RMA request bad RAM

[eswimm]

I've been getting a lot of random system crashes, so I decided to test my memory. SDRAM Timing is set to automatic [2.0-3-2-6].

 

The 1st module only seems to have one intermittant single bit error, but the 2nd module reports dozens of errors. The modules were purchased as a TwinX matched pair, so I don't know if both need to be replaced or not. The BIOS on the Intel board doesn't allow a lot of fiddling with memory settings.

I swapped DIMMs around in slots and tested them one at a time, to rule out a bad slot.

 

Intel D875PBZ

Intel Pentium 4 2.8/800 HT

BFG nVidia 6800GT OC

SB AUdigy 2 ZS

 

Memtest86+ v1.40

 

Settings: RAM 200 Mhz (DDR400) / CAS : 2-3-2-6 / Single Channel (64 bits)

 

3 passes per run

 

Module 1 - CMX512-3200LLPRO

 

Slot: Channel A DIMM 0

 

Tst Pass Failing Address Good Bad Err-Bits Count Chan

--- ---- ----------------------- -------- -------- -------- ----- ----

8 0 00005c26fe4 - 92.3MB ffffffff fffffeff 00000100 2

 

 

Slot: Channel B DIMM 0

 

Tst Pass Failing Address Good Bad Err-Bits Count Chan

--- ---- ----------------------- -------- -------- -------- ----- ----

8 0 00005c26fe4 - 92.3MB ffffffff fffffeff 00000100 1

 

Module 2 - CMX512-3200LLPRO

 

Slot: Channel A DIMM 0

 

Tst Pass Failing Address Good Bad Err-Bits Count Chan

--- ---- ----------------------- -------- -------- -------- ----- ----

2 0 0001ea6a3c8 - 490.6MB ffffffff ffefffff 00100000 6

6 0 0000cd623c8 - 205.1MB fffffff7 ffeffff7 00100000 1

6 0 0000f4623c8 - 244.1MB fffffff7 ffeffff7 00100000 1

6 0 0000fd623c8 - 253.1MB fffffff7 ffeffff7 00100000 1

6 0 0001e06a3c8 - 480.6MB fffffff7 ffeffff7 00100000 1

6 0 0001ea6a3c8 - 490.6MB fffffff7 ffeffff7 00100000 1

6 0 0001f1623c8 - 497.1MB fffffff7 ffeffff7 00100000 1

6 0 0001f56a3c8 - 501.6MB fffffff7 ffeffff7 00100000 1

6 0 0001ea6a3c8 - 490.6MB ffffffef ff3fffef 00100000 2

6 0 0001f46a3c8 - 500.6MB ffffbfff ffefbfff 00100000 1

3 1 0001ea6a3c8 - 490.6MB fefefefe feeefefe 00100000 2

... and many, many more

 

Slot: Channel B DIMM 0

 

Tst Pass Failing Address Good Bad Err-Bits Count Chan

--- ---- ----------------------- -------- -------- -------- ----- ----

3 0 0001e86a3c8 - 488.6MB fdfdfdfd fdedfdfd 00100000 1

6 0 0001ee623c8 - 494.1MB ffffffef ffefffef 00100000 1

6 0 0001f9623c8 - 505.1MB ffffffef ffefffef 00100000 1

6 0 0001ea6a3c8 - 490.6MB ffffffbf ffefffbf 00100000 2

6 0 0001f56a3c8 - 501.6MB ffffffbf ffefffbf 00100000 1

... and many, many more

[RAM GUY]

First thing, if you are getting errors with both modules that would suggest some other problem. In addition, our Pro modules are tested 3 times at their rated speed on a MB before they leave our factory, so it would be hard to get a bad module let alone 2. Please make sure that you have the latest bios and then please set the Dim Voltage to 2.775 Volts and test the modules one at a time with http://www.memtest.org. If you still get errors, I would suggest testing them on another MB that will allow you to set the timings and voltage manually

[eswimm]

Latest BIOS for MB, motherboard is an Intel D875PBZ, it doesn't support changing the DIMM voltages, unless there's some way other than the BIOS.

 

Only 1 DIMM is giving significant errors, regardless of slot. The other DIMM only gave 1 error during 3 passes of the tests.

 

I've tested with both Memtest v3.2 and Memtest+ v1.40, both gave the same results.

 

While I'm sure it's incredibly unlikely to get bad PRO modules, each module generates the set of errors regardless of which slot it is in.

 

Module 1 generates an intermittant error at 00005c26fe4 in each of the 4 memory slots.

 

Module 2 generates a whole list of errors in the top 64MB with addresses ending in xxxxx3c8.

 

The results I attached the first time were to show that each module was getting the same errors no matter where they were installed, not that 2 modules were getting the same errors.

 

I don't have another DDR400 machine to test these DIMMs in, but I have tested DDR333 DIMMs in this machine without error.

[RAM GUY]

Based on that, I would see if a friend can maybe help you test the modules on another MB. In addition, if you have the latest revision of D875PBZ with bios P27 or newer you should be able to set the Dim Voltage. You might check with Intel to be sure. Also if you do have an older Rev of that MB I would try and set the timings manually to Cass 2-3-3-6 or Cass 2.5-3-3-7 and see if it will pass http://www.memtest.org.

[eswimm]

I'm running BIOS P33, I have an older rev. board which has fixed memory voltage at 2.6V.

 

I lowered CAS to 2.5-3-3-6 and the DIMM which was reporting only 1 intermittent error passed without error (at least in the few tests I ran).

 

The 2nd DIMM, which was reporting numerous errors continued to error at the slower timings, in every slot.

 

It's not the board. I've got a pair of 512MB 3200XLPRO modules on order for this machine and I didn't plan on returning the 3200LLPRO chips for replacement until the XLs came in anyway. I assume when I get the XLs and they test out properly without errors, I won't have any trouble getting an RMA for the faulty LLs?

[RAM GUY]

Well we can sure try replacing the modules for you. But XMS3200XL would not be suggested for that MB. The Minimum voltage for XMS3200XL is 2.75 Volts. You can try them with this MB but I would set the timings to Cass 3-3-3-8.

Please send us a email with a copy of the form or all of your info name address and phone# and the Module part# and copy the link to this post and email it to Warranty@corsairmemory.com. Or you can also use the On Line RMA Request Form as well.

If after 1 day or 24 hours excluding weekends you do not get the RMA please email the same to rma@corsairmemory.com and we will help to resolve it.