Is my RAM failing?

[robot zombie]

When bad ram is in the mix it can cause all sorts of problems from bsods to resets.

So I've seen. It makes sense, since basically everything that happens between components and software is dependent on memory having all of the right instructions and feeding the information whenever asked. Everything stops when it does.

Resets tell me that Windows is set to automatically restart upon crashing. To get a bsod to appear you must disable windows from auto restarting upon failure.

Ahh, makes sense. I guess I'm a little confused because I get both the restart and BSODs. I'm guessing it depends on the nature of the crash. I just found it strange that when it restarted without a BSOD, I couldn't even find any information on why in the event viewer. All it would tell me that it rebooted to recover from a crash. And I'm sitting here wondering "Well, yeah. But WHY did it crash?"

I'm sure its in there somewhere. Maybe I'm filtering things down wrong.

So the problem was fixed on your rig when you swapped ram and it moved to the other machine, it is the ram there bucko.

Yeap. Kind of a bummer. I generally like kingston. Incidentally for their reputation (and my experience) of good reliability. Just got unlucky I guess. Gonna have to contact them. Hopefully I can keep the working one... ...once I figure out which one it is. Since they both pass memtest and just generally don't show measurable signs of failure, I'll have to try running them separately and see which rig crashes. <_< That's why I was really hoping for some definitive proof basically showing me, "This particular module is bad."

I guess if I can't keep the working one, the living room rig goes down for a while. Or maybe I run both on 4gb lol.


Gotta say, this is about the least eventful hardware failure I have ever dealt with. I once had a PSU literally go up in sparks and smoke, followed by very loud and constant humming and whining. I've seen a couple of HDD's fail in a dramatic fashion, with nasty scraping, chirping, and clicking/clacking. I've seen capacitors vent rather violently (for what they are anyway.)

Thanks to all for your patient input... ...especially eidairaman1. I know it must've felt like a test at some points, like I was just wasting your time. If it did I swear I never meant to and I really do appreciate it! I'm just getting back into this stuff and for the past several years I've used nothing but linux, so it's all basically new to me at this point. I really appreciate the point-to-point guidance. I learned a few things about troubleshooting on windows that I think will be worth a stick of ram in the end! Glad we could get this figured out.

 

[eidairaman1]

So I've seen. It makes sense, since basically everything that happens between components and software is dependent on memory having all of the right instructions and feeding the information whenever asked. Everything stops when it does.


Ahh, makes sense. I guess I'm a little confused because I get both the restart and BSODs. I'm guessing it depends on the nature of the crash. I just found it strange that when it restarted without a BSOD, I couldn't even find any information on why in the event viewer. All it would tell me that it rebooted to recover from a crash. And I'm sitting here wondering "Well, yeah. But WHY did it crash?"

I'm sure its in there somewhere. Maybe I'm filtering things down wrong.


Yeap. Kind of a bummer. I generally like kingston. Incidentally for their reputation (and my experience) of good reliability. Just got unlucky I guess. Gonna have to contact them. Hopefully I can keep the working one... ...once I figure out which one it is. Since they both pass memtest and just generally don't show measurable signs of failure, I'll have to try running them separately and see which rig crashes. <_< That's why I was really hoping for some definitive proof basically showing me, "This particular module is bad."

I guess if I can't keep the working one, the living room rig goes down for a while. Or maybe I run both on 4gb lol.


Gotta say, this is about the least eventful hardware failure I have ever dealt with. I once had a PSU literally go up in sparks and smoke, followed by very loud and constant humming and whining. I've seen a couple of HDD's fail in a dramatic fashion, with nasty scraping, chirping, and clicking/clacking. I've seen capacitors vent rather violently (for what they are anyway.)

Thanks to all for your patient input... ...especially eidairaman1. I know it must've felt like a test at some points, like I was just wasting your time. If it did I swear I never meant to and I really do appreciate it! I'm just getting back into this stuff and for the past several years I've used nothing but linux, so it's all basically new to me at this point. I really appreciate the point-to-point guidance. I learned a few things about troubleshooting on windows that I think will be worth a stick of ram in the end! Glad we could get this figured out.

You learn something today so no worries, you managed to shorten your troubleshooting time and resolve your problem fairly quickly by the suggestions. Now your other machine I will say right now that it most likely is that memory but you need to disable the auto restart function of Windows in order to determine if you're getting a bsod, now if it does not I would highly suggest the plausibility of temperature being a problem.

 

[robot zombie]

You learn something today so no worries, you managed to shorten your troubleshooting time and resolve your problem fairly quickly by the suggestions. Now your other machine I will say right now that it most likely is that memory but you need to disable the auto restart function of Windows in order to determine if you're getting a bsod, now if it does not I would highly suggest the plausibility of temperature being a problem.

Well that won't be necessary now. Now I know for sure.

Came home to find the build with the bad RAM in it caught in a hard lock-up. After a hard reboot, DRAM light on the mobo is going crazy and it's caught in a pre-boot loop. Cut power for 30 seconds and it posts, but only shows 4gb of ram. From there I booted and loaded up CPU-Z to see which slot was actually registering. Slot #2. OK. I then shut down, removed both, and put a sticker on the stick that was in slot #4. I then proceeded to try that stick in every slot, on both machines, in a single configuration. It no posty post. Other stick posts fine in both machines. So there it is. It finally up and quit after all of this time.

So now, I've got the working set of... DDR4 2133 from my other machine in this one, and the working stick from the bad set in the other one. I'm sitting at my desk holding a bad stick of ram with a big flourescent pink sticker on it, praying to the gods of tech support that kingston goes easy on me and doesn't make me go through 1000 motions in order for them to understand how much time and effort I have gone through to find out for sure that yes, the stick is toast. I then pray that they'll let me keep the good stick. Cannot stress how much I am not a fan of this game. I am good at it, which is why I'm the chosen one in my family... ...doesn't mean I like it lol

 

[eidairaman1]

Well that won't be necessary now. Now I know for sure.

Came home to find the build with the bad RAM in it caught in a hard lock-up. After a hard reboot, DRAM light on the mobo is going crazy and it's caught in a pre-boot loop. Cut power for 30 seconds and it posts, but only shows 4gb of ram. From there I booted and loaded up CPU-Z to see which slot was actually registering. Slot #2. OK. I then shut down, removed both, and put a sticker on the stick that was in slot #4. I then proceeded to try that stick in every slot, on both machines, in a single configuration. It no posty post. Other stick posts fine in both machines. So there it is. It finally up and quit after all of this time.

So now, I've got the working set of... DDR4 2133 from my other machine in this one, and the working stick from the bad set in the other one. I'm sitting at my desk holding a bad stick of ram with a big flourescent pink sticker on it, praying to the gods of tech support that kingston goes easy on me and doesn't make me go through 1000 motions in order for them to understand how much time and effort I have gone through to find out for sure that yes, the stick is toast. I then pray that they'll let me keep the good stick. Cannot stress how much I am not a fan of this game. I am good at it, which is why I'm the chosen one in my family... ...doesn't mean I like it lol

Just explain to them that you're getting blue screens of death one of them related to hardware and also lockups and you move the same module to a different computer and it did the same exact thing.

 

[robot zombie]

Just explain to them that you're getting blue screens of death one of them related to hardware and also lockups and you move the same module to a different computer and it did the same exact thing.

It actually looks fairly easy and hands off, they actually have an option to send out the replacement to you before they receive your return. You have I think 10 days to send them out after receiving your replacement! Of course they put a hold on your credit card, but that's still pretty sweet.

But I haven't gone through with it yet. I just keep making more interesting discoveries. I dunno if I'm unlucky or just stupid now.

I was getting ready to push the button on an RMA when I decided to take one last look at the bad stick. That's when I saw this...


It's got schmoo on it! Holy crap! I went and pulled the "good" stick out of the other machine and sure enough, same schmoo, same spot...

So now, I know that at least one of my ram slots has shmoo in it. How this happened, I don't know. I promise you I handle all of my components very carefully. I don't think I ever even set these sticks down... ...they went straight from the package to the freshly opened mobo. I never even touched the contacts. Such a tiny window for anything to get in there. I wonder if one of them had it on it when I got it... ...or maybe it was in the mobo slot... ...sounds unlikely but looking back on when I did this build I am kind of in disbelief. I'm pretty meticulous when it comes to environment, handling, and esd.

It looks weird in the picture, but in person it looks and behaves like burnt dust with a dash of gook. After cleaning with alcohol, most of the nastiness is gone, but there are still small burn marks on the affected contacts. The good news is that after cleaning, both sticks post and have been running with zero problems since my last post. The problem is that they have to be in a specific configuration. The "good" module will post in any slot. The "bad" one only works in slot #3. I'm assuming the contacts inside EVERY slot are now dirty and possibly damaged too. They must line up just right for shorting out in certain slots. I'm suddenly regretting rotating them around so much earlier. Never stopped to think maybe it was a different sort of contact problem... ...just checked to make sure they were seated properly. I do everything I can to avoid contamination like this. Not enough though, apparently.

At least the contacts on the other RAM I swapped in are perfectly clean. I can deduce from this that the majority of the shmoo adhered to the sticks and not the contacts. May just be some leftover residue in the slots combining with the hampered conductivity of certain parts of the shorted contacts that's causing the worse-off stick not to post in certain slots.

I did go ahead and take the duster to all four slots, but I think it's going to take more than that to clean them out... ...what I wouldn't do for an ultrasonic cleaner about now. I used to have one :/

Not even sure it's cool to send these back now. Technically not normal operating conditions. Nothing was wrong with them until dust got into the slots...

 

[eidairaman1]

It actually looks fairly easy and hands off, they actually have an option to send out the replacement to you before they receive your return. You have I think 10 days to send them out after receiving your replacement! Of course they put a hold on your credit card, but that's still pretty sweet.

But I haven't gone through with it yet. I just keep making more interesting discoveries. I dunno if I'm unlucky or just stupid now.

I was getting ready to push the button on an RMA when I decided to take one last look at the bad stick. That's when I saw this...
View attachment 102783

It's got schmoo on it! Holy crap! I went and pulled the "good" stick out of the other machine and sure enough, same schmoo, same spot...

So now, I know that at least one of my ram slots has shmoo in it. How this happened, I don't know. I promise you I handle all of my components very carefully. I don't think I ever even set these sticks down... ...they went straight from the package to the freshly opened mobo. I never even touched the contacts. Such a tiny window for anything to get in there. I wonder if one of them had it on it when I got it... ...or maybe it was in the mobo slot... ...sounds unlikely but looking back on when I did this build I am kind of in disbelief. I'm pretty meticulous when it comes to environment, handling, and esd.

It looks weird in the picture, but in person it looks and behaves like burnt dust with a dash of gook. After cleaning with alcohol, most of the nastiness is gone, but there are still small burn marks on the affected contacts. The good news is that after cleaning, both sticks post and have been running with zero problems since my last post. The problem is that they have to be in a specific configuration. The "good" module will post in any slot. The "bad" one only works in slot #3. I'm assuming the contacts inside EVERY slot are now dirty and possibly damaged too. They must line up just right for shorting out in certain slots. I'm suddenly regretting rotating them around so much earlier. Never stopped to think maybe it was a different sort of contact problem... ...just checked to make sure they were seated properly. I do everything I can to avoid contamination like this. Not enough though, apparently.

At least the contacts on the other RAM I swapped in are perfectly clean. I can deduce from this that the majority of the shmoo adhered to the sticks and not the contacts. May just be some leftover residue in the slots combining with the hampered conductivity of certain parts of the shorted contacts that's causing the worse-off stick not to post in certain slots.

I did go ahead and take the duster to all four slots, but I think it's going to take more than that to clean them out... ...what I wouldn't do for an ultrasonic cleaner about now. I used to have one :/

Not even sure it's cool to send these back now. Technically not normal operating conditions. Nothing was wrong with them until dust got into the slots...

Clean the ram modules, clean the slots using contact cleaner and air dusting. I still think 1 still fails if you xan only put it in slot 3 (whereever that is) i was thinking slot A1, B1, A2, B2

 

[robot zombie]

Clean the ram modules, clean the slots using contact cleaner and air dusting. I still think 1 still fails if you xan only put it in slot 3 (whereever that is) i was thinking slot A1, B1, A2, B2

Yep, gonna pick up some contact cleaner tomorrow. And yeah, A2 would be the slot.

See, I'm not sure if it is an outright failure or a contact issue. If, after cleaning the slots better it works in any slot, passes stress tests and runs stable at all times, I wouldn't say that's a failure. I mean, looking at these contacts, they're still all there structurally, as in not marred. Pretty sure it's just crud baked on. I know that surface is just microns thin, if that, but I've got to figure that if I can reduce the tarnish on the surface enough it should still make good contact. And from there I can't see an issue with it. It's not like the gold is gonna wear or corrode just sitting in the slot.

Honestly, I think I'd rather just be able to clean it up and get it working than go through warranty RMA. I'm not super hopeful but I'd just feel kind of silly sending out a module for replacement when all I had to do was clean everything up.

Though I do see your point. If I can't get it to post across different slots, I will be sending it out. Right now I'd like to test it in the other build, but it's running its nightly scan/purge/backup routine.

 

[itsakjt]

To my experience, most BSODs and random system lock ups mostly always points to faulty RAM or RAM with loose contact especially when the BSODs are frequent. And among those, most of the problems are from corrosion on the gold plated contacts of the RAM.
@OP - The reason why the first stick was warmer than the other one is because it is closer to the CPU. Also assuming that you are using an air cooler with omni-directional airflow, the warm air will reach the first stick of RAM.
Also, many times these corrosion are due to a combination of dust and moisture. And the air cooler actually is kinda responsible for that as the RAMs act as a dust barrier to the air cooler which spreads dust. A good way to minimize it is putting a piece of paper between the CPU cooler and the RAM (you need to improvise on how you are going to do it but I bet, it will help you a lot).

 

[Caring1]

Try using an eraser to gently rub over the contacts on the dim.
That should remove any gunk built up on them.

 

[robot zombie]

Welp, armed with a natural gum eraser, a small artist's brush, a can of electronic cleaner, a can of air, some alcohol, and microfiber, I set out to clean everything off.

First thing I did to clean the slots was run the artist's brush up and down to loosen everything up. After that, I hit em up and down with the air. After that, I sprayed the areas with schmoo in em with CRC QD Electronic Cleaner (which seems pretty great btw - safe on delicate electronics, evaporates fast and leaves no residue.) I was a little generous lol. I sprayed more air in before the solvent evaporated to push out whatever the solvent had caught. Sprayed in more after they dried to ensure that everything was out.

For the shmooey contacts on the ram, I decided to try the eraser trick as recommended. Used natural gum simply because I read that others contain abrasives. It actually worked very well. I used the alcohol and cloth to remove any residue left behind. The majority came right off.

Unfortunately, what's left on there appears to be actual damage to the plating... either tarnish or complete burn-off. Actually, it looks like copper underneath. I never knew they were gold-plated copper. Dunno what I thought was under there though lol. I sprayed the cleaner on to see if it was just corrosion. The cleaner would remove it if so. But alas, the gold's just worn through in a few tiny spots. Lucky, the actual contact points in the centers are intact. Most of the wear appears to be towards the outer edges of the fingers. It'll probably spread to the center eventually, I suppose.

I'm hoping this damage is mostly limited to the contacts on the ram and not the contacts inside the slots. Whatever damage was caused in there doesn't seem to pose a problem, since other RAM worked fine. But I'm worried that corrosion set off by this event may come back to haunt me in a coupla years, when the slots potentially become corroded enough to brick the board. Last-ditch I could toss it into an ultrasonic bath, but if the material is gone, that's kinda that.

The good news is that either stick will now post and pass memtest in any slot. So it appears that the cleaning worked as I'd hoped. For now, I'm going to keep both sticks. The damage to the contacts definitely shaved down their lifespan and its only a matter of time till that shows, but seeing as how they seem to be working for now and carry a lifetime warranty, I'm going to keep them, right along with my proof of purchase... If the problems return, I'll send them both back.

I'm considering a thin layer of dielectric grease on the damaged contacts, just to protect those areas from the inevitable corrosion. For now, I'm gonna leave it be. Not something I want to delve into or later have to purge out of my RAM slots. Potentially hampering conductivity further seems iffy. Just kind of a thought in my head.


Thanks again to all who contributed. This ended up being wayyyy more of an adventure than I ever thought it'd be! And to think it ended up being something so basic. All I had to do was actually LOOK at my RAM