So how do you know it's UTT? my 2x512mb pc3200 was $105 @ newegg and its A4AT and does ~235 memtest stable on 2-2-2-5 with 3.3v
Printable View
So how do you know it's UTT? my 2x512mb pc3200 was $105 @ newegg and its A4AT and does ~235 memtest stable on 2-2-2-5 with 3.3v
If you have AA4T, then it is the CH5-UTT. Since you are able to to 235, 2-2-2-5, pretty sure you have it.Quote:
Originally Posted by DamienKC
well, this is my result in DFI Infinity NF2
only can run SuperPI 1 M :( with vdimm @ 3.7V. But there is no difference between 3.6V and 3.7V because i think my system was reach the limit.
Twinmos SP AA4T 04484 KO6633 512x2
http://i4.photobucket.com/albums/y10...d/super256.jpg
That Kingston HyperX3200 _is_ the NEW BH5. Pop the covers and you'll see new BH5, not the old Kingston relabeled/packaged BH5 chips...D328DW45's.
Everyone is getting on the bandwagon of the new production run.
What I havent seen is 2x512M of the NEW BH5 run through the NForce2 ringer to see if the CPC on/off issue still remains. If they are just more BH5 production then they probably do still suffer from it.
The question becomes... is the new UTT-CH5-VX the better choice for maximum performance (even though it requires high Vdimm) or can new BH5 at lower Vdimm equal the UTT-CH5-VX stuff? Dunno.
Quote:
Originally Posted by uwackme
Are your claiming that those are new bh die UTT on the KHX at ZZF as opposed to just some old inventory that was located somewhere? I'm not asking this question to argue with you, but just to try to get more information. I've not seen Kingston claim to have a new bh based product available. Did I miss this or are you just speculating that its UTT? Do you have a link to a news release or a product description for new Kingston bh ram? I really think it is just old stock that turned up somewhere. And not that you stated all KHX was relabeled, but for what its worth, all of my KHX PC3200 is original bh-5.
Keep in mind that Mushkin has been selling original bh-5 in their Mushkin Blue dual-packs. They claim that in a recent order Winbond sent them a box of original bh-5, along with new bh die UTT. They used these, as well as some stock that they had for RMA, to produce a batch with original bh-5, then they produced the rest of the Blue dual-packs with the new bh die. This information was posted by one of their reps on several sites.
Again, at $240 per gig there are many other choices to consider. I just saw the earlier post from Bennah asking if the stuff he linked to at the egg was old school bh-5, so I posted that link to ZZF.
As far as I know Kingston was cleaned out of BH5, and didnt even have RMA BH5's....ie: if you had to RMA a BH5 stick you werent going to get back a BH5 replacement.
So I hardly think some miraculously appeared....meanwhile NEW BH5 has come into the world and suddenly some clearly BH5 spec'd sticks show up... If I were a betting man, I'd bet they were new BH5 based sticks.
Moreover, a quick peak under the heatsink will confirm, the code numbers on the chips will have new date codes. I doubt they would have arranged again for "custom" packaging from Winbond like they did in the past....BH5 HyperX came in a custom TSOP with fewer pins (the unused pins for beyond 32Mx8 chips were deleted from the stock JEDEC TSOP).
But Im not wasting my money buying a pair to confirm. Im sure someone will buy a pair and maybe we'll find out then,
More importantly.... noone has done a NEW BH5 2x512M thorough test to find out if any "improvements" has been made. If not, UTT/VX/newCH5 is a better stick to be using for all NForce2 applications... since 1T operation will work, while BH5 rrely if ever did 1T on the NForce2's. At $54/512M the Twinmos remains an obscenely good deal.
As Reefa_Madness said, Winbond had just some month ago sent Mushkin a shipment of da old BH-5.Quote:
Originally Posted by uwackme
Maybe Kingston too got a shipment.
What RAM is better TwinMos UTT WInbond or the new TwinMos 14AT (BH-5)?
UTT Winbond is UnTesTed. This means that if it is Winbond is either BH-5 or CH-5. I heard that lately CH-5 is clocking much better then the new BH-5.Quote:
Originally Posted by xavicl
By the way...just saw on another site where one of the guys posted that Mushkin is coming out with a "Redline" named set of PC3200 and PC3500 that are supposed to do 2-2-2 timings...more binned bh UTT perhaps?Quote:
Originally Posted by uwackme
Edit:
Never mind about Redline...just saw where he started a thread here as well.
o just saw u reply my post :p:Quote:
Originally Posted by 4X4N
here is my result:
twinmos UTT 2x256
260 fsb 1:1 3.7vdimm
got alot of errors...damn it
So its better Winbond UTT AA4T than 14AT?
My Twinmos UTT at 3.2 volts 2 x 512mb 2-2-2-5-7-15-1-2-1-1 @ DDR 500
No probs!
Will max it out and post CPU-Z and Pi fast etc latter
Edit DDR 510 only 3.2 volts DFI
My Twinmos TMSP AA4T from newegg.
250@1.5-2-2-5 with 3.4v :D
http://img213.echo.cx/img213/9389/tim0ro.jpg
i had a kit that did 260 2-2-2, died on me within 4 days, then a kit that did 240, one stick has died on me. Just complete failure, just bang and the comp wouldn't turn on. Disappointed to say the least, but i will still keep dumping my money into it to find a pair that won't die! :p
Skip, what kind of voltage and cooling on those that died?
Do you think i should go for a pair of
Twinmos Speed Premium pc3200 - confirmed new BH-5
instead of Samsung TCCC or UCCC(single sided)?
Some says the new CH5 is better than the new BH5 but
then another says thats because of trouble with the old
nf2 mb. Anyone with comparable info regarding the new nf4?
[I´m building a system with DFI ultra-D nf4 and AMD64 3000+]
Is it the DFI that is killing Rams ?
I had my twinmos on 3.8-4 volt burn in for some weeks and its still ok.
Have you read this thread?
"Very important, read it before you kill your RAM"
http://www.xtremesystems.org/forums/...4&page=1&pp=25
Ive got one SP stick that is flaking out at HIGH Vdimm. It will run ok for a while then ....beeeeep. If I unplug it, sit it on the desk for a day, its back to working fine. No issues @ 3.2Vdimm but 3.4-3.5Vdimm triggers this effect.
If you didnt know it, you might think the rams were DEAD.... so those with rams that up and failed...like failed to even post/boot... you may have similar situtaion occuring and the RAM isnt actually "bad".
My hypothesis on this is that it may not be the UTT chips responsible at all, it may actualy be the SPD rom chip that at high Vdimm flakes out and latches up.
Im going to use this stick to experiment.... Ill put a diode or two into the power supply to the SPD rom (the little 8pin gizmo on the end of the stick) and see if lower voltage to the SPD rom while high voltage on the UTT's works fine with no malfunction.
Its not a matter of high errors reported, when this "latchup" occurs the system wont even post/boot with the stick installed. This screams SPD rom malfunction, and the bios flips out since it cant get a response from the DIMM telling it WHAT kind/size/etc ram is installed.....since the bios needs this info to setup the memory controller registers. And we never get to boot if the rom malfunctions.
its the only logical conclusion I can ferret out so far.
It's killing me cause this stick is one AWESOME stick that does 250Mhz in dual channel on NForce2 8,2,2,2,2.0 1T and I want it in action.
This would be one of those things where Twinmos might cheap-out while other makers using the same UTT chips dont, and so we see "failures" fom Twinmos that arent really failures.
Though I have actually had BAD sticks, with 1-2-3 fixed errors in test #1 or #2 or #8 regardless of voltage/settings/etc. These are actual defective sticks.
Thoughts? Anyone with "BAD" "DEAD" sticks please try them out again in LOWER voltage sytem/settings to see if they come alive again.
Twinmos SP bh-5 seems to clock better with cas 1.5
managed to benc 01se @ 272-1.5-2-2-5 @ 3.7v , 2x512
Just received my two 512 SP sticks from Newegg....
AA4T 44D :toast: :banana: :woot:
i had a an 80mm fan directly on them, and a 90 on the mosfet. They were cool to the touch, i had 3.5V running through them. I'm not sure what was killing them. it might be this board for all i know.Quote:
Originally Posted by Toro 45
Okay, maybe some of you seasoned overclockers can give me an answer to these questions:
I have 2 sticks SP 512 AA4T 44D
Shuttle AK37GT motherboard. NO PCI lock, so I'm stick at 166fsb. Voltage is 2.6
In the bios I set 2-2-2-5, but in cpu-z it shows 2-2-2-6. is this a motherboard issue?
I set cas to 1.5 and I got BEEEEP BEEEEEP black screen, had to clear cmos, etc...
So, unable to hit 1.5-2-2-5, did I get crappy ram or is 2-2-2-6 the best I can expect at 166fsb?
I am going to upgrade to A64 system in the near future with a nice, overclock-friendly motherboard...I would like to get 1.5cas but worried that I won't since I can't @ 166fsb.
Thanks in advance for any help or advice
You need to raise your Vmem ;) Also, as I remember that MB did have dividers for the FSB/PCI... use 5/1 divider for above 166Mhz.Quote:
Originally Posted by Hemingway
One other note: On an A64 system the memory controller is part of the CPU so CAS1.5 is more likely... and on a good OC board much higher Vmem is available than on that Shuttle.
Peace :toast: