appreciate the effort you have expended on my behalf. But you're doing this out of the goodness of your heart. I would be quite willing to read something that would give me instructions how to interpret these dumps. I am a programmer with plenty of experience. That's all on IBM mainframes. But I think I'm quite capable of understanding documentation that would be relevant here. Please point me to something that would be educational for me. I don't expect you to solve my problems. That is my job. But I would love to get some help. Not you doing it for me. Help. My work. Your help.

Please post a new V2 share link.

Then reboot the computer into safe mode, safe mode with networking, or safe mode with command prompt.
(shift + restart)


Run safe mode continuously including overnight.

The longer the continuous run in safe mode the better.

For any / all unexpected shutdowns and restarts including BSOD post new V2 share links.

Are you an opera lover too? I check the Met's calendar pretty much every day.

Here's the latest batch of dumps:


I'm going to run CCleaner now to clean up these dumps. I think this file I've just uploaded includes dumps you've already looked at. So I may as well get rid of them so if I upload any more, they won't be repeats.

This whole exercise has been quite stressful & exhausting. I took a bit of a break. I also did what I threatened to do earlier. I completely erased Firefox & installed it over again. A bit of a hassle, helped greatly by Mozilla Synch. Since I did that, I have experienced fairly long periods of useful service from my system. At the moment, it's been nearly 5 hours since my last reboot. That's easily & by a large margin the best I've had in a week. This tells me that this problem is most likely a software problem, which has been my best guess from the start.

But things still blue screen on occasion & freeze solid occasionally. The blue screens like to happen in rapid succession. I'll reboot & before I've had the chance to properly start using the system, it will blue screen again. It seems like now, with a reinstalled Firefox, things are fine after the second blue screen . . . until the next mishap occurs. I am finding that the most common next mishap seems to be a total freeze of the system. It stops responding to my keyboard & mouse and the clock on the Start Bar stops ticking over. I have seen this as a prelude to a blue screen. But it also happens as a permanent freeze with the only out being the power button.

My next candidate for replacement is Macrium Reflect. It's something that has its hooks deep into the file system. I'm going to remove it & leave it off the system for a couple of days before I reinstall it. I wasn't planning on doing any backups in the next few days anyway, so it's no hardship to dump it.

At this point, the only thing that is overtly failing on me is File Explorer. Spontaneously, when File Explorer is in background & has been for quite a while, my screen flickers & I see all the tasks on my Task Bar get scrambled. At the same time, File Explorer simply disappears from the active tasks. I can reopen it & open whatever tabs I had open prior to its crash. But some unpredictable time later, it crashes again, with more scrambling on the Task Bar. However, as I mentioned before, I seem to be far from alone with this problem. I am seriously considering investigating alternative file managers. I already have Disk Genius, but you can't just directly launch an application by its file association by double clicking a file of the right type, something we take for granted in File Explorer. So I need something else a little less fancy than Disk Genius. I have been running a Notepad replacement for so long I don't even remember now all the little annoying things Notepad did badly. I ran Metapad for several years on W7. Then it started doing something strange & somebody recommended Notepad++. I've been running that for 2 or 3 years now, first on W7 & now on W11. It is really quite nice. So convincing me to replace File Explorer would be quite easy.

Every once in a while I will see Task Manager or Resource Monitor slide down to the end of the Task Bar from their normal assigned positions as 1 & 3 from the left. That's normal assigned positions as defined by the way I choose to use them, not any rule defined by Microsoft. I don't know what causes this. They don't actually crash. I think this happens when they suffer a Not Responding condition, then return from it. Returning from Not Responding seems to want to move them to the end of the Task Bar. I just drag them back to where they belong & things run fine after that.

I really would like a recommendation from you for a reference that teaches how to read crash dumps. I could do a web search & that would turn up thousands of results. I would then just have to plow through them looking for one that seems to be helpful. But you're an expert in these things. Surely you can recommend one for me that would shorten my search. I don't mind spending time, even a lot of time, reading reference & tutorial material. But I'd like to start with something that I would hope is useful without my having to read untold amounts of stuff that ends up being not useful. So please, I ask again, tell me about something that would teach me how to analyze crash dumps.

On the subject of testing my RAM, something I may never get around to doing, I think I could shorten things a bit. I don't need to run 16 tests. I can test with a single DIMM in each of the 4 slots. If those all pass, that proves the slots are not faulty. I could then test each of the other 3 DIMMs in a single slot, thus reducing the number of tests from 16 to 7. Does this sound reasonable to you?

Sign in​

to continue to Google Drive


Make sure that links do not require sign in


There are a lot of trial and error steps that you can perform.

Please complete all of the software uninstallations and reinstallations.

Then comment into this thread if you're able or not able to perform the next steps.



If possible

1) please plan to run the computer in safe mode with networking for as long as possible (one to three days) (continuously) > for any BSOD post a new V2 share link

2) plan to run the computer with one RAM module at a time in the same DIMM.

a) With just one RAM module if there is no computer instability for a week then it's likely that the RAM module and DIMM are not malfunctioning

b) Ditto for each of the remaining RAM modules

3) Plan software confirmation testing as needed to be performed overnight and day hours that do not conflict with opera activities
(just 1 error is a test fail and testing can be aborted)

4) Unless there is a problem where free space is reduced to < 30 GB please do not use disk cleanup or third party cleaning software during the troubleshooting (allow windows error reporting, dump files, event viewer, dxdiag, msinfo32 data, etc. to grow during the troubleshooting)(allow all V2 share links to remain available in the thread)

Oops. Sorry about that. Try this:

The 03/25/2024 collected BSOD bugchecks were:

4E
7E
50
1A
21



The other BSOD and live kernel bugchecks were:

139
DE
EF
1E
7F
A
133
3B
124

I had appointments & errands to take care of today so I was not home for several hours. Counting from when I went to bed last night & the time I was out, I had my system booted into safe mode for over 12 hours. No, I did not do it with only 1 DIMM installed. More on that below. I had all 4 DIMMs installed, the full 128G. The system stayed up the whole time. I left one task running. It was a command prompt executing this command:

Timeout /T 99999

The Task Bar in safe mode doesn't have a clock on the right end of it. As I've mentioned, my system has been throwing in a few freezes scattered among the blue screens. I wanted to have some way to get an idea of when a freeze might have occurred. The counter displayed by the Timeout command would have frozen at some point & I could have just subtracted whatever it had frozen at from 99999 & I would have had a rough estimate of when the freeze happened. If it had crashed, it would have booted back into normal mode, & I do have a way of knowing when that might have happened. Of course, if it had gone to blue screen multiple times, I wouldn't have known. I would have known only that at least one blue screen crash had occurred.

But that was all unnecessary because it never froze & it never did an automatic reboot after a blue screen. It just sat here idling, counting down the Timeout command timer.

What does this tell us? Anything? I have some speculations but you'll have to tell me whether they are even plausible. I gather that safe mode puts the system into some kind of semi-disabled state. Certain drivers are not loaded. My conclusion from this failure to fail is that I need to reinstall some drivers. But which ones? This is the list of drivers I can think of that I have had to install onto this system myself:

Intel video driver
Intel audio driver
Intel networking driver
NVidia video/audio driver

This reflects the hardware I have. My motherboard has an Intel chipset. There are video, audio, & networking adapters built into the motherboard. I also have an NVidia video adapter. The details are all in the My Computer profile you can click on in any of my posts.

So is it a reasonable stab in the dark for me to reinstall these drivers?

I'm going to rant a bit here. This rant is not so much directed at you as it is at Microsoft. I must say that I am quite disappointed that these famous crash dumps don't clearly state which piece of hardware suffered the alleged fatal error. All we are told is, "A fatal hardware error has occurred." Could that possibly be any LESS helpful? Do they at least tell us what program was issuing the request that caused the hardware error to be encountered? That would be something a user could take action on. No. We get the vaguest & most useless error reports. Not only that, it takes some skill that looks like magic to even extract the dumps, never mind interpret them. If I had written software like that when I was a programmer, I would have been fired. I get the feeling they write code to feed their egos & to ensure their job security. They are the only ones who know anything about this sort of thing so they are indispensable to Microsoft. This situation is reprehensible. You should have been able to look at my first dump & tell me exactly where the problem was.

By the way, is a fatal hardware error distinguishable from a software driver bug?

I am curious why you won't tell me where to get some education. I've asked you more than once to point me to a reputable, reliable document that would let me figure some of this out on my own. Why so secretive?

I've never booted into safe mode before. I never did it in W3, NT, W98, W2K, WXP, W7. This is the first time I've ever done it. I thought maybe I could just run my normal activities in safe mode for a while & see what happens. But there were a number of things that didn't work.

The Task Bar doesn't have the clock on it like it does in a normal boot.

The Start Menu doesn't work properly. Normally, you can hit the Win key, then DownArrow, then navigate around your pinned apps with the Arrow keys & Page keys. In safe mode, that doesn't work. You have to click on an icon with your mouse to open that app. What does someone with no mouse do?

The command prompt is automatically elevated. Plus it doesn't execute simple .bat files. I tried to run a couple of my .bat files, meaning scripts I have written that work great when I'm booted up normally. Instead of just running, I got some weird PowerShell command that failed & a File Explorer window opened. I did verify that File Explorer did recognize all my partitions on my various HDDs.

I tried to activate my second monitor. But the detect function in Settings failed to detect it. It sure as shooting is detected when I normal boot.

I opened a PDF. I was listening to tonight's opera in the XM app on my phone, Bluetoothed to my wireless headphones. Quite good sound. Infintely better than trying to listen on the phone's internal speaker. I've had a lifetime subscription to XM for over 10 years, mostly because MetOpera Radio is in the package. I had downloaded the playbill program for La Rondine from the Met's web site Monday. It's a PDF, as all their programs are. Acrobat didn't work properly. Normally, you can scroll pages using the Arrow & Page keys. In safe mode, they don't work. You are forced to resort to mouse interaction, something I try to avoid as much as possible. Annoying.

Firefox would open but I couldn't visit any web sites. But, you say, you were running safe mode with networking, weren't you? Yes, that is what I did. You'll have to trust me on this one. I selected #5 from the relevant bootup menu. But I got an error message. I tried to make a screenshot of it but the PrintScreen key didn't do anything. It's supposed to activate the Snipping Tool, like it does in normal boot, but it did nothing. Fortunately, my image processing program (FastStone Image Viewer) did work. It includes a screen grabber. Here's what I grabbed:



This fits in with the contortions I had to go through to get W11 installed back in late November, early December. It seems that despite the fact that I have a motherboard with an Intel chipset, a model of board that was about a year old by the time I was trying to install W11, the 23H2 install media did not have a driver for the network adpater on this motherboard. I believe this error in safe mode is just another manifestation of that. The upshot is that I did my best to boot into safe mode with networking, but I was thwarted by circumstances beyond my control. This network adapter works fine . . . in normal boot. I'm here posting because it works fine.

From where I sit, safe mode is not very useful. Either that or my expectations are just too high.

One of the errands I was running today was to purchase a 128G thumb drive. You see, I have not totally rejected your idea of testing my RAM. I think it is a sensible thing to do & I will do it. And now, I can do it, while I couldn't before. So I'm not going to waste my time on safe mode one DIMM at a time. I'm going to do a proper test.

Speaking of that, I still believe my RAM is fine. So, being optimistic, I expect to do only 4 tests with Memtest86+. I will test each DIMM in turn in the slot in which it normally resides. I am betting each test will approve the DIMM I test. That will have the side effect of also proving there's nothing wrong with any of the 4 DIMM sockets.

We've verified with sfc that my Microsoft programs are correct. We've run chkdsk on my boot partition & it's clean. I also, by the way, ran chkdsk on all my other partitions & they are all OK. I'm going to test my RAM. But that's not all my hardware. How do I test my CPU? How do I test the rest of the chipset? I suppose the facts that I have 2 working monitors, that I get sound out of my receiver, & that I'm here online posting this, combine to prove that those parts of my motherboard are beyond suspicion. But what about the rest? How can we even test that? What other hardware is there to test that can even be tested? Plus, I ask again, is a hardware failure distinguishable from a software failure in a hardware driver?

There are many steps that could be performed.

Safe mode testing is useful as you believe there is a software problem.

Another method is making free or pay backup images.

Then hardware testing can start with RAM testing.

You can also probably consider an in-place repair upgrade if it's not a hardware problem:

I have been quietly working on this without making daily posts. Here is what I've done.

First, a bit of terminology. My motherboard manual labels the DIMM sockets A1, A2, B1, B2 starting closest to the CPU & counting away. I imagine lots of motherboards use the same terminology. The manual also says that if you're running only a single DIMM, it should go in socket A2. For 2 DIMMs, they should be in sockets A2 & B2. They don't show a possibility of running 3 DIMMs. I might have experimented with that but it turned out I didn't go there.

I suppose we all know that RAM is packaged & sold in pairs. So let's call my first package of DIMMs D1 & D2, my second package D3 & D4. Making things easy to keep straight, each DIMM has a serial number printed on it.

I have read that you're supposed to use matched sets of DIMMs. In my case, following the recommendations, my RAM has been installed like this: A1|D1, A2|D3, B1|D2, B2|D4.

I started testing with A1|D1. During the first pass, Memtest86+ reported 5 failures scattered across 3 of the 11 tests it executes. Despite the errors, I did let the pass run to completion. There was no use bothering with any further passes. The one pass I ran took about 34 minutes.

After I killed that test, I ran A1|D2. It had 0 failures during pass 1, which also took about 34 minutes. Sadly, pass 2 reported 7 failures across 2 of the tests. So I stopped that test run there. I did let pass 2 complete. Strangely, pass 2 took about 90 minutes. I read the documentation on the Memtest86+ web site & I didn't see anything that would explain why I observed the second pass taking 3 times as long as the first pass.

At that point, I didn't want to be offline any longer. So I installed only A2|D3 & A4|D4. Were they good? Were they bad? I hadn't tested them yet. But it's what I had & it was the only way I could get my computer back up. I had no illusions that things were cured. Testing these last 2 DIMMs was still necessary. Plus, I need to get a DIMM, either of the remaining 2, to run 8 clean Memtest86+ passes in a socket for me to say the socket was fine as well as the DIMM. After 2 failing DIMMs in socket A1, I had no conclusion to draw about the functioning of socket A1. At that point, I was facing at least 5 more tests. I had to find a DIMM that would pass 8 tests in socket A1. That would prove socket A1 was fine. But I now also would have a known good DIMM. Any failures of that DIMM in the other 3 sockets would point the finger at the socket. So after a first successful test, I would need to run that DIMM in the other 3 sockets, plus the last DIMM in any socket. Of course, further failures would change the picture, but I didn't want to get ahead of myself.

You might say just run with your 2 DIMMs & see how that goes. No. That's not a proper test. I've been able to run this system as long as 12 hours before getting a blue screen. Did that prove my RAM was good? The answer is obvious. So I did start booting with A2|D3 & A4|D4 & hoped for the best.

Speaking of hoping for the best, I had been running with at least 2 faulty DIMMs for about 2 weeks. I first encountered problems when Macrium Reflect wouldn't verify backups it had just taken. This was behavior that I had not observed before. I initially thought Reflect had caused the problems I had been seeing. But after discovering that my RAM was faulty, I suppose Reflect was just the first application I ran that was sensitive to the faults in my hardware. After that, everything just went down the toilet. I don't believe I have any corrupt files. I haven't encountered any, which I consider fortunate in the extreme. That doesn't mean I don't have any. Plus, I did not trust this system enough at this point to run any backups. So I was taking a lot of chances. I had done only some of my backups when these failures started occurring. The bulk of my system at that point was not backed up, which made me very uncomfortable.

It appears, then, that these 2 faulty DIMMs went bad while I was running my backups. Kind of coincidental. Or maybe they had been bad for a while but Reflect was the first victim. On the other hand, I had no blue screens before that day when I did manage to get Reflect to verify a couple of backups of smaller amounts of data. The first failed backup was on a partition of a substantial size. And the first blue screen happened after that, with ever increasing frequency. I originally blamed Reflect for this mess. I see now that I should have taken the hint & started investigating a hardware problem sooner.

So why don't I just return the faulty RAM? Micro Center has a 30-day free return policy. I bought this RAM in mid-October. I am, as they say, SOL. Or maybe not. The original packaging says there's a "limited lifetime warranty." But I won't be checking into that. The warranty means just that they'll replace the RAM with more of the same. I don't want more of the same. It failed after a bit more than 3 months in service. Plus, I don't want more RAM with LEDs. When I bought these, I had no idea that RAM could even have lights. I never even conceived of such a possibility. I have since read that gamers are deep into the bling in the their systems. My system sits on the floor beside my desk & I mostly can't see inside the box. Such frivolity is simply wasted on me. I also have a probably impossible to prove hypothesis that Corsair wasted a lot of effort on their app to make the LEDs practically sing, something you will be not at all surprised to learn I never even considered investigating. I think Corsair put effort into the bling that should have gone into quality control. So I'm done with Corsair. I will be replacing this RAM with something else.

So as I just said, after testing the 2 DIMMs D1 & D2, I reconfigured to A2|D3 & B2|D4. Things ran fine for a few hours & I was quite optimistic this was finally a stable system. Of course not. Why would it be? Things started getting very weird. At one point, all my disk drives disappeared. Even the BIOS wasn't recognizing them. After a few power offs & crossing my fingers, I finally got booted back up & things ran fine for a few more hours. But I knew I was on borrowed time. So I shut down & powered off, disconnected the non-essential components . . . yet again, & reconfigured to A2|D3 alone.

After a couple of days running with D3 alone, I booted up Memtest86+ & started to run it on A2|D3. I fully expected this to give me failures within the first 5 or 6 tests, like I'd seen with the first 2 DIMMs. The universe decided to surprise me.



As you can see, I lost the greater part of the day to this. I have a few questions about this.

The middle part of the screen says it's a DDR5-4000. The bottom part of the screen says it's DDR5-4800. Here's the box it came in.



DDR5-6000. So which one should I believe? Plus the bottom part of the screen says it's only 2G. That's clearly not true & it treated this DIMM, as it treated all of them, as 32G. Still, that's weird.

I find it curious that the motherboard manufacturer labels the DIMM sockets A1, A2, B1, B2 but Memtest86+ labels them 0-3. Everybody has their own way of doing things, apparently.

So I started running this A2|D3 configuration for Windows. Nobody can be surprised to learn that this was a very stable configuration. I did still intend to procure a full 128G of RAM. But at least I was able to run like this for a time.

I can't give my entire life over to testing my RAM. I have to use my computer for at least a little while. But I was still determined to test my last DIMM, for which I pretty much already had proof it was faulty. I just ran with the one DIMM (D3) for a couple of days to get some relief & to appreciate a computer that wasn't going to blue screen on me at any moment & exhibit other weird behavior. But when I did get around to testing the last DIMM, it actually passed 2 passes of Memtest86+ without any failures. I thought it really wasn't useful to put potentially as long as 11 hours+ into testing a DIMM whose presence caused my system to be totally unstable but whose absence allowed my system to run just fine for days at a time.

But despite 8 clean passes through Memtest86+, this D3 DIMM still caused the occasional anomalous behaviors. Running with just the one DIMM, I never had any blue screens & I never had any freezes. But every so often my system would just spontaneously detach all the HDDs, necessitating a shut down & powering all the HDDs off for a few seconds. I also had occasional, like once every 10 or 15 times, weird errors during downloads with ffmpeg. I would get an error in ffmpeg saying it couldn't demux a packet. It looked like corrupt input. But a retry would work fine. The error seemed to be more likely to occur if I were trying to do 2 or 3 concurrent ffmpeg downloads. Sometimes, this error seemed to happen on the Internet side of things. Sometimes, it seemed to be a result of my system detaching at least the one HDD onto which I was trying to download. All very disturbing.

Even more disturbing is how the system was rebooting. Even if I did an orderly close of all running programs & a successful shutdown from the Windows desktop, the reboot always did not go as it was supposed to. The power on sequence does some system tests, reflected by certain LEDs on the motherboard that would light up in sequence. If I sit at a certain position, I can see the LEDs through the mesh top of my system case. When the first LED goes on & off, that means the CPU is working fine. The second LED indicates whether the RAM is fine. I would see the second LED go on & stay on for about 30 seconds before going off & back on again. This would happen about 6 times. Finally, a green LED would flash to indicate that the system had found a monitor. There is a fourth LED to indicate the BIOS has found a bootable device. But it never got that far.

Then the system would boot up into safe mode. I don't mean it would boot into the Windows safe mode desktop. I mean the BIOS booted into its safe mode. I would get a black screen with white type. It would show the name & version of the BIOS, the make & model of my CPU, the make & model of my motherboard, the amount of active RAM (always 32G at this point, for the 1 DIMM of my original 4 I was still using), & a list of all my hard drives. Then it would say it had booted into safe mode because of some reason I didn't pay too much attention to. Then it recommended I hit F1 to run setup & maybe restore my BIOS settings to a state I knew from experience was a stable configuration. At that point, there was no option. I had to hit F1 or hit F1. Nothing else would be accepted. That got me into the BIOS. From there, I made no changes, then I used the BIOS boot menu to tell it to boot from my NVMe SSD, where I have Windows installed. That would then bring up my Windows desktop & everything would be fine for a couple of days.

Clearly, the one remaining DIMM, which Memtest86+ had approved, was failing the POST phase of bootup. Even th0ugh I could run Windows without any problems for a few days at a time, the BIOS didn't like this one remaining DIMM.

The bottom line is that definitely my first 2 DIMMs, & probably my fourth DIMM, had serious hardware errors that Windows couldn't tolerate. The one remaining DIMM, the third one, seemed to have fooled Memtest86+ & managed to hide its faults from Windows well enough. But the BIOS didn't like it. So all my RAM went bad suddenly. The blue screen dumps weren't lying. I had a hardware problem.

I am now running 4x32G Teamgroup TForce Vulcan DDR5 6000 DIMMs. They don't have any lights. They pass the bootup POST on the first try & the BIOS doesn't put the system in safe mode. This Teamgroup RAM is cheaper than the Corsair RAM that failed. I can only hope these last. I've had computers before this one. The Windows 7 machine that preceded this W11 box ran essentially 24x7 for EIGHT YEARS without a RAM failure. I have never encountered bad RAM before in yet earlier computers. I sure hope failing RAM isn't just a modern trend.

I know you're thinking I should have subjected this new RAM to Memtest86+ testing. But I can't be without my system for 4x11+ hours just to verify brand new RAM. It's brand new. If Windows runs without errors, that's all the test I need for now. If I had a spare computer sitting in the corner on which I could let Memtest86+ do its thing for 2 days while I limp along on the 1 Corsair DIMM, I would do that. But I have only this one computer. I couldn't have used my W7 box for the memory testing, either, because it was so old it took only DDR3. Plus I no longer have that machine. So I'm just going to run like this until something happens that makes me suspect I need to test my RAM. May that never happen.

On the subject of my HDDs spontaneously detaching, I accidentally figured that one out. It was an entirely unrelated problem. One day after I had installed my new RAM, I reached over to do something & noticed that there was an awful lot of heat coming up from my HDDs. Each HDD is in an enclosure that has its own individual fan. I had been running since early December until mid-March without any problems with the HDDs. Suddenly, coincident with my RAM problem, my HDDs were detaching. I though the detaching was related to the bad RAM . . . until it happened with my new RAM. So what's the big difference between December & March? Look out any window. It's a new season. Now it's spring. It's getting warmer. The heat of the HDDs wasn't getting dissipated so well. The little fans in the enclosures are apparently not quite up to the task. I bought 4 little personal fans. At least, that's their original design intent. They have clips, & batteries that can be recharged from a USB port. You can run them while recharging. So I have arranged them strategically around my hard drives & I have not had a single detachment of my HDDs since I added these fans to my configuration. So the changing seasons added a coincidental overheating problem to the mix.

So. Bottom line. Everything is fine now. My computer has returned to 100% stability & reliability. I have even taken brand new backups of all my data.

Case closed.