agreed^^
6-8-6=PSC's
Printable View
agreed^^
6-8-6=PSC's
they should give you 2k stuff or 1866 stuff upgrade or = not downgradeQuote:
Originally Posted by LightSpeed
i received my reapers with hypers from a 2000mhz plat stuff rma
ocz has a bad habit of downgrading us dont let them
I have now had to return my Blade ocz3b2000c7v6gk twice. (costing 50gbp a time might I add) They are offering me in exchange the new blade kit which is 2000mhz cas8. only problem is my original purchase was for 370gbp and the new kit is just short of 180gbp.
Should one just put up and shut up?
edit. as of today they have now just offered me reapers 2133 cas9!!! I figure thats still 2000mhz cas8 ram. I am at a bit of loss as of what to do?
take the one that you like the look better they both have same icsQuote:
Originally Posted by Enzo78
i'd take the reapers :D
be glad ocz making ram again lol
they been offerin 1600mhz stuff for 2000mhz rmas :rofl:
Think that we wonīt get Hyper back from RMAs from most of the Dram houses is a fact we have to accept as the ICs are EOL and therefore hard to get, even for the manufacturers- was the same with D9s at DDR2, I was lucky to get D9s back at two RMAs(ebay buys that were defect) but at the other rmas I made for friends I got other ICs, performing worse- but always were the best the manufacturers could offer as replacement
New Blades = 2000 8-9-8 = non-Hyper.Quote:
Originally Posted by Enzo78
I also wouldn't get too excited about the Reaper 2133 cas9...
OCZ has no hypers left- no matter which kit you take, you wonīt get Hypers-thatīs a fact- normally same with Corsair as I heard and a lot of other DRam houses
i'd go for new blade 2000 8-9-8, they aren't as good as your old blade elpida hyper but they are better than nothing...
it's a pita that now - except for STT i think - we have only PSC crappy ICs...
p.s. luckily my a-data 2133X MNH-E and HyperX 2000 MGH-E still work like the first day
i still see other brands other than stt selling hypers...
Thatīs true- but only a few- in europe you randomly get G-Skill 2200 Perfect storm, you get the remains of Corsair 2000 C8-8-8(2x2GB and 3x2GB) and a lot of 2200 C8 A-Data- thatīs all, except from a few STT 2200 C8 and with luck 2000 C8- maybe at your place the offer is better:D- but doesnīt change facts on RMAs, and especially in this case with OCZ:(
is there any info on these http://www.ocztechnology.com/product...e-channel.html or am I still better off getting the blade st's?
odds are the 2k cas8 will do 2133 cas9...Quote:
Originally Posted by Enzo78
i personally like how the reapers heatpipes work
my hypers with reapers run much cooler that my other ST hypers
and if you rma again wont get severely downgraded lol
don't know in usa but here - in europe - i can find:Quote:
Originally Posted by Hondacity
g.skill perfect storm 2x2 gb - 260 euro - http://compare.eu/?in=&fs=gskill+2200
a-data 2200 2x2 gb - 230 euro - http://compare.eu/?in=&fs=adata+2200
corsair 2000 3x2 and 2x2 gb - 360 and 200 euro - http://compare.eu/?in=&fs=corsair+2000
and some supertalent - http://compare.eu/?cat=ramddr3&asuch=supertalent
it's clear that if you want a good 3x2 gb kit the best thing to do is buy STT from usa unless you wont spend at least 400 euro :D
STT kit 3x2 gb 2000 cas8 or 2133 cas8 from usa is about 210 euro shippment included, if you have to pay the fee it will cost 250/260 euro but you still save at least 140 euro.
Believe it or not guys Eplida is in the news again attempting to produce the thinner version DIMM. We'll see if Corsair, and OCZ will stick their necks out for this one this time. If they do, I suspect heavy testing for longevity will ensue before their released for sale again.:rofl:
about the 30nm?Quote:
Originally Posted by firerx
supposedly 23nmQuote:
Originally Posted by Hondacity
23nm makes it thinner?...lmao
DDR3-1600 6-8-6 to replace DDR3-2000 MHz? Wow, that's quite pathetic.Quote:
Originally Posted by LightSpeed
if it was 2000 999... then that sounds good.. if it was 2000 888... then its... bad... but if it was 2000 777 or 787 then thats really unacceptable...Quote:
Originally Posted by CryptiK
cas7Quote:
Originally Posted by saaya
if it was me i wouldnt accept it
ocz gettin pathetic in rma
Looks like my STT 2200 C8s are degrading... I've only had them for a month and a half or so, been running them at 2000mhz 7-7-7-20-64/1T at 1.66v, 1.25v VTT. Shame...
symptoms?Quote:
Originally Posted by Kain665
What was 100 passes LinX stable last week can't pass a single pass of LinX now. I'm testing further to make sure it is the ram.
Test with memtest in windows to rule out CPU and chipset instability.
It's not that. I know it's ram related, I'm trying to figure out whether it's IMC related or RAM - raising neither the vTT nor the vDimm helps. It seems to be running fine at 2000mhz at XMP timings - 8-8-8-24-with a TRFC of 88, and I was able to drop my vTT down to 1.225v. Interesting indeed...
Yep, it's ram I set it to run memtest 86 for a while, looked fine for a pass or two, left it running, came back on pass 7 and had 7000 errors all from test 5. This is from 1600mhz 8-8-8-24-88-1T with 1.65v. Need to get that RMA going, I'm not sure if I want another set of these sticks if they're just going to die - what's the best PSC's got out nowadays?
2500s of a lot of Manufacturers if you donīt go for GTX- I have my STT 2200C8 since 4 month- no problems;)- but I also never lost a ramstick til now
We'll see. Depends on what kit they ship me, I'm looking at Malabs' website (STalent is the memory division of MaLabs), and they have both the 2200 C8 in stock, and I guess a left over kit of Project X 2000 C7, I may see if they can send me that instead.
Yeah IIRC I've had them for 3 months now, running 1.66v. Maybe that's what killed them?
Good luck- hope you get a good replacement that serves you well- and longer;)
P.S. Donīt know- mybe- but EVGAs Boards are notorious(some of them at least) for overvolting- maybe yours does that as well- hard to say from the distance, can also be that you just ran out of luck and have a glass kit^^
Still, 1.67v? So PSC chips are more durable? Is there much bandwidth difference in having TRCD a bit higher? What are some other characteristics of PSC I should know about?
Higher TRCD, higher TRFC- and I am not sure that PSC are more durable if you feed them with 1,67V 24/7- maybe someone else can tell about that, and I think we are getting a little bit OT:rofl:
Hopefully I don't jinx it, but I think switching the DIMMs around and cleaning the contacts worked for me, suprisingly... There was nothing visible on the contacts, maybe some sort of residue? Prime Blend at 2000mhz 7-7-7-20-64-1T, Tfaw at 20 as per before with 1.66v.
Hehe only 5 minutes prime so far, but hey, it's better than before where they used to error out instantly :D
You've gotta be kidding me... The second I posted that one thread crapped out... :rofl:
Alright, I think I got it stable with a bios update and 1.67v (1.664 -1.665 Real) vDimm at 7-7-7-20-64-1T, I think the whole problem, was some bios issues. Also, I think EVGA solved the overvolting problem with the new A72 bios (I used A70 previously), so my voltage needs to be set at 1.67v. Testing more ATM.
Sounds good and is great news if the ram stands the tests- so not another victim for the graveyard:D
Alright, I'm stumped. Last night, I set my ram to 7-7-7-20-64-1T (and the rest of the subtimings that I had before), and it passed 45 passes of LinX, whereas last week it could do 100 passes no problem. However, This morning, with the same settings it has trouble passing 1 pass. I'm fairly sure it's the ram, because as you loosen the timings, I get more stability. Adding vTT/vCore doesn't help.
weird behavior here
have you set the RTLs manually?
when after reboot i'm unstable 9/10 is because i've the RTLs in auto and the mobo change it.
never ever had in my life a memory that can do a frequency stable and after a reboot goes completly unstable with the same bios settings.
No reboots, I left it doing LinX overnight, came back this morning and tried it again. Timings are the same as in my sig (click on the cpu).
the hypers...are not good for 2000mhz daily...fast degradation...Quote:
Originally Posted by Kain665
so you runned linx for a night, when you get up without reebot you have retried to run it but it gave errors?Quote:
Originally Posted by Kain665
if this is the case it means that linx sucks badly. :D
for memory stability check i would use memtest in windows, for cpu stability check i would use prime95.
+1Quote:
Originally Posted by Hondacity
and you don't need them for DU...you can't see the difference, except with few specific programs, between 2000 cas7 and 1333 cas9.
hypers are for benching purpose, for DU if you ask me they are time and cash waste.
I got a really good deal on them, I get them straight from S.Talent. CPU passes Prime SmallFFT with flying colors, The ram passed 10 passes of memtest86 3.2 at 2000mhz C7 :shrug: I'm going to go fish out another set of ram to test and make sure that it is, in fact my memory and not something else, lol.
i'm talking about this memtest http://hcidesign.com/memtest/download.html
bootable memtest using X58 - never tryed with 1156 but i think that it's the same - gives me strange results.
i can pass it at a frequency that gives error in less of 5 minutes of prime95.
Yeah, that's what I've heard. Well regardless, I decided to drop it down to 1600mhz for longevity, running 6-6-6-18-52-1T at 1.55v with a minimal drop in bandwidth. I may still pick up some of the remaining higher binned Project X 2000 7-7-7-21 though, I'm curious to see what that can do at lower voltages.
Update! After having it stable at 8-8-8-20-64-1T at 2000mhz and 1.6v, I'm starting to see signs of instability, just over 20 days down the line. A 124 here and there, and Prime Blend errors with a specific rounding error (Don't see those very often, lol) after a few hours of Blend + games. I guess I need to RMA, after all... :(
Kaine what kit was that? please dont tell me it was the new revision elpida Super talent speed :S 2200 cas 8... or ill scream if mine degrade.. i dnt think i can RMa from canada to the states....Quote:
Originally Posted by Kain665
almost feeling like going for those new Gskill eco's everyone is speaking highly about... im soo scared that i spent 300$$ 220+100 shipping/importing fees to canada :shakes: never again if these fly into the garbage...
Yeah they are the 2200 C8 sticks. I ran them at 1.66v though for 3 months.
Anyone knows A-Data XPG Plus 1866+ CL8 what chips do they use?MGH-E or MNH-E?Mems are this ones:
http://hardverapro.hu/dl/upc/2010-07...6cl8-2-phw.jpg
The 1866+ A-DATA, are with MGH-E, equipped.
so whats the deal guys i really dont get it, sorry to ask, first elpida hypers died bad, and where replaced by the MGE-H wich dont suppose to die unless you strike them with bad volts...i run my 2200 C8 Stalent kit 2x2 at 1,66 daily as it was needed on P45 at 1600 CL6, now on I7 they need same for 1720 C7 stable and 1,68 or bit more for 1720 C6, real volts i mean
1720 7-7-7-20 1T 1,66 real
linx is solid can run it many times passes always this is my 24/7
http://img18.imageshack.us/img18/671...linxx5pass.jpg
1720 6-6-6-19 1T 1,68 real
linx goes random passes then some loops then no then again yes so need tweaks but i dont want to go over 1,68
http://img825.imageshack.us/img825/2...linxx5pass.jpg
cheers guys :toast:
Sergio
Put a powerfull fan over them dimms. These hypers chips love to be cool. I can get a few extra MHz out of mine at the same voltage with a good fan over them.
i have fans pointing to them i dont like over as heat gets trapped inside...but mainly im asking why they say new hypers also die...its supposed not at reasonable volts...Quote:
Originally Posted by Zeus
I don't think it's volt related, most old Hypers seemed to have died at stock volts. It's just Hyper's nature to die out of nowhere i guess.
weird tought they fix that on MGE-H chips :shrug:Quote:
Originally Posted by Zeus
Hi,
my question might be misplaced, but does anybody know if all DDR3 Corsair Rev2.1 are equipped with Hyper? I have bought a 1600 C8 Kit that is rev 2.1 and wonder if these might be Hyper, for I wanted something else for testing purposes
Yeah, I think so, Corsair used 2.1 for Hypers and 2.2 for BBSE as far as I know.
How come you bought some Corsair memory? :D
Aaargh- more Hypers:eek:- I already have a closet full of them:rofl:- I just wanted to test some Corsair so I can rant on a higher level about them:p:- if they work like the GMH on DDR2 @ Corsair I had I might be able to rant about the rams nonetheless:ROTF:
Ahhh, that makes sense! :rofl:
I'm looking forward to seeing your results! :p:
Sudden death :DQuote:
Originally Posted by Zeus
But there is a small Polish manufacturer which hasn't got any dead Hyper yet:
http://translate.google.com/translat...o_2000mhz_cl7_
Strange? Or simply well designed.
I actually wonder if its a PCB flaw rather than a IC flaw?
How many chips now are put into PCBs designed by people other than the company who makes the ICs?
i'm sure its ic relatedQuote:
Originally Posted by Panther_Seraphi
i know than if my stix get hot they get unstable
i try keepin my stix below 28c even on load
does hyper still better than current 40nm ICs?
depends... from what i heard, most people seem to think so...Quote:
Originally Posted by PatRaceTin
Does Elpida Hyper MGH-E 2GBx2 will run great on 1155 SandyBridge System?
(i decide to buy one kit to use with new system)
does someone have this ocz memory kit? http://preisvergleich.hardwareluxx.de/a420456.html
I could get it very cheap, seller said those are hyper ics...but i cant rly believe it...good bbse or pcs should also do this.
Does anybody know whats underneath this memory ? (black pcb seller said)
thanks in advance :-)
its verry possibleQuote:
Originally Posted by SuicidePhoenix
if it can run 1600+ 6-6-6 its 100%
my 1866 reapers are the best hypers i ever tested
even if good bbse still worth it
black pcb odds went higher
but has it degraded
I have my dead Elpida hyper: G.SKILL Perfect Storm. (PC3-16000 CL8 Triple)
I used it on AMD 890FX board (ASUS CH3F) running at 2000MHz / 7-7-8-21 or 1600MHz / 6-6-5-15.
Timingwise it's undoubtedly the best while the OC headroom is not that roomy... Anyway it's gone last summer.
1.) After how much time did your memory fail? hours? weeks? months? -> less than a month. I bought it July 2010 and sent it for RMA August 2010.
2.) Did you stress it a lot? (memtest/orthos/gaming) -> (is Super PI stressful? then yes) I never run memtest / prime / linx - and similar torture program.
3.) How did the memory die? Idle, load, reboot? -> freeze and hardly boot
4.) What exactly happened? freeze, errors, crash, system turned off, system rebooted, blue screen, black screen -> freeze and sudden power-off.
5.) What board, pcb rev and bios version did you use? -> Crosshair 4 formula, sorry but i can't remember then version of bios.
6.) What speed and timings did you run the memory at when it died? -> 2000MHz / 7-7-8-21
7.) what speed and timings did you run the memory at most of the time? -> equally 1600MHz 6-6-5-15 & 2000MHz 7-7-8-21 (before it died, it was running at latter status)
6.) What vtt did you run when the memory died? -> 1.7V
8.) what vtt did you run the memory at most of the time? -> 1.75V
At a glance I only see these dying on Intel, is it possibly memory controller related?
Higher frequencies pushed with tight timings at exuberant voltages :p ( for the bench rams that failed) Most Hyper kits barely reached 1950Mhz on AMD... going over 2200 CL7 on Intel...
Have a double kit, and it come`s with Micron KPT`s. Black PCB.Quote:
Originally Posted by SuicidePhoenix
Really not bad stuff, but C6 is a no-go. Early triple kit though, serial ending with 000000000035 and 00000000036
Except that scanning through this thread, most failed at stock DDR1600 or DDR1866?Quote:
Originally Posted by Leeghoofd
Why not buy regular kits then ? Don't always believe what you read :p Who bought 500 dollar rams to run them sub specs ? at least I wouldn't :p Either they pushed timings very tight, usually went wild with voltages or high speeds... and tested the crap out of them with LinX, Prime95 and memtest sortlikes...
I lost my OCZ BLades and the Kingston rams... never pushed above 1.65Vdimm; but timings were tight at a higher frequency (both lost on I7) Till now holding my breath with the GTX2's...
Mine is running 1.78v with World Community Grid and GPUGrid hammering the RAM 24/7 for a few months now without so much as degradation :shrug: I don't know about there, but in South African my Hypers were the cheapest sticks I could find above DDR1333 :p:
Quote:
Originally Posted by Alex-Ro
Just saw one of those kits, here, and it was equipped with MNH-E chips.Quote:
Originally Posted by Ramguy
hi,
i want to buy some rams with hypers, but i am afraid of this hyper problem...
i have started to read the thread from the beginning, but sorry it is to much to find the correct information to buy second hand hypers in a safe way.
so, can someone give me a conclusion, some advises which facts i have to take care ? e.g. batches, dates, rev., versions, pcb, brands and so on ?
sorry for my english, i am german and thank you for help !
greetings H.
Me too.Quote:
Originally Posted by Hirogen
I have just bought these mems.:
A-Data XPG Plus V2.0 1866+ CL8 (4*2GB).
They should be Hypers. Should I give them back immediately?
No point in reading this thread from the beginning since it was started in 2009., back in the day when MNH-E was king of the hill.
Elpida was aware of the issue and released the updated version of the chips, MGH-E, lower clocks compared to predecessor, but increased
long time reliability. It is safe to use them, just make sure those are newer batches with newer chips and you'll be fine.
Thanks for your reply man. So is it confirmed that MGH-E don't die? And how can I be sure my A-Data would mount MGH-E and not MNH-E without removing the HS at least?Quote:
Originally Posted by Erklat
Thanks.
Goodbye!
All ICs can die :p: but MGH-E less than MNH-E however you can't be sure without remove heatsink.Quote:
Originally Posted by Punisher!
Non puoi saperlo senza rimuovere il dissipatore, in ogni caso ad oggi dubito ci siano in commercio kit "nuovi" con MNH-E
submited