Jump to content
Corsair Logo Corsair Logo
Corsair Community
CORSAIR COMMUNITY Survey, we want to hear your feedback!

DX48BT2, X38/X48 and Corsair DDR3

NicholasDX48BT2

By NicholasDX48BT2
in Memory

Recommended Posts

Dave England798

Dave England798

Posted
Posted
I have tried all settings using all the recommended settings and also the settings quoted by your many forum users i still cant get this memory to run at the correct speed. it is not too much of an issue as i can still use my pc at 1066mhz it is just a nuisance to know i paid for memory that will not run at the speed it is supposed to. i have done extensive research into this problem and it seems to me either one of 2 things either the memory or the motherboard is goosed and since i dont have a replacement of either to further test my theory it is summat i will have to live with any further help would be greatly appreciated since the memory is quoted as being able to run at the correct speed in this motherboard
Link to comment
Share on other sites
  • Replies 81
  • Created
  • Last Reply
mlhearn

mlhearn

Posted
Posted

RAM Guy,

 

Thanks for the recommendation to test with memtest.org. I had originally tested with the memory test supplied with OS and found no issues and was suspecting the motherboard or incompatibility.

 

Per your suggestion I tested each module (labeled MEM1 and MEM2) one at a time with memtest.org latest test. In addition, I tested each channel of the memory controller to see if errors moved with the memory. I was still suspicious of MB. The only setting that had to be manually set was voltage to 1.7V, everything else was correct including frequency (though in one place it reports 1332Mhz versus 1333Mhz).

 

MEM 1, Channel A - 36 test passes, ~12 hours, one single bit error

MEM 1, Channel B - 34 test passes, 11 hours, 0 errors

MEM 2, Channel A - 3 passes, 1 hour lock up, 30K+ errors

MEM 2, Channel B - 9 passes, 3 hours, 200K+ errors

 

So MEM 2 experienced 100,000s of errors. MEM 1 experienced one single bit error in 24 hours. Since these are a matched pair of DIMMs, should I send both back?

 

Please advice,

mlhearn

 

mlhearn please set the tested settings and test the modules one at a time with memtest.org and let's make sure one is not failing.

Link to comment
Share on other sites
NicholasDX48BT2

NicholasDX48BT2

Posted
  • Author
Posted

All -- yes, there are issues with the DX48BT2. The conclusion is that Intel released a board that has some bugs/issues in some cases, and is not always compatible in a "plug-and-play/bug-free" manner with all configurations that it lists as compatible. Which is not what people expect from Intel (and is the reason often people would choose Intel over other brands, to avoid those kind of issues, the "beta-tester" feeling).

 

Some users have no issues. Some have minor bugs (which some people ignored or did not notice and classified as "no issues"). Some have severe issues. Also note that people that post reviews saying that it's all good when they've used the system lightly for only 3 days, or after they see an issue they make a setting change and use the computer one evening and do not see a BSOD again or crash and say it's OK, is not good. People have found BSOD and severe errors the first time after a week of running and traced it to incompatibility, settings issue, hardware issue, software/driver bug. OK to me means no crash, no BSOD, no boot issue/bug, etc, for at least > 1-2 weeks of use, and also > 8-12 hrs of Memtest with each module. (I'm only talking about non-overclocked boards here, not overclocking). There are also tests like Prime etc that stress the CPU and cache, etc, that people do when overclocking.

 

In my case, I pointed out two bugs to Intel, that they have independently confirmed (Intel itself duplicated the bug on their own PC), and are in the process of investigating (it's taking a long time), and in particular the unsucessful boot POST issue that some have experienced (I'm NOT talking about the case where the message pops everytime if you have a problem with settings or failed hardware). I also had in my specific case a bad memory module that I have replaced. I have an occasional BSOD that is linked to the ATI display driver that has been experienced by many many people with Vista in general and not specifically my confirguration and it's very very unfrequent so it may take me a while to track and may just be another Vista/ATI bug.

 

Also, as shown above, several hours of Memtest are needed before saying that "the memory is OK".

 

RamGuy and mlhearn - does this mean that both your modules are defective, or could it be the board itself? More precisely, which slots did you check? How come the memory 1 tests OK for 11 hours but on another test it has one error? Does that mean that really we should test memory for super super long (like 24 hours?) to ensure that it is fine, since you had a 50% chance of not seeing the error with your first module over a long 11 hour test? Would that mean also that your memory 2 which has tons of bugs made your system crash, and memory 1 has less bugs so it's more stable but still has issues?

 

My understanding is no matter how long the test is, there should be NO errors ever? I tested my new memory modules 8 hours only, so now I'm wondering...Does it hurt to test for like 24 hours?? It's stressing the components...

Link to comment
Share on other sites
boba303

boba303

Posted
Posted

Hi Nichloas

 

I've been reading you thread earlier but now I thought I could give some input regarding mems showing good test at some time and 2-3 errors at another.

 

I got the problem that running memtest with Dual channel mode I can get errors around ~56 000. When running them apart from each other, one gives more errors then the other. Follow link to the thread I currently having in case you find this relevant.

 

http://www.asktheramguy.com/v3/showthread.php?t=71399

 

Im using TWIN3X2048-1600C7DHXIN memory's on a MSI X48 Platinum board. (Just see specs for more info)

Link to comment
Share on other sites
  • Corsair Employee
RAM GUY

RAM GUY

Posted
  • Corsair Employee
Posted

I cannot comment about Intel and their problems they may have other than to say I have not seen problems like you are describing when using their MB's.

And the reason I suggested replacing the memory to take that out of the equation. I would suggest we go ahead and replace the modules and if you still have the same problem then I would get the MB replaced.

Link to comment
Share on other sites
mlhearn

mlhearn

Posted
Posted

RAMGUY,

 

I've started the RMA process and will send both DIMMs back.

 

Like NicholasDX48BT2, I also purchased an Intel MB with the expectation that it was more likley to be fully tested and stable. Especially for a none tweaked/not OC application. I'm still not sure if that's the case, but I was surprised how quickly Intel has offered to replace the MB. I'm waiting on the new memory and hope that will work.

 

NicholasDX48BT2 raises a good point, how many memory errors are acceptable and over what period of time? Especially since I'm running without ECC. And, should ECC support be the default going forward on performance builds?

 

Thanks Mike (mlhearn)

Link to comment
Share on other sites
Wired

Wired

Posted
Posted
how many memory errors are acceptable and over what period of time?
None. A low amount of errors COULD infer that it's a setting issue, but not always.

 

And, should ECC support be the default going forward on performance builds?
Haven't seen ECC memory since DDR1, and "ECC" wasn't listed on NewEgg's Power Search for memory.
Link to comment
Share on other sites
  • Corsair Employee
RAM GUY

RAM GUY

Posted
  • Corsair Employee
Posted
For testing memory 2-3 passes will be enough but for testing the system to ensure it will be stable I would suggest 12-24 hours and yes you should not get any errors, but say for example you got one error after 20 hours I would not be too worried but would check for a possible thermal issue.
Link to comment
Share on other sites
NicholasDX48BT2

NicholasDX48BT2

Posted
  • Author
Posted
For testing memory 2-3 passes will be enough but for testing the system to ensure it will be stable I would suggest 12-24 hours and yes you should not get any errors, but say for example you got one error after 20 hours I would not be too worried but would check for a possible thermal issue.

 

Thanks a lot! If I test both sticks together for 2-3 passes, is it also good enough? (as opposed to separately, individually)

For my new memory, I tested 2 passes with each module, and then 11 passes (8 hours) both together in the mobo -- good enough?

 

For those having the boot issue, I posted an update or "solution" from Intel there:

http://www.houseofhelp.com/forums/showthread.php?p=361313&posted=1#post361313

Link to comment
Share on other sites
Rodger

Rodger

Posted
Posted

Hi there, I definitely think it's a memory problem. My config is a DX48BT2 + Q9300 + 4x OCZ 2GB 1333MHz 7-7-7 anx I experience the same problems which al seem related to the memory. At first only one module worked. After BIOS update 1, 2 and 3 modules allowed POST. 4 modules results in 3 beeps = memory problems. I tried installing Vista x64 with 2 modules. Result sudden reboots at random. Even tried with one, still random reboots. Then I saw the voltage not being correctly set in the BIOS. 1.5V in stead of 1.8 which OCZ states is the correct voltage for those modules. So I manually switched to the 1.82 setting in the latest BIOS.

 

I am really dissapointed in the memory support from Intel on this board. If you want to go DDR3, you want to go as fast as possible. Let's hope they will fix the memory support in their next BIOS.

 

Monday I will contact our Intel technician to ask if he has a beta BIOS for me or if he is willing to check my config. I work for a large system builder in Holland and I have some technical contact.

Link to comment
Share on other sites
Rodger

Rodger

Posted
Posted

By increasing voltage to 1.82 I managed to install Vista x64. After that I found things not running as smooth as it could be. I noticed a lot of hard memory errors in windows taskmanager, so I relaxed timings a bit. No result. After that I increased the voltage again to 1.90. This time definitely way better results. From a stunning +/- 8000 hard memory errors a minute to +/- 20 a minute.

 

So try to play with voltage a bit.

Link to comment
Share on other sites
  • 2 weeks later...
mlhearn

mlhearn

Posted
Posted

RAMguy,

 

I received the RMA replacement memory and have repeated the earlier memory tests. Tested both DIMMs as a pair for 48 hours straight. Not a single error was reported, yea! I'll be testing the system for stability over the next few days, but things now work much better.

 

Thanks,

mlhearn (Mike)

Link to comment
Share on other sites
mamow

mamow

Posted
Posted

So I RMAd the boards that were failing the memtest. Ran memtest on both sets of 2 overnight with no error in either. No random restarts.

 

Put the four together, however, and I'm back to unstable. Not as bad as before, but the machine restarts overnight invariably. Sometimes during the day, too.

 

Any suggested next steps before I ramp down to slower memory? Can't afford to be worried about restarts.

Link to comment
Share on other sites

Archived

This topic is now archived and is closed to further replies.

Go to topic listing
×
×
  • Create New...