DDR3 IC thread

[Shiranui Gen-An]

That's pretty much how all the Asus Z87 boards behave when they don't like something about the memory config. They don't auto recover either, you have to do a hard power down. Asrock and Gigabyte are a lot better in that regard.

[ZenEffect]

Made no difference.
I've already tried to loosen up all of the timings, it changes nothing.

The way it shuts down, it is almost like the board is hitting some sort of hardware protection (OVP/OCP).

The CPU should be fine as it is a brand new part, never molested.
Surely no IMC can be this bad, can it?

you have already verified that the imc is capable of more on cfr (if i am reading correctly)

sounds like the board is having issues training the memory and something that is not exposed to end users is getting set way too tight. if you have some samsungs, i'd pop them in just to see what its doing with different ICs. Really sounds like the motherboard is not cooperating as i have never seen a haswell cpu that bad ever. im not going to suggest "try this or that" because it seems you have already tried everything and i know how annoying it can get to be told to try something you already did. if you do switch boards, try the asrock oc mini, nick has been pretty good about getting the products with his name on them compatible across a bunch of different types and brands of memory, and its relatively cheap... and the boards are efficient as hell to boot.

[Shiranui Gen-An]

Looks like that board doesn't have latency boundary settings, which makes it easier to set all of the tertiary timings to loose values. I would suggest setting DRAM CLK Period to 14.

[The Stilt]

Looks like that board doesn't have latency boundary settings, which makes it easier to set all of the tertiary timings to loose values. I would suggest setting DRAM CLK Period to 14.

Made no difference either.

It is some sort of a compability issue.
After some tinkering I was able to run a CFR kit at DDR-2800 so the hardware itself is good.

Decided to try something different and plugged in one Samsung Q module and one CFR module.
The system posted right away at 2666MHz. However after making some changes to the settings and saving them I could not get it running again.

With a CFR module plugged in the board does not shut down as fast as it does with the single Samsung module.
Go and figure.

[Shiranui Gen-An]

I wonder if these Ver2.2 Dominator 1600C9 are the same BBBG ICs as my ADATA 2000X. Below are the max clocks/volts I've been able to reach on both.

ADATA:


Corsair:


According to Sam, Ver2.2=BBxx. Same clocks, similar timings, similar VDIMM. Hmm....

[Reefa_Madness]

Given what you've just posted makes me wonder if that's what was also being used on those Avexir 2000 C9 kits that were available from Mwave last April.

http://www.xtremesystems.org/forums/...=1#post5182497

[Shiranui Gen-An]

I think those were BDBG.

[Reefa_Madness]

Is there an easy place to get the physical IC measurements of the various Elpida offerings?

[Shiranui Gen-An]

The best I've found is on this old Japanese PC blog:
http://blog.ark-pc.jp/?eid=104312#sequel

[Reefa_Madness]

Thanks for the link.

[smsmasters]

Does anyone know what ICs are used in these? http://www.overclockers.co.uk/showpr...odid=MY-042-TG

[Shiranui Gen-An]

Could be lots of things. Hynix MFR, Samsung 4Gbit B-die, even some kind of Micron IC.

[smsmasters]

Could be lots of things. Hynix MFR, Samsung 4Gbit B-die, even some kind of Micron IC.

Is it possible to tell from the TRFC?

[TaPaKaH]

I don't think so. Lowest both MFR and Samsung B-die can go would be around 260-280, so you won't get a clear IC indication from tRFC alone.

[Shiranui Gen-An]

So I finished binning all of my BBSE sticks, using Sam's method.



Geil Evo 1600C7 Stick #1 > No boot, Bunny Extraction
Geil Evo 1600C7 Stick #2 > Not stable any voltage, Bunny Extraction
Corsair Dominator GT 1600C7 stick #1 > no boot, Bunny Extraction
Corsair Dominator GT 1600C7 stick #2 > 1.64v 32M stable PASS
Corsair Dominator GT 1600C7 stick #3 > 1.685v 32M stable PASS
G.Skill Ripjaws 2000C9 stick #1 > no boot, Bunny Extraction
G.Skill Ripjaws 2000C9 stick #2 > no boot, Bunny Extraction
G.Skill Trident 2000C9 stick #1 > 1.705v 32M stable PASS
G.Skill Trident 2000C9 stick #2 > 1.75v 32M stable PASS
G.Skill Perfect Storm 2133C9 stick #1 > Code 55, Bunny Extraction
G.Skill Perfect Storm 2133C9 stick #2 > Code 55, Bunny Extraction
G.Skill Perfect Storm 2133C9 stick #3 > Code 55, Bunny Extraction

Pretty atrocious. The biggest surprise/disappointment was the Perfect Storm sticks. None of the three would POST at 2200 7-9-7 regardless of voltage used, I would get a hard Code 55. The other failures would at least attempt to boot and a few could at least get into Windows.

[TaPaKaH]

To be honest, this is not unexpected. BBSE vary in quality a lot, only ICs on 2133C8 and 2200C9 RipjawsX are consistently interesting, elsewhere you have to be lucky.

[Shiranui Gen-An]

OCZ Flex II DDR3-1600 C8 2x2GB OCZ3FXT1600LV4GK PSC XE

[Don_Dan]

Nice find! Have you tested them yet?

[Shiranui Gen-An]

Nice find! Have you tested them yet?

They're decent, not great but not bad. They can pass the PSC tight 2400 preset for Z77 OC Formula with 1.775v BIOS set.

[Shiranui Gen-An]

Haven't tried for 2600 yet.

[darkage]

hi
does this Corsair dominator kits worth any Money (can buy 2)
TR3X6G1600C8D / 8-8-8-24 1.65V R2.1
thanks for any help
for use with RIVE and or CHVF

[websmile]

This would be a risky purchase if you buy without any results, most of the non GT Dominators rev2.1 used bad clocking ICs by Elpida like BASE which failed 1000 at any Timings configuration. It is very doubtful these will give you good clocks you search for, but not impossible. Personally, I wouldn?t buy unless these are really cheap, which Corsair usually are not ^^

[darkage]

yes they are cheap and i could get around 12 sticks (4 3 chanell kits) but to binn all of them it will get expensive ; -)
thanks for your help

[Splave]

anyone else have AE-F BBSE :P cant find any info out about it

[TaPaKaH]

I guess it's the same IC with a different JEDEC bin to DJ-F.