BSOD when 16GB RAM installed (but not 8GB), what can I do?

[Aaargh Zombies]

Re: BSOD when 16GB RAM installed (but not 8GB)

Load between the two sticks is configured by putting the RAM in the same coloured memory slots, the OS can't mess with that.. CPU-Z says you're in dual mode which is good.

Your BIOS version does not look good. It says you are on F5b, but support for 4th gen started with F7. I would go with F8 as it has enhancements for your K processor.

If Intel burn test runs fine and it only blue screens on shutdown, its probably not a hardware issue. BTW I have 2500k with a Gigabyte board and 32GB of Crutial RAM and it's rock solid running Windows 7 through to 10. I did switch to UEFI when moving to 8.0. Don't install any of the Gigabyte utilities.

Hi

Sorry, a small mistake on my part. The CPU-Z is from when I first began trying to fix this problem. I've since upgraded the BIOS and have been on the latest version for a couple of weeks.

The BSOD is on startup only, rather than shutdown, it's rock solid and stable during usage. Even intense gaming doesn't produce any instabilities.

 

[pkcable]

DAMN you still haven't resolved this!

 

[Aaargh Zombies]

The error only seems to happen if the PC is switched off for a while, if I shut down now and come back in an hour it will boot perfectly. If I leave it off overnight it will BSOD when I switch the PC on first think in the morning.

This means that I'm essentially only able to try one possible solution per day, and it's why I'm sure that this is a software issue.

It's almost as if the PC BSOD if the date that I switched it on is different form the date that I switched it off. For example, if a service detects that it hasn't updated for 12 hours and then check for a new version and brings the system down with it.

 

[Jason Rosenthal]

You have the weirdest problem I've ever heard of. With 8GB of RAM installed, you never get a BSOD and with 16 GB it will BSOD but only after a very cold boot. I'm back to thinking it is power related. Are you by chance turning the power bar off? Perhaps there is a capacitor that is getting drained overnight and isn't getting charged in time during first boot. Will it boot properly after pressing reset or does it require a full power off reboot?

I would still try swapping power supplies. What power supply do you have in there now and how old is it?

I always have a spare power supply on the shelf ready to go, sadly I've had to buy new ones more times than I like to admit. I can't say which one is a good one, I just go by bronze or better efficiency and then physical weight. The heavier the better.

Have you disabled the onboard Intel video? It shares system RAM that could be affected by the second stick.

 

[Aaargh Zombies]

You have the weirdest problem I've ever heard of. With 8GB of RAM installed, you never get a BSOD and with 16 GB it will BSOD but only after a very cold boot. I'm back to thinking it is power related. Are you by chance turning the power bar off? Perhaps there is a capacitor that is getting drained overnight and isn't getting charged in time during first boot. Will it boot properly after pressing reset or does it require a full power off reboot?

I would still try swapping power supplies. What power supply do you have in there now and how old is it?

I always have a spare power supply on the shelf ready to go, sadly I've had to buy new ones more times than I like to admit. I can't say which one is a good one, I just go by bronze or better efficiency and then physical weight. The heavier the better.

Have you disabled the onboard Intel video? It shares system RAM that could be affected by the second stick.

I originally thought that it was the paging file, and that the file for 16GB was getting corrupted due to it's size and it not writing properly.

Yes, the PC is connected using a power bar which is unplugged at the wall, but if I shut the PC down for a couple of hours it boots up OK, it's only after being shut down for long periods of time that is BSOD.

If it's the PSU I'm going to have to wait until it totally fails, as I can't justify spending the money on it right now on the off-chance. It always BSOD in the exact same place, right to the absolute second. If it were a PSU problem I doubt that ti would be so predictable.

I've tried enabling and disabling the onboard graphics and the onboard sound. Neither effects the BSOD.

 

[Jason Rosenthal]

If it's the PSU I'm going to have to wait until it totally fails

You can do what you want, but changes in power are affecting your BSOD and not software. You know that there is no guarantee that there won't be a fire and/or a large CPU/GPU/SSD killing surge should it totally fail. It doesn't take much to weld a 22nm transistor together.

I did see a 500w Silverstone 80 plus gold certified for ~70 regular price online. Avoid getting an oversized PSU as they will create more heat and problems. I usually prefer ~400w but NVidia recommends 500w for your 760.

 

[Aaargh Zombies]

You can do what you want, but changes in power are affecting your BSOD and not software. You know that there is no guarantee that there won't be a fire and/or a large CPU/GPU/SSD killing surge should it totally fail. It doesn't take much to weld a 22nm transistor together.

I did see a 500w Silverstone 80 plus gold certified for ~70 regular price online. Avoid getting an oversized PSU as they will create more heat and problems. I usually prefer ~400w but NVidia recommends 500w for your 760.

The BSOD literally happens at the exact same time. every time. I mean to the exact second from when the power is switched on every single time that it happens.

It's also rock solid stable during gaming, which is when you'd expect to see a PSU related problem, and doesn't happen after a dirty boot. The issue also didn't happen under Windows 7.

I really don't see how this can possibly be a PSU issue. I've never heard of a PSU acts differently between a clean and a dirty boot. Or which only happens under Windows 10.

 

[Jason Rosenthal]

The BSOD literally happens at the exact same time. every time

Sorry, I thought it was only happening after long periods of your computer being unplugged.

So how are you able to boot with both memory sticks in the system currently?

 

[Aaargh Zombies]

Sorry, I thought it was only happening after long periods of your computer being unplugged.

So how are you able to boot with both memory sticks in the system currently?

I think that I put it in the first post.

I only get the BSOD if 3 conditions are met

1) Both sticks of RAM are in
2) The last shutdown was clean
3) The PC has been switched off for some time (Typically, overnight. A couple of hours during the day doesn't seem to trigger it).

If all 3 conditions are met it BSOD, I reboot by pressing the power button which means that condition 2 and 3 are not met, so it boots absolutely perfectly.

I literally only get one of these BSDO per day, and it's always the first time that I boot that day. No matter how many times I shut down and switch back on again from total power off during the day it remains stable.

 

[midnightfrolic]

Re: BSOD when 16GB RAM installed (but not 8GB)

Sounds like you have faulty hardware. New computer time unless you want to further invest more time troubleshooting the issue, and put up with bsod daily.

 

[Aaargh Zombies]

The fact that the error is only present on a clean boot says that this is software.

The fact that changing the startup apps and services changes the error also says that this is software.

If this was a OSU or RAM issue you'd expect an error when gaming or performing other high stress tasks, if it were an MB error you'd see no change between a clean and a dirty boot.

 

[a5cent]

Re: BSOD when 16GB RAM installed (but not 8GB)

Let me know if you ever get around to using Ghost.

A concise and accurate description of how behaviour changes using different OS/driver settings may also help. It's never been 100% clear how frequent which errors occur under which circumstances.
I understand you don't always get the same errors, just not what changes or when.

 

[Aaargh Zombies]

Re: BSOD when 16GB RAM installed (but not 8GB)

Do you have a download link?

 

[a5cent]

Re: BSOD when 16GB RAM installed (but not 8GB)

Sorry, GhostBuster:
https://ghostbuster.codeplex.com/
Use it after deactivating drivers in Device Manager as we once discussed.

 

[Jason Rosenthal]

The fact that the error is only present on a clean boot says that this is software.

Software absolutely cannot react differently to the length of time your computer is unplugged as per condition 3). It could be a software problem as you say only if it BSOD on each and every clean boot. It's likely the very common hardware problem of a bad capacitor in your power supply or motherboard. I would try to isolate the power supply first. My gut feeling is that it is a motherboard cap or voltage regulator for the memory modules unfortunately due to condition 1), but hopefully not.

You don't have special hardware that isn't well supported by a clean install of Windows 10 out of the box. It think what you are trying to achieve in changing your boot sequence and settings is like the dramatic scene in the Apollo 13 Movie where there trying to find the start sequence that doesn't blow the breaker for the capsule. Since you're not in the capsule or forced to use a suspect power supply, just swap it out.

 

[Aaargh Zombies]

I'm going to have to disagree with you quite vociferously on this one. Software frequently - By Design - will check the current date\time and will react differently depending on what it sees.

My computer probably performs several dozen time\date checks on boot.

For example, Windows 10 has many scheduled maintenance events, such as checking for updates, or contacting it's DHCP to check the lease on it's network address. Some maintenance functions are also performed on a scheduled basis. If you switch your computer off during one of the events they can activate when you switch it back on again.

I don't have a spare PSU to swap out, and I can't justify spending the money on a new unit on the off chance that my current one has a problem. It's also a Chilblast PC which means that it's fully cable managed. So removing the PSU would essentially entail dismantling and resembling the PC. Including removing the MB from the case to get at the cables routed behind it. That's pretty much an entire weekend to break it down and reassemble it, and the risk of damaging a component is one that I'm not willing to take on somebody else's hunch which I think is inaccurate.

I really don't see how this can be a motherboard\RAM\PSU problem as the error is so specific.

I've run stress routines on the memory and the CPU. Both in Windows and form a boot disc. I loaded both up to the hilt for several hours. They got hotter, and ran more intensively than they would during normal us, and the PC remained stable as a rock.

I only get this error on clean boot. ONLY clean boot. A hardware error would manifest itself more frequently, and more randomly.

What's more, the error always happens at the exact same place in the boot cycle. I can literally start a stopwatch and time when it will BSOD to a fraction of a second.

Is there a capacitor that is only used when the logon screen is displayed?

 

[a5cent]

Software absolutely cannot react differently to the length of time your computer is unplugged as per condition 3).

No, but then again, nobody is claiming that his problem is dependent on how long the computer was UNPLUGGED. His problem is dependent on how long ago the computer was TURNED OFF, after having been properly shut down. Software very well can and does react to that.
There is absolutely nothing in his problem description that is typical of a PSU related problem.

 

[Jason Rosenthal]

Is there a capacitor that is only used when the logon screen is displayed?

I'm just trying to help after most stopped. I've been through similar problems dozens of times; I'm a Deer PSU fiasco survivor, however I can't say the same about the computers. RAM related issues can often be traced back to poor power. Or am I in wrong about it being RAM related now from point 1)?

If you are unfamiliar with how RAM and power are related; RAM requires a refresh cycle hundreds of times per second in order to retain its values. If it loses power, a refresh is missed and then your computer is going to crash eventually. It is usually expected to crash at random, but not always. RAM requires good stable power at all times to operate. Also more RAM requires more power. Your stress tests are likely passing because the drained capacitor I propose is fully charged by the time the tests are stared and/or the power supply is in its efficiency window under the higher load. Either way, your RAM at that point is getting the power it needs.

I'm thinking the problem likely starts within the first second powering on from cold boot after the caps have drained and not a specific logon screen capacitor. Generally, the first stage of the application lifecycle is initialization of variables. The initialization of the UEFI may have failed in some weird way (like RAM losing its refresh power) that isn't discovered until later in the boot process. The UEFI or any software isn't going to back and check that the initialization phase worked. Perhaps it is the same address space(s) in RAM that are losing their refresh each time and the reason why you are seeing the identical blue screen logs. A defective/drained capacitor can cause all this on first cold boot because the power is used to finish charging the capacitor instead of refreshing the RAM.

If it there were a software issue or anything to do with the real-time clock like you say, it should have failed with one stick of RAM as well. If were a driver issue relating to RAM size, then it should fail every boot with 16 GB installed. There is nothing unique about your hardware and largely the same design Intel has been using for the last 5 years sadly. I have the same brands in my system, just slightly older. A clean install of Windows 10 is all that is required for a reliable setup provided you don't have a hardware/firmware issue. Recheck your motherboard firmware, perhaps its only supplying DDR3L 1.3 volts instead of the required 1.5 volts for your RAM. Check the bios to see if it reporting the correct model of CPU, if not it probably isn't getting the correct power either. A marginal cap is really the only thing that I can think of that covers all your symptoms.

 

[Aaargh Zombies]

I've checked the firmware, and the BIOS. All are at the most current version.

I checked the voltage, the BIOS confirms that it is at 1.5v, and the Windows diagnostics software says that it is also at 1.5V

CPU and RAM are both being correctly reported in BIOS and in Windows.

If it were a capacitor issue, and it isn't fully charged, then shouldn't I simply be able to boot to BIOS for 5 minutes, and then boot to Windows, as the time in BIOS would allow the capacitor to charge?

The motherboard is set to Auto, so if the voltage dips the RAM is clocked down. Even if it were only getting 1.3V it should remain stable.

I think that this is software because when I add\remove software from the startup routine the time that the error happens changes. Something is being loaded\read earlier\later and is causing the BSOD to happen earlier\later.

As far as I'm aware Windows 10 will split the load between two sticks if it detects that they are correctly paired. Could it simply be an older 32 bit driver or service that doesn't like being split up between two 8GB sticks? If the software can't address 16GB then it won't like being split across 16GB.

If they is only one stick then it simply resides at the highest address that it can register on the 8GB stick.

I'm going to remove and uninstall my sound card today as it wasn't factory installed. It's an older model Sound Blaster and is the most likely component to still be running older 32 bit packages.

 

[a5cent]

I'm going to remove and uninstall my sound card today as it wasn't factory installed. It's an older model Sound Blaster and is the most likely component to still be running older 32 bit packages.

Good. Once removed also use GhostBuster to rid yourself entirely of the corresponding driver, as it otherwise stays on your system.