Ballistix Performance Memory

[cpm73]

I'm hoping someone can advise me with a problem I have.
My system has been working fine until recently and now it will crash/segfault about 1-2 times a day (running Linux 24hrs a day).

I've got 2 x 2GB sticks of Crucial Ballistix (BL25564UK80A) DDR2 800MHz memory.

After yet another crash the other day, I decided to run Memtest86+ v2.11 on it right after rebooting and it was throwing up loads of errors so I shut it down and decided to test each stick separately.
The first stick looked healthier until the Random Number Sequence test which raised some errors, but nowhere near as many as when I first tested with both stick installed.
I then tested the 2nd stick and it passed with no faults.
I decided to re-test the first stick again and this time it passed too.

So I'm a little baffled. Either its the rest of my system thats at fault, or the memory was overheating perhaps and had cooled down with me opening the case and swapping bits around to do the testing.
After the memory passed, I restarted the system. Each time it comes up fine, but usually within 24 hrs it'll crash again.

I've luckily got another system with 2 x 1GB sticks of slower DDR2 memory (not Crucial memory) that I could switch over to using, and with that memory installed, my system has been up and running quite happily for the last 3-4 days under basically the same loads it was experiencing before.
I switched everything back again last night and it crashed in the night yet again.

So it looks like its the memory to me and under heavy load its failing, but its hard to be conclusive given it will pass the Memtests eventually.
If I send them back as faulty - the chances are they'll pass any RMA testing and get sent right back to me.

Has anyone got any other ideas as to what the cause might be or what I can do?


Thanks for any help in advance

Chris

[Crucial_maulmod]

Hi Chris,

If the modules are passing Memtest then it could be a set up issue rather than faulty parts.

The first thing to do is check that you have the latest version of the BIOS for the motherboard from the manufacturer's website.  Then ensure that the following timings are set within the BIOS:

Speed: PC2-6400 (800MHz)

Timings: 4-4-4-12

Voltage: 2.0v

Try the modules individually with the above timings, in normal use and also through Memtest again.  If both modules pass, try them together again with these timings.

Please also ensure you have appropriate case airflow to help cooling within the system.

If this does not help, please confirm your motherboard details in your reply and also let us know if anything else hasn't recently changed within your system.

Thanks.

[cpm73]

Thanks for the reply.

The motherboard I'm using is an ASRock A780GM-LE with an AMD X2 5050e processor.
My case has pretty good airflow, but my board only has the two RAM slots so the memory sticks are right up against each other which restricts the air on one side of each stick.

There hasn't been any changes to the kit since it was setup around May/June this year and like I say - its been working fine until now. I had recently downloaded the latest kernel update which I originally thought may have introduced the stability problem, but I re-booted under the original kernel and still had a crash, so I assumed a hardware fault of some sort. The fact that I saw some Memtest errors seemed to also suggest that fact. I suppose I'm focusing on the memory as the cause at the moment purely because of the Memtest errors and the lack of a problem when I swapped in the 2 x 1GB Elixr DDR2 533MHz sticks from another machine.

I've just upgraded the BIOS to the latest version as per your suggestion and I've attempted to set the correct values in the BIOS although I notice you say its speed should be set to 800MHz, but my BIOS only lets me select a max setting of "400MHz (DDR2 800)". Is that going to be a problem or does the two sticks in dual channel mode operate at 800MHz? Also, my "DRAM Voltage" setting doesn't have 2.0v exactly, so I've set it to 2.02v which I think is the nearest.

I had previously tried setting the timings and speed but not the voltage (couldn't find any details on it at the time) and it still crashed so I'll see how it goes this time around.

The system starts OK as per usual and I'll run some similar loads on it over the next day or two and see if its more stable.

Chris

[cpm73]

Although my system didn't crash overnight, it did raise a few segfaults in other services that were running, so it was well on its way to falling over and Memtest once again showed lots of errors.

I now suspect this is an overheating problem as I discovered today one of my case fans had stopped working.
I have a 4U rack mount case with 2 x 120mm front fans and 2 x 80mm back fans and it was one of the rear fans that had stopped.
Combined with my two RAM modules being right next to each other, its quite possible they were not getting cooled enough.

I've now swapped in another fan from a similar case, which although doesn't seem to be as fast with quite the throughput, it will have to do until my replacement arrives.
Hopefully this fix will resolve my stability issues.

[Jody21]

Think about the facts. Running good for a long time, then crash, crash, crash. Other memory seems to work fine. No config problems here, The memory is failing. Replace it. If your ambiant is less than 104F your cooling your case fine. That doenst mean the memory isnt overheating.

[Marauder]

I've been experiencing the same issue. 4 x 2G Modules (BL25664AA80).

They had been running fine for about five months and now they are randomly generating errors. If I shut down the system and test each module with memtest86, they all pass. I had 4G (4x1G) of Corsair XMS memory in the system for several years without any issues.

It seems that recently a number of people on NewEgg are reporting similar issues (often in similar time frames) with the 2G DDR2 Ballistix memory which makes me believe it is more of a faulty design then an isolated incident or two of improper ventalation or something else environmental like that.

http://www.newegg.com/Product/ProductReview.aspx?Item=N82E16820148211