RAM Issues on X58 Boards

Found this post whilst looking for a solution.

Sabertooth x58 board
i7 960 3.2ghz
3X4gb Corsair Vengeance

In triple channel it will only find 4gb, when fitting Ram to A1 A2 + B1 finds 8gb. No matter what Ai Tweaker settings, reseating Cpu/Ram most it'll ever post is 8gb. I've tested each separately and they all post 4gb so they are all working. It's driving me literally insane..

Any suggestions?
 
David-Duc said:
It's been a long time but if I recall correctly, one of the possible cause of the X58 memory issues is that the 1366 socket latch wasn't strong enough and the pressure is distributed unevenly which in turn caused the CPU to "lose memory slots/channels" due to bad contact. Intel remedied the issue by using the dual latch mechanism on socket 2011 and forth. If possible, you may want to reinstall the processor and make sure the contacts between the pins and pads are all good.
Click to expand...

Very belated X58 enthusiast here.

I registered an account here, just to offer a solution. I had tried everything mentioned in this thread, but no dice. Then I read the above and I have found exactly this to be the cause of banks going "missing".

The solution? Add some padding between the cpu and the small part on the latch pushing down on (making contact with) the CPU. Specifically, I only did this to the left. I used a tiny piece of cardboard.

I was consistently able to reproduce RAM gone missing (no cardboard) and RAM showing up (cardboard). This was tested/reproduced with a GA-X58-UDR3 rev 1.0 board breadboxed on a flat surface with a stock LGA 1366 heatsink sitting loose on the CPU (e.g. not mounted).

Depending on your CPU cooler and/or motherboard orientation you might be able to apply enough pressure in a different way. This is likely why people in this thread report tinkering with (e.g. reseating the CPU) seems to "solve" the problem.

Hope this helps anyone so many years later! Love this platform. :)
 
j49 said:
Very belated X58 enthusiast here.

I registered an account here, just to offer a solution. I had tried everything mentioned in this thread, but no dice. Then I read the above and I have found exactly this to be the cause of banks going "missing".

The solution? Add some padding between the cpu and the small part on the latch pushing down on (making contact with) the CPU. Specifically, I only did this to the left. I used a tiny piece of cardboard.

I was consistently able to reproduce RAM gone missing (no cardboard) and RAM showing up (cardboard). This was tested/reproduced with a GA-X58-UDR3 rev 1.0 board breadboxed on a flat surface with a stock LGA 1366 heatsink sitting loose on the CPU (e.g. not mounted).

Depending on your CPU cooler and/or motherboard orientation you might be able to apply enough pressure in a different way. This is likely why people in this thread report tinkering with (e.g. reseating the CPU) seems to "solve" the problem.

Hope this helps anyone so many years later! Love this platform. :)
Click to expand...


Kyle Bennett liked your very first post.

WTF
 
Reefa_Madness said:
Well, while it may be a first post, I would add that it certainly included some pretty interesting information.
Click to expand...
Guys, I'm in the process of necroing my old X58 1920 Bloomfield rig, and I ran into this problem tonight. I have the Gigabyte X58 UD5 Rev 2 board. I also have the Gigabyte X58 UD5 Rev 1 board.

Guess what? I had to RMA the Rev 1 back in like 2012. I got the RMAed board back, but never fired it up because it took a month to get back. In the mean time, back in 2012, I bought the Rev 2 board, and after reinstalling everything, I got all of my RAM back (12GB). Now, the new board is again showing 8GB of Ram. I have the cooler screwed down with pressure spring screwed with a mounting plate on the back of the CPU (Thermaltake cros member that fits across the cooler CPU mount. That thing is tight too. I just remounted the cooler after 7 years. I jsut noticed after trying to get my BIOS, which was wiped do to a dead CMOS battery, back in shpe.

Good thing all of the old ocing setting are still there backed up in the BIOS. However, I just loaded optimal defaults to make sure all was well, and I noticed two banks of ram not posting. Any ideas? BTW,, I am a longtime meber. My old sig was "DOUGWD" here at HARDO
WP_20191208_19_40_28_Pro.jpg
 
That bios screen makes my eyes bleed...

lol

good times. I miss my i7 920...
 
86 5.0L said:
That bios screen makes my eyes bleed...

lol

good times. I miss my i7 920...
Click to expand...
HELP ME!!!!!!!!!!!!!!!!!! LOL I'm so glad I had my old settings saved in the BIOS. Anyway, you can see the slots showing blank in the "Enabled Slots" line. How do I enable them? GD it!

Might want to reply in the new thread I created. I dunno. just need help. This old shit. I wish I had the money to just GD wel upgrade it all. Too many old memories.
 
[SOLVED]
Hi everyone
I believe this problem has been solved:
It's the aftermarket CPU coolers that put too much force on the motherboard around the CPU when they use non-stock tightening mechanisms (not the stock plastic push-pins of original stock Intel CPU coolers).

I bought myself 4 GIGABYTE X58A-UD5 motherboards.
On ALL boards all memory worked flawlessly when I got them. Then I installed an aftermarket CPU cooler (by quiet TF2) with hard metal screws to secure the cooler via a backplate.
After tightening the CPU cooler on tightly, at first everything worked. Then I had to change some RAM modules and had to push quite a bit to get them in without removing the cooler.
Then I ran the PC again, and as the board heated up, memory in the last two slots (5 and 6) suddenly started failing. I worked, then it didn't then it worked again and then it gave up for good.
Eventually the memory in the slots 5 and 6 was never enabled again. BIOS detects the memory, but it's never available for usage again.
I tried everything, nothing helped. Not even installing the stock original Intel CPU cooler again. The board is toast / broken.

Then I replaced the motherboard with an identical one that also worked flawlessly. Testing the same memory with a stock original Intel CPU cooler and everything was perfectly fine. Like before.
Then I put on the same aftermarket "be quiet" CPU cooler with the hard metal screws to tighten it on. This time I did not tighten them quite as hard.
When testing memory with Memtest86+, suddenly one memory module (of 6 in total) gave an error. Before the aftermarket cooler it always all worked flawlessly.
I tried testing each memory module individually then, to find out which one was failing, and to my horror I noticed, the NO MEMORY was properly working anymore in the dreaded slots #5 and #6 of the board.
Same phenomenon. Going into BIOS it looked like this board was toasted as well. BIOS recognized all memory, but the memory in slots 5 and 6 was no longer enabled and not usable.
Then I removed the aftermarket CPU cooler and put the stock original Intel CPU cooler back on.
And voila: Luckily, immediately all memory was working and enabled again!!! Memtest86+ reports no more errors and all is detected and enabled as it should! Even after running about 10 minutes during Memtest.
Thank God! I was lucky, the board was not quite toasted yet!!! :-O

Conclusion: All these X58 chipset motherboards are EXTREMELY EXTREMELY EXTREMELY SENSITIVE to the tightening of CPU coolers and how much the CPU is pushed down on the motherboard.
If there is ANY force pushing down on the CPU that is greater than the force which a stock original Intel CPU cooler does, the ones with the black push-pins, the motherboard will short out memory slots number 5 and 6 and actually cause permanent damage. This means, DO NOT USE ANY CPU coolers on these boards that do not have the stock original Intel type CPU fastening push-pins made of plastic. NEVER EVER use CPU coolers on these boards that are actually screwed on with metal screws! NEVER. Best: To be safe: NEVER USE ANY OTHER CPU COOLER on these boards than ORIGINAL STOCK Intel CPU coolers!
That's what I am doing!

Keep in mind, even if you do not tighten on metal screws very much, they are still a lot more unyielding than the plastic push-bins of original stock CPU coolers. When the CPU heats up, it expands quite a lot. So does the CPU cooler and the backplate on the other side of the CPU. All this makes the fastening points around the CPU to be pulled up, or the CPU down very much, greatly increasing the pressure, as the CPU and cooler heat up. Only the plastic push-pins of original stock CPU coolers seem to give enough tolerance to not put too much bending force on the motherboard around the CPU.

Well, that's what I found out.
Good luck everyone on this!
- Mark

PS: Never use Windows Memory Diagnostic! It does NOT detect 95% of memory errors! Always use Memtest86+, included on the free Clonezilla Live CD!
 
Last edited:
Tyler3 said:
[SOLVED]
Hi everyone
I believe this problem has been solved:
It's the aftermarket CPU coolers that put too much force on the motherboard around the CPU when they use non-stock tightening mechanisms (not the stock plastic push-pins of original stock Intel CPU coolers).

I bought myself 4 GIGABYTE X58A-UD5 motherboards.
On ALL boards all memory worked flawlessly when I got them. Then I installed an aftermarket CPU cooler (by quiet TF2) with hard metal screws to secure the cooler via a backplate.
After tightening the CPU cooler on tightly, at first everything worked. Then I had to change some RAM modules and had to push quite a bit to get them in without removing the cooler.
Then I ran the PC again, and as the board heated up, memory in the last two slots (5 and 6) suddenly started failing. I worked, then it didn't then it worked again and then it gave up for good.
Eventually the memory in the slots 5 and 6 was never enabled again. BIOS detects the memory, but it's never available for usage again.
I tried everything, nothing helped. Not even installing the stock original Intel CPU cooler again. The board is toast / broken.

Then I replaced the motherboard with an identical one that also worked flawlessly. Testing the same memory with a stock original Intel CPU cooler and everything was perfectly fine. Like before.
Then I put on the same aftermarket "be quiet" CPU cooler with the hard metal screws to tighten it on. This time I did not tighten them quite as hard.
When testing memory with Memtest86+, suddenly one memory module (of 6 in total) gave an error. Before the aftermarket cooler it always all worked flawlessly.
I tried testing each memory module individually then, to find out which one was failing, and to my horror I noticed, the NO MEMORY was properly working anymore in the dreaded slots #5 and #6 of the board.
Same phenomenon. Going into BIOS it looked like this board was toasted as well. BIOS recognized all memory, but the memory in slots 5 and 6 was no longer enabled and not usable.
Then I removed the aftermarket CPU cooler and put the stock original Intel CPU cooler back on.
And voila: Luckily, immediately all memory was working and enabled again!!! Memtest86+ reports no more errors and all is detected and enabled as it should! Even after running about 10 minutes during Memtest.
Thank God! I was lucky, the board was not quite toasted yet!!! :-O

Conclusion: All these X58 chipset motherboards are EXTREMELY EXTREMELY EXTREMELY SENSITIVE to the tightening of CPU coolers and how much the CPU is pushed down on the motherboard.
If there is ANY force pushing down on the CPU that is greater than the force which a stock original Intel CPU cooler does, the ones with the black push-pins, the motherboard will short out memory slots number 5 and 6 and actually cause permanent damage. This means, DO NOT USE ANY CPU coolers on these boards that do not have the stock original Intel type CPU fastening push-pins made of plastic. NEVER EVER use CPU coolers on these boards that are actually screwed on with metal screws! NEVER. Best: To be safe: NEVER USE ANY OTHER CPU COOLER on these boards than ORIGINAL STOCK Intel CPU coolers!
That's what I am doing!

Keep in mind, even if you do not tighten on metal screws very much, they are still a lot more unyielding than the plastic push-bins of original stock CPU coolers. When the CPU heats up, it expands quite a lot. So does the CPU cooler and the backplate on the other side of the CPU. All this makes the fastening points around the CPU to be pulled up, or the CPU down very much, greatly increasing the pressure, as the CPU and cooler heat up. Only the plastic push-pins of original stock CPU coolers seem to give enough tolerance to not put too much bending force on the motherboard around the CPU.

Well, that's what I found out.
Good luck everyone on this!
- Mark

PS: Never use Windows Memory Diagnostic! It does NOT detect 95% of memory errors! Always use Memtest86+, included on the free Clonezilla Live CD!
Click to expand...
Are X58 motherboards even a 'thing' anymore? I mean, they're dinos from 2010. Interesting find, though. This shit drove me crazy back in the day.
 
Never had any problem running those on Easter blocks and used to torque the hell out of those. Maybe over time?
 
FrgMstr said:
Never had any problem running those on Easter blocks and used to torque the hell out of those. Maybe over time?
Click to expand...
What are "Easter blocks"?

Additional info on my finding:
X58 motherboards have a CPU backplate on the reverse side of the motherboard, on the other side of the CPU socket. In the middle of the backplate there is a hole through which you can see many tiny resistors.
What I found is that what these X58 motherboards *REALLY* can't stand, is if you have an aftermarket CPU cooler with an additional backplate, that goes over this CPU backplate on the motherboard, and then squeezes down on that. That's exactly what my "be quiet!" Shadowrock TF2 cooler did. Essentially, it sandwiches and squeezes together the CPU and the CPU backpate on the MB. That's what X58 MBs are EXTREMELY sensitive about. It seems there are a lot of very sensitive circuits in between the CPU or its socket, and the CPU backplate on the other side of the motherboard. And pushing these to together, as aftermarket CPU coolers do with backplates that push on the CPU backplate simply disrupt and damage these fragile circuits. Even if the board initially works, as soon as the CPU and backplate heat up, they expand and the pressure on the fragile electronics between CPU socket and backplate probably increases by magnitudes if they are sandwiched together by CPU coolers that push down on the backplate. That's why, if you're very quick in cutting the power when you first start getting memory errors, sometimes you can save the board before connections are permanently damaged, as in my last case.

In my case, I was able to save my last board that I put my "be quiet" cooler on by immediately turning off the power when memory errors started to happen due to the system getting up to operating temperature.
What's more, I now have the exact same cooler on the system and it's running flawlessly. Only I am no longer using the cooler's original backplate to fasten it to the motherboard (the one that goes over the motherbaord's CPU backplate), but instead, simply two zip ties that go through the original CPU cooler holes in the motherboard. So the CPU backplate on the motherboard is totally free and naked.
It seems as long as a CPU cooler, even aftermarket coolers are fastened *ONLY* to the holes in the motherboard, and not also push down on or obstruct the motherboard backplate, there is no problem. As if the motherboard backplate really really needs free space to expand when it heats up.
Click to expand...
 
Last edited:
DWD1961 said:
Are X58 motherboards even a 'thing' anymore? I mean, they're dinos from 2010. Interesting find, though. This shit drove me crazy back in the day.
Click to expand...

eh, when I upgraded to a Ryzen system I kept my x58 system and installed XP on it to play old games, recently I spend a lot more time on it than my modern machine

but I am old
 
I had this issue on my asus p6x58d-e with 6 x 8gb sticks. I was getting about 32gb of memory showing up when I definitely had 48gb available. I would sometimes have to move pci-e cards around to prevent memory from disappearing or use different types of cards.

I took apart all the components one day, including the back bracket on the cpu (had to use a hair dryer), re-assembled it, and the 48gb of ram showed up. I placed the case sideways this time and it's been fine since. Overclocking it even higher now too.
 
You must log in or register to reply here.
Back
Top