What would cause a DRAM slot to stop functioning?

[napster100]

Hi guys... So about 3 days ago I installed a custom water cooling loop, blah blah etc... And after I had problems booting up, which has since been solved, it turns out to be a problem with RAM.

I originally thought I fried one of my RAM sticks, and because they were a rather cheap set, I thought I'd buy a set of Corsair Vengeance Pro and grabbed 4 x 4GB. Brand new they should work out the box, and they seem too quite nicely, though the problem is I can only use 2 sticks... It seems the DRAM slot closest to the CPU is either playing up, or is just dead. Since its dual channel (of course), I'm guessing that may have an effect on bay 3 (assuming bay 4 is closest to the CPU), is this correct? I'm not sure, but it seems only bays 1 and 2 (furthest from the CPU) are operable.

So here's what I've tried. Basicly every stick configuration possible. I popped all of them and I get a power up loop where it stays on for maybe 4 seconds, if I take out the stick from bay 4, it stays on for a mere 30 seconds before power looping again (nothing on screen during this time). So I took out sticks 3 and 4, It booted fine, I thought I'd take this chance to jump into the BIOS and set the RAM speed and voltage to what was advertised for it, rebooted popped the 2 sticks back in, again, power loop. So I reset BIOS to default config, but still a boot loop. I tried a single stick in each slot, 1 and 2 booted past POST absolutely fine, 3 had the 30 seconds power on time before a reboot and repeat. The last slot (number 4) gave the same 4 seconds power on time before also rebooting and repeating. I tried each stick with the same outcome.

It can't be down to reseating the sticks, as you can imagine the amount of times I've pulled those out and back in, I even tried a little extra pressure just to be sure... So it looks like either I've damaged it some how, though I'm always gentle when it comes to my precious metal box ah, or its just failed...

What I'd like to know is if you maybe able to give an explanation as to how this may have happened, or if there's some mystical thing in the BIOS that maybe stopping dual channel memory or something, I dunno.

Shall I just accept my fate and maybe seek out a new board? I mean I don't want to be wasting 8GB of RAM that could certainly be put to use right now!

 

[holyprof]

I had a similar problem on my old socket 775 system, had 2x2GB DDR2. Adding one more 2GB DDR stick caused random blue screens, lockups and reboots. Adding 2 more 2GB sticks caused reboot loops past POST screen. I tried everthing except swapping to a new motherboard. I suspect power problems (motherboard RAM VRM) or BIOS compatibility problem.
I've learned my lesson, my new system has 2x8GB instead of 4x4GB to avoid that type of problems.

I'd suggest trying to force lower DRAM frequency from BIOS (1333MHz for example) and see if it works. Try running test programs like MEMTEST on every single module and then on pairs to try and detect any incompatibility when matching 2 of them.
Also ask a friend with compatible motherboard to test your RAM sticks. It might be a problem of your motherboard with particular RAM modules and the motherboard and RAM modules could work OK when separated.

 

[napster100]

I had a similar problem on my old socket 775 system, had 2x2GB DDR2. Adding one more 2GB DDR stick caused random blue screens, lockups and reboots. Adding 2 more 2GB sticks caused reboot loops past POST screen. I tried everthing except swapping to a new motherboard. I suspect power problems (motherboard RAM VRM) or BIOS compatibility problem.
I've learned my lesson, my new system has 2x8GB instead of 4x4GB to avoid that type of problems.

I'd suggest trying to force lower DRAM frequency from BIOS (1333MHz for example) and see if it works. Try running test programs like MEMTEST on every single module and then on pairs to try and detect any incompatibility when matching 2 of them.
Also ask a friend with compatible motherboard to test your RAM sticks. It might be a problem of your motherboard with particular RAM modules and the motherboard and RAM modules could work OK when separated.

Well, my old set of 4 sticks worked perfectly fine, prior to installing my new loop and also a new PSU. Available watts for sure isn't a limiting factor, its the Corsair HX1000i PSU... So plenty avalible aha there. But after that was when the problem arose. In the past the last slot has always seemed a little bit temperamental, but it kind of comes normal to me, as when first installing new sticks, I've frequently had to reseat them in any machine I've worked on. But this time, reseating and extra pressure just isn't quite cutting it, sadly.

The motherboard manual says it supports OC'ing RAM up to something ridiculous like 3200MHz. The board defaults the settings to 1333MHz with a voltage of the standard 1.5v. Elsewhere online, people have had theirs set to 1.65v so I gave that a shot, but still no luck.

After installing the third stick, it won't even POST. I guess I could do memtest just to check, but I can't see how all 8 sticks that I have would be affected by the same issue? Considering its the last 2 slots that never work, despite what stick is placed in there.

Also there shouldn't be incompatibilities, they were brought at the exact same time, and I haven't been mixing the old cheap set with the Vengeance ones. They appear to be from the same batch too as the serial numbers are within 10 away from the other pair. And likewise, I tested each stick in a slot that is for sure working and it booted past POST successfully every time. Only when they're placed in the last 2 slots, I never get POST successfully.

Thanks for the reply too, I did notice I wrote a rather long essay haha. Just thought I'd try to cut out as much Q&A as possible within the first post.

 

[Tradesman1]

Might be the cooler is too tight in a corner or side of the CPU so that the pins aren't getting full contact which can throw off the CPU's MC (memory controller). Try loosening the cooler then snug it back down.

 

[napster100]

Might be the cooler is too tight in a corner or side of the CPU so that the pins aren't getting full contact which can throw off the CPU's MC (memory controller). Try loosening the cooler then snug it back down.

Got each side quite firmly down... Wanted to get that snug fit, hoping for optimal heat transfer, surprisingly working well. But this is something I did not think of! I'll give it a shot and see what happens. Thanks.

 

[Tradesman1]

No worries, happens quite often 😉

 

[napster100]

No worries, happens quite often 😉

Didn't want to loosen it too much, I maybe able to go a bit further just don't want a frying CPU aha. Well, no changes after about 5 full turns... It has springs under the screws too just an fyi if that would make much more difference? I'm assuming the applied pressure to the CPU would be less because of that?

Well, temps still seem the same, so I may drop a few more turns and see how it goes... Just a thought, would it be possible that I've permanently damaged the MC? I hope not If so, thats going to be a costly misfortune, I brought a EST wristband to save the hell of frying any hardware (mainly RAM). Just my luck haha. 🙁

 

[Tradesman1]

Shouldn't No, big scare might be if too, too tight that it might bend pins. Just want it about screw tight and may want to reset CMOS then check the DRAM again

 

[napster100]

Shouldn't No, big scare might be if too, too tight that it might bend pins. Just want it about screw tight and may want to reset CMOS then check the DRAM again

Dude!, you was right! It was too tight! I unscrewed all threads, and the turned them 5 times each and now it boots with all the RAM sticks installed!!

Thanks sooo much man! You have probably saved me money haha.

On idle, my max and average CPU temp has dropped by 1C too haha.

 

[holyprof]

Overtightening the cooler can indeed make the thermal contact worse due to bending. Glad your problem was solved and kudos to Tradesman1.

 

[Tradesman1]

Shouldn't No, big scare might be if too, too tight that it might bend pins. Just want it about screw tight and may want to reset CMOS then check the DRAM again

Dude!, you was right! It was too tight! I unscrewed all threads, and the turned them 5 times each and now it boots with all the RAM sticks installed!!

Thanks sooo much man! You have probably saved me money haha.

On idle, my max and average CPU temp has dropped by 1C too haha.

__________

I'll send you a bill 😉 Just kidding, glad all is well, have some fun now !