Samsung generic HCH9 dbl. sided 2X4GB

[Don_Dan]

Most likely they're newer 10XX or 11XX batches BFBG.

Yes, they're 1100 batch.

...and you never will be, either.

Why you giving me a hard time...haven't you been taught to respect your elders?

Oh, I'm sorry. It's time for a hug!

[aussie-revhead]

Hey, I joined the Dumo generic club and have 4x4gb HCH0 on the way for my z79 rig, what should I try for? I know I'm not going to get much above 2400 with my 3930k so can I get any decent timings from these? Or 2400-10-12-12 at decent volts?

[AVulgarPicture]

Probably only one way to find out hey.

[Shiranui Gen-An]

I stated playing with my Corsair Vengeance 2133 C11 sticks with HCH9 ICs finally, so far they don't want to do any tight timings at 1200MHz with 1.65v. I was hoping I would get some good overclocking headroom since they're default 1.5v, but I'll keep plugging away at them.

[Shiranui Gen-An]

Well I can definitely say that the generic ones the OP has are much better than my Corsair HCH9, these things won't do any decent timings at DDR3-2400 with 1.65v. I've tried 11-11-11, 10-12-12, heck even 11-12-12.

[K404]

Are they 4GB sticks? What's your tRFC and tRAS?

[Shiranui Gen-An]

Yes, they're double-sided 4GB sticks. tRAS I usually set to 30. tRFC is set to AUTO and is usually over 200 although if I use the Rampage Tweaks it gets set much tighter, 120 in some cases.

[aussie-revhead]

Guys I have four 1147 batch here, what timings should I aim for with x79 + 3930K for 2400 24/7 use?

I have 4x4 Team 2400-9-11-11, benched 2450c9 but not prime stable even at 2400. Looser timings didn't seem to help, I think my imc is struggling, xmp 2133 seems much more stable with four sticks in.

Testing now ....

Edit: first boot successful at 2400-10-12-12-28-1T 1.65v but not 32m stable, second boot 2400-11-12-12-31-1T passes 32m, passed again at same timings 1.60v. Should I try cas10 again with tras 31? Memtweakit is blank so what prog do I use to save 1000 reboots?

Edit: Memtweakit needed update ... priming now at 2400-10-12-12-28-1T 1.62v

[Zeus]

Just ordered 2x4GB of this stuff. Real cheap and most of all, available in my own country! Will post my results here.

[Zeus]

The sticks came in today but instead of HCH9 chips they had HCK0 chips. Even though this looked promising as they seemed to be the 1.5V variant of the famous HYK0 chips, they were not good. At all!

Tried 2600 10-12-12-28 up to 1.75V code 55. Everything over 2500 was always code 55, no matter the timings or voltage. They could do 2400 9-12-12-25 at 1.7V but don't know how stable that was, it passed memtest 86 test 5 for 3 loops at least.

Tried the sticks individual but non booted at 2600 cas 10 or 11.

If you see them, skip 'm.

[Viss]

Think mine are the exact same Eus, wil check tonight. nvm, these are HC9:

[Reefa_Madness]

Zeus,

From your picks, the latter part# of your label reads "CH0-CK0", but the ones that Dumo has, per his pic in post #35, end in "DH0-CH9". Those are the parts # that you need to look for, not just generic 4GB Samsung.

That said, your sticks seem to do about as well as those that aussie-revhead posted just above you, and about as well as the single-sided GSkill Trident X 2400 C10 sticks that list for about $90.

Granted, they aren't doing what Dumo's sticks have done, but let's face it, Samsung ICs seem to have a crush on him. That or he sprinkles pixie dust on them prior to firing up his rig.

[Zeus]

I guess HCH9 is better. Just tried 32M at 2400 9-12-12-25 110-1T but i wasn't faster than my 4x4 HYK0. These HCK0 sticks can do TRRSR =4 though which gives a nice boost in read bandwidth but it hardly makes any difference in 32M.

Your sticks are week 6-12 Bart, mine are week 26-12. Seems they changed over to HCK0 now.

Reefa_Madness, i think it has to do with the chips itself, the serial number last digits are derived from this i guess. K0 means 1600 cas 11 binned by Samsung where H9 means 1333 Cas 9.

[ZoLKoRn]

Wow amazing clock with price Dumo and thanks for sharing man

[Reefa_Madness]

Reefa_Madness, i think it has to do with the chips itself, the serial number last digits are derived from this i guess. K0 means 1600 cas 11 binned by Samsung where H9 means 1333 Cas 9.

Yes, I know that it has to do with the ICs themselves (it always starts with the ICs)...that was my point...that the "DH0-CH9" modules are the ones sporting the HCH9 chips that are the better clockers. The modules that you bought were a different parts number and hence, were sporting different ICs.

[Zeus]

Fortunately i can return them without extra charge. The sticks were advertised as the part number same as Dumo has, the ones i got have different part numbers so the store accepted rma.

I just want some 1300+ action!

[Shiranui Gen-An]

Aren't they all the same ICs, just binned differently? Like Micron D9GCT/GMH/GKX?

[Reefa_Madness]

Fortunately i can return them without extra charge. The sticks were advertised as the part number same as Dumo has, the ones i got have different part numbers so the store accepted rma.

I just want some 1300+ action!

Well, that's good that you could return them and only fair, after all, they weren't what you intended to purchase, parts # wise.

[Dumo]

Samsung 4Gb M378B5273DH0-CH9 batch production 11XX~1152

[Reefa_Madness]

Perhap, but it is those attributes that allowed them to be binned as "HCH9" which result in the higher OC potential.

I would suspect that it is very similar to the DDR1 days and Samsung's own TCCD vs TCC5. The TCCDs were all TCC5s that had undergone further binning by Samsung for specific attributes. After Samsung decided to stop binning for TCCD the manufacturers had to start the binning of TCC5s on their own, but it was the same IC. Within those TCC5s were plenty of ICs with "TCCD like" qualities, but not all of them. Even within the TCCDs were some that could run tight timings, but not nearly as high clocks and some that were not capable of keeping the CL2 timing very long, but could scale very well and reach the DDR600+ range. I'm sure you guys all remember the wide range of parts available from OCZ, Mushkin and others, all based on the same TCCD part.

These HCK0s may be the equivalent of the TCC5s...binned to a certain point, but not binned as far as the HCH9s are. If that is the case, then there could very well be some gems in those HCK0s, but you just have to look thru more of them to find them. Starting out with the HCH9s just gives you a leg up on the "binning-generics-for-gems" game.

[Zeus]

Talking about TCCD brings back old memories of 300+ 2,5-3-3-7. Man, those were the days. Recently going through my old HDD i found some screens from back then. I think i had 316 2.5-3-3-7 running 32M iirc.

I don't have a good feeling about the HCK0's as non of both sticks did boot at 2600, not even at cas 11 and 1.7V. All four of my HYK0 sticks boot at 2600, just one is actually stable but boot they will.

Pretty happy i can return them, hopefully they have some HCH9 sticks left but i don't think so.

[Dumo]

Even HCH9 needs binning, most can do 2400 C10 but not all can do 2600~2800 with any settings/voltages

[Reefa_Madness]

Even HCH9 needs binning, most can do 2400 C10 but not all can do 2600~2800 with any settings/voltages

Without a doubt...after all, it is a CL9 DDR3-1333 part to begin with.

I still think that you have some pixie dust that you use on yours.

[Shiranui Gen-An]

I'd say OnePageBook is the guy with the pixie dust!

http://www.ocxtreme.org/forumenus/showthread.php?t=6481

[Shiranui Gen-An]

Perhap, but it is those attributes that allowed them to be binned as "HCH9" which result in the higher OC potential.

Didn't catch that the sticks he got were C-die. I guess what I was trying to say was all K4B2G0846D ICs are the same just with different binning by Samsung and the overclockability on average I'd imagine to be pretty close with some golden chips of every bin and some duds as well.