6gb memory installed, 4gb working

[briligg]

hi

i have been having memory related problems since i bought my new core i7 system, an i7 920 on an asus rampage ii gene mb with 6gb ocz 1600 mhz ram (currently running at 1066mhz) (ocz3p1600lv6gk). with some effort, i got it running stably, but starting about 3 weeks ago the system stopped seeing all three dimm sticks.

bios and windows 7 x64 see 4gb ram. all three sticks check out in memtest86+ from boot. i have removed the cpu and check the mobo's pins, and they seem fine. i tried swapping the dimms into different slots, but swapping c1 to a2 caused the system not to post, b1 to b2 caused only 2gb to be visible, and then i gave up on that line of thinking, and spent some time getting my 4gb back, which reappeared after i removed and replaced the cpu, and took c1 back out and put it in again.

this memory is not on asus's approved list for this board, which they claim is why it is misbehaving. the store says others who bought the same package deal as i did haven't had this problem. getting new ram from the approved list would be quite expensive where i am, and i'm not even sure it would solve the problem.

any input?

 

[darkguset]

The fact that swapping the (same i guess) sticks in the slots gives you different results could indeed mean that this memory has problems with the specific motherboard. I would also think that Asus is probably right this time and your store wants to make a sale and that is why it is claiming that it works fine. Or it could indeed run fine in some cases, but that does not make it the rule and that is why ASUS does not have it in the approved list.

 

[briligg]

there is an opinion i have seen two or three different places that bent pins on the mobo's contacts for the cpu can cause this problem, but talking to our techie today, he couldn't see how that could be true. how could bent pins cause such a specific, and actually rather trivial, problem? wouldn't the whole system be screwy if there was bad contact with the cpu? and which pins are supposed to be involved? surely if the pins were to blame, one could say very clearly which ones, no? but even the store i bought the system from suggested faulty pins could be the problem.

this problem appeared shortly after i replaced the psu. the old one failed, our techie said it was probably struggling to maintain voltage on this system, because it wasn't intended for a machine so powerful. the system just stopped posting one day. once the new psu was installed, the techie said he had quite a struggle to get the system stable again. a few days after that, when i was adjusting bios parameters for the ram - it causes random bsods if the back to back cas# delay isn't set at 8 or higher - i noticed there was only 4gb of ram.

 

[darkguset]

Regarding the CPU bent pins, it is true that it could cause ANY kind of issues, from specific ones to more general driving you pulling your hair out one by one, but to tell you the truth i highly doubt that is the cause. It is very simple though to check, just remove the heatsink then the CPU, get a small torch and look at the pins on the mobo carefully.

You didn't mention the PSU earlier. What sort of PSU did you have before and what do you have now? If it does indeed not provide enough power to the system then it is true that all sorts of problems could arise (including bad memory detection). It could also be (more far fetched) that when the older PSU died it killed one or more of the memory modules. When i say killed i do not necessarily mean completely. They could be damaged in many ways therefore behaving strange.

 

[briligg]

I did remove the cpu and check that and couldn't see any problem. I have wondered if the old psu caused the problem, but so far asking around in forums hasn't elicited any comments in that regard. The old psu came with an Acteck Kendall case - it's a Mexican brand, which is where i'm living. It was 500W, but it was a generic brand, and the case was a mini-tower, poorly cooled - I modified it and then chucked it. The new psu is a 460W CoolerMaster Elite. I remembered it wrong before. Thinking back now, it's that our techie struggled to get the system stable while the old psu was still in it. And I should also confess a complete bone-headed act on my part, in which I plugged in the power leads to a dvd drive backwards, fried it and the psu, which was then repaired. (Generally in Mexico things often get repaired that would get replaced in a more moneyed country, and I believe this is a case in point.) But the modules themselves check out. The techie wonders if one of the sockets is damaged, and suggested i check that, beyond that I now plan to take the system in to him in a week or so when i have a chance.

 

[j12david]

I actually had the same problem yesterday (and the past 2 weeks) after I installed Windows 7 and did a bios update (not sure when it changed from 6 to 4). Except i'm using Corsair 1600 mhz, and a DFI lanparty-dk-t3he6. The only improvement i've actually gotten was when I was told my mobo's manual was written wrong, and that another 3 slots were actually triple channel. After I switched my ram over to triple channel I was able to pick up another .5gb in windows, and my bios recognizes all 6gb now.

 

[briligg]

J12david, i don't know what to say about that. The manual says to use a1 b1 c1 for 3 dimm sticks, which is what i've done. But, i opened up cpu-z this afternoon, and what it's telling me has me totally confused. It says my memory is dual-channel, and that i have sticks in slots #1 and #5. I assume that with my sticks in a1 b1 c1, it should be saying that slots 1, 2, and 3 are occupied.

I did some testing of the slots yesterday. I don't know how to do that, and i couldn't find a guide, so i just kept swapping the sticks around and recording what bios saw. I was advised that in some cases, sticks will perform differently in different slots, but i think that is only really relevant in overclocking instances. anyhow, after trying a bunch of different combinations, it seemed that slot b1 isn't working. the sticks were all checked with memtest86+ when this first cropped up, so it isn't a bad stick. but now, with cpu-z reporting slot 1 and 5 occupied, i don't know what to think about that result. it's weird, isn't it? it makes me think of bad pins again, but i really couldn't see anything wrong with the pins.

i replaced my aftermarket heatsink with the stock one, in case there was an uneven pressure on the pins that was causing bad contact. i was advised elsewhere that could be a cause. but there is no change. and now i have another bit of weirdness - when i boot into bios, my cpu heats up immediately. i've done it twice this afternoon, and both times the cpu temp had already risen more than 10c by the time bios loaded, jumping from about 45c - 50c to 63c. then, within a minute in bios, cpu temp had risen another 6c to 69c. and then i quit bios.

so yeah, i've been thinking just to take the whole system in to the techie, but really, i would really like to know what is going on and try to fix it myself, because at this point i just don't know if the techie is going to grok this.

what's going on?? any advice??