Unstable system with Ripjaws Z Series 64GB (8 x 8GB) F3-12800CL10Q2-64GBZL

[Tanquen]

Not most but all, DIMMs 1 through 5 all had errors in Prime95 then I upped the voltages and tested DIMMs 5-8 and then back to DIMM 3 and no more errors. I went back to DIMM 3 because I had run it 4 nights and it always had errors at stock. If I put the voltages back to stock I’m think they will all have errors. I only had the one DIMM from a friend but it had no errors running at its stock settings. I don’t know anyone with a 2011 CPU.

I did look over the CPU socket as I’d seen some stuff online about bad 2011 pins killing CPUs and the motherboard makers telling people they must have damaged them. I feel confident there is no compound passed the heat spreader.

Also, I was able to run Prime95 with all 8 DIMMs and not get errors at first. The first thing I did after getting the system together was run Prime95 for two nights and a run of Memtest86+. It also makes me wonder because there are three posts (two verified owners) on Newegg that say the DIMMs start to fail after a month or so. Then there is one more that simple says it was bad and he returned it. I thought maybe they did not know what they were doing or were crazy but now I’m thinking they were onto something. At the same time, in 20+ years and many builds I’ve only seen bad RAM like two times.

[Tradesman]

I'm the same, started building back w/ the XT era, and have had very few chips, SIMMs, or DIMMs that were bad, and to have a few bad in a set is really strange, which is why I lean towards mobo, CPU, PSU or GPU, any possibility of swapping some component with somebody to test? Im sure GSkill would have no problem with an RMA, but would hate to see you lose a week-to ten days, get new sticks and the same old problem

[Tanquen]

So the fact that all 8 seemed to work before and that a friend’s DDR3-1600 DIMM seems to run ok and that others seem to have a failing over time issue with this kit doesn't bother you? It reminds me of guys that OC there CPU and its stable for a time but then they need to up the voltage, like the CPU is slowly degrading. Trust me, if I had access to other components I’d swap them. I have other power supplies but this is the only one I have with 2 extra (I think) 8 pin connections for the motherboard.

[Tradesman]

Was just offering ideas, by all means, as I said I have no doubt that there would be no problem with GSkill and doing an RMA, just trying to offer alternatives that might save some time. Can contact them at ram@gskillusa.com or rma@gskill.com

[Tanquen]

Ok, I got a Kingston (KHX1600C9D3K8/32X) DDR3-1600 8x4GB kit yesterday to test with. Right now I'm at 14 hours running Prime95 with no errors with all 8 DIMMs installed and with the BIOS at all its default settings with only the XMP profile selected. The timings are different but again like my friends 2GB DDR3-1600 DIMM the DRAM voltage is 1.65V vs. 1.5V.

The night before last I was able to do a 10 hour run with DIMM number 8 at default settings and it passed so maybe it is just a few of the DIMMs that require the higher voltages. After I?ve let the Kingston kit run as long as I can I stand I?ll got back to DIMM number 3 at stock and see if it errors again.

Maybe its just so random that with one DIMM overnight is not enough. But the last time I tried all 8 DIMMs I had the first Prime95 error at just over one hour and then many after.

Took out the extra "passed!" rows.

[Fri Apr 13 22:48:36 2012]
Self-test 448K passed!
...
[Fri Apr 13 23:04:07 2012]
Self-test 8K passed!
...
[Fri Apr 13 23:19:08 2012]
Self-test 512K passed!
...
[Fri Apr 13 23:34:59 2012]
Self-test 12K passed!
...
[Fri Apr 13 23:50:18 2012]
Self-test 576K passed!
...
[Sat Apr 14 00:05:18 2012]
Self-test 18K passed!
FATAL ERROR: Rounding was 0.5, expected less than 0.4
Hardware failure detected, consult stress.txt file.
Self-test 18K passed!
...
[Sat Apr 14 00:21:30 2012]
Self-test 672K passed!
...
FATAL ERROR: Rounding was 0.4997711182, expected less than 0.4
Hardware failure detected, consult stress.txt file.
Self-test 672K passed!
[Sat Apr 14 00:36:38 2012]
Self-test 21K passed!
...
[Sat Apr 14 00:51:38 2012]
Self-test 768K passed!
...
[Sat Apr 14 01:06:39 2012]
Self-test 25K passed!
...
[Sat Apr 14 01:22:28 2012]
Self-test 864K passed!
...
[Sat Apr 14 01:28:28 2012]
FATAL ERROR: Rounding was 0.5, expected less than 0.4
Hardware failure detected, consult stress.txt file.
[Sat Apr 14 01:37:44 2012]
Self-test 32K passed!
...
[Sat Apr 14 01:52:51 2012]
Self-test 960K passed!
...
[Sat Apr 14 02:08:26 2012]
Self-test 36K passed!
...
[Sat Apr 14 02:24:11 2012]
Self-test 1120K passed!
...
[Sat Apr 14 02:34:23 2012]
FATAL ERROR: Rounding was 0.5, expected less than 0.4
Hardware failure detected, consult stress.txt file.
Self-test 48K passed!
[Sat Apr 14 02:40:07 2012]
Self-test 48K passed!
...
[Sat Apr 14 02:54:41 2012]
Self-test 1200K passed!
...
FATAL ERROR: Rounding was 0.4983215332, expected less than 0.4
Hardware failure detected, consult stress.txt file.
Self-test 1200K passed!
...
[Sat Apr 14 03:09:58 2012]
Self-test 60K passed!
...
[Sat Apr 14 03:25:22 2012]
Self-test 1344K passed!
...
[Sat Apr 14 03:33:03 2012]
FATAL ERROR: Rounding was 0.5, expected less than 0.4
Hardware failure detected, consult stress.txt file.
[Sat Apr 14 03:40:23 2012]
Self-test 72K passed!
...
[Sat Apr 14 03:45:39 2012]
Self-test 72K passed!
...
[Sat Apr 14 03:56:11 2012]
Self-test 1536K passed!
...
[Sat Apr 14 04:11:39 2012]
Self-test 84K passed!
...
[Sat Apr 14 04:16:45 2012]
Self-test 84K passed!
[Sat Apr 14 04:26:55 2012]
Self-test 1728K passed!
...
[Sat Apr 14 04:32:02 2012]
Self-test 1728K passed!
[Sat Apr 14 04:42:04 2012]
Self-test 100K passed!
...
[Sat Apr 14 04:47:28 2012]
Self-test 100K passed!
...
[Sat Apr 14 04:57:34 2012]
Self-test 1920K passed!
...
[Sat Apr 14 05:02:49 2012]
Self-test 1920K passed!
Self-test 1920K passed!
FATAL ERROR: Rounding was 0.4999961853, expected less than 0.4
Hardware failure detected, consult stress.txt file.
[Sat Apr 14 05:13:29 2012]
Self-test 120K passed!
...
[Sat Apr 14 05:18:34 2012]
Self-test 120K passed!
[Sat Apr 14 05:28:30 2012]
Self-test 2240K passed!
...
FATAL ERROR: Rounding was 0.5, expected less than 0.4
Hardware failure detected, consult stress.txt file.
Self-test 2240K passed!
[Sat Apr 14 05:33:36 2012]
Self-test 2240K passed!
[Sat Apr 14 05:43:33 2012]
Self-test 140K passed!
...
[Sat Apr 14 05:49:00 2012]
Self-test 140K passed!
...
[Sat Apr 14 05:59:09 2012]
Self-test 2400K passed!
...
FATAL ERROR: Rounding was 0.5, expected less than 0.4
Hardware failure detected, consult stress.txt file.
[Sat Apr 14 06:04:19 2012]
Self-test 2400K passed!
...
[Sat Apr 14 06:16:22 2012]
Self-test 160K passed!
...
[Sat Apr 14 06:31:36 2012]
Self-test 2688K passed!
...
[Sat Apr 14 06:47:05 2012]
Self-test 192K passed!
...
[Sat Apr 14 07:02:26 2012]
Self-test 2880K passed!
...
[Sat Apr 14 07:17:58 2012]
Self-test 224K passed!
...
[Sat Apr 14 07:33:26 2012]
Self-test 3200K passed!
...
[Sat Apr 14 07:48:53 2012]
Self-test 256K passed!
...

[Tanquen]

I did a clean install of Windows 7 on a spare hard drive with no updates or drivers. I was worried that the SSD drive may have been causing/adding to the instability. I only installed the latest version of Prime95. With all 8 Kingston 4GB DIMMs installed I did a few overnight runs and one that went for 28 hours with no errors. I then put all 8 of the 8GB GSkill DIMMs in and I had two threads error within 30 minutes. I reset the BIOS each time only selecting the XMP profile.

I’m going to do the RMA now. I hope it works.

[Tradesman]

Thanks for the update, keep us informed

[Tanquen]

That is odd. My email to ram@gskillusa.com was not delivered because I don’t have an account?


Sorry, we were unable to deliver your message to the following address.

<ram@gskillusa.com>:
Remote host said: 554 delivery error: dd This user doesn't have a gskillusa.com account (ram@gskillusa.com) [0] - mta1009.biz.mail.sk1.yahoo.com [BODY]

[Tradesman]

It should be RMA@ you show RAM@ must have been thinking about the sticks

[Tanquen]

Oops, I got it from your earlier post. That's what I get for copying and pasting. :P Thanks.

[Tanquen]

Ouch, looking at the invoice for the RMA I just saw that they went from $699.99 to $495.00 on Newegg. It’s kind of depressing when the price drops that much before you can even get a new setup working.

[Tradesman]

My fault on the email, sorry, get going to fast sometimes and my fingers don't do what my brain says to do......and I know the feeling on the DRAM and price, I've often purchased stuff and seen the price drop before it even gets delivered......and even more often, I'll make an order and it will be on a good sale the next day

[Tanquen]

Update time!

I am now 13 hours into the first Prime95 run with the replacement GSkill kit. I did a BIOS reset and selected the XMP profile and have had no errors yet. It seems like there was something different about the original kit. It also seems like it changed over time because I know I did a Prime95 run when I first got it and had no errors. By the time I set it back I was consistently getting errors (with all 8 DIMMs) within the first 30 minutes or so. This is kind of a scary precedent as I’ve never had RAM fail over time before, not in the first few weeks anyway. The fact that others on Newegg had the same thing happen makes me bit uneasy.

Is there a way to find out what was wrong with the original kit? Just wondering if they have some super RAM tester hardware that can say: Yes, something was off with this kit.

Unfortunately I’m still getting random lockups. Even though I did not get Prime95 errors with the Kingston kit I still had four or five lockups. I even got a second video card to test with and I still had the lockups. That just leaves the motherboard, CPU and power supply. I went ahead and got an ASUS motherboard to test with but I’ll see what happens with the new GSkill kit first.

My worst build… ever!

[Tradesman]

Well keep us updated, it's sounding better, might be something in the advanced timings, tRFC or something. Know what you are going through am playing with a Z77/3570K and 32 GB 2133/9 and having trouble taking it over 4.4

[Tanquen]

But I'm not even trying to OC.