TCCD 2-2-2-5 @ 255mhz @ 3.3v

Wow.. this is such an interesting thread.

read through most of it :)

Well.. the reason i was reading, because I got soem Geil 3200 Ultra-X to complement my 240FSB Max Stable (3.2e @ 3.84Ghz WaterCooled)

With my first inital overclocking I was only able to get around 220 2-2-2-5 w/ 2.9v (tested only few min on memtest #5).

PCSTAT showed over 250Mhz @ 2-2-2-5.

So what type of OC should I be expecting from my Geil Ultra-X?
How high should I push my vDimm?
How are you guys checking for Stability? MemTest or SuperPi?

SAD

EDIT:

I plan to do more OC'ing of the mem this weekend, and wanted some pointers on these TCCD chips, esp Geil's UltraX.

Well... do a volt mod to your board... or if you have a booster.... pump up the voltage to see if Chris's Geil chips weren't handpicked... or.... to see if all Geil / Mushkin TCCDs can accomodate high voltages (past 3.0v) for some intense 5-2-2-2 action.

drunken master read yourself. maybe mushkin has modified the chips? thats what crucial said about the chips they use on their ballistix line.

really it is possible that mushkin sent had picked sticks to colin, but even so they wouldnt be able to hit speeds like that with those timmings without something else majorly different.

also from what ive read in the specs of the tccd(the ones i posted) there is no mention of internal voltage regulators.

EDIT: no flaming! pm sent
-saaya

Quote:

---

Originally Posted by drunkenmaster

oen other thing, similar voltages he threw into those dimms will get you same timings on bh-5, wasn't tccd's only advantage that at the lower timings and up at 2.5,3,3,7/8 it could do 300Mhz and surpass performance at lower timings lower fsb?
Either way, i've got some pqi on order, hopefully coming tomorrow but probo not being sent out till monday/ Mushkin as for last 18 months is very expensive for same thing, corsairs just corsair not even a option. OCZ's outta stock everywhere and not much else is available in the uk. :(

PS, theres a lot of stuff thats not very accurate in your reviews, the latency vs fsb thing is very much a grey area, it used to be 20-30Mhz more vs lower latencys, now its 40-50Mhz, sometimes more and not as much higher latencys and quite a few benchmarks/apps show better performance at higher latencies and much higher fsb's.

Also i very much disagree with the much heard theory that timings matter less on any specific platform. Its just wrong, lower timings or higher fsb's are both very much a performance issue on all platforms. running cas 2.5,3,3,7 at same speeds will cause a decent performance loss on any platform, i can't understand why the performance loss would matter less on a P4 than a ath 64, that makes zero sense.

IF you would argue anything its that the p4 is already a high latency memory subsystem due to talking through a chipset that every extra latency penalty makes every other subsystem wait even longer which is even worse than on a system thats much lower latency in the first place.

EDIT:- ps if ocz or gskill have said they have limiter in another thead, with details please link, not got time to read every thread.

---

the quad pumped bus is very unefficient in certain situations, this means theres a lot of bandwidth wasted in some situations, so the p4 gains more from a higher bandwidth than the a64 (in theory, to a certain point).
this would explain why the p4 would benefit more from a high badnwidth even if its with bad timings, than the a64.

but thats just the theory, no idear how its in the real world :D

Quote:

---

Originally Posted by ryanpgroovy

The voltage limit on TCCD would typicaly be set inside the chip .... unless there was a recent die revision that changed that ... those type of voltage limiting are used to limit the power consumption on the chip.

Not to be confused with OCZ EVP , which is a way to allow higher voltages to run more safely and is implemented on module .

---

so is there something on the ocz sticks that limits the vdimm? so 3.2v actually result in 3.1v to the ic's? yes or no? :P

Quote:

---

Originally Posted by SuAsDu

Wow.. this is such an interesting thread.

read through most of it :)

Well.. the reason i was reading, because I got soem Geil 3200 Ultra-X to complement my 240FSB Max Stable (3.2e @ 3.84Ghz WaterCooled)

With my first inital overclocking I was only able to get around 220 2-2-2-5 w/ 2.9v (tested only few min on memtest #5).

PCSTAT showed over 250Mhz @ 2-2-2-5.

So what type of OC should I be expecting from my Geil Ultra-X?
How high should I push my vDimm?
How are you guys checking for Stability? MemTest or SuperPi?

SAD

EDIT:

I plan to do more OC'ing of the mem this weekend, and wanted some pointers on these TCCD chips, esp Geil's UltraX.

---

cool! you could help us all to solve this riddle by overvolting your memory and seeing how high it goes! :D

im sure 3.3v vdiimm is 100% safe if you run it for just some hours and have a fan on the memory to cool it :)

Quote:

---

Originally Posted by WeStSiDePLaYa

drunken master read yourself. maybe mushkin has modified the chips? thats what crucial said about the chips they use on their ballistix line.

really it is possible that mushkin sent had picked sticks to colin, but even so they wouldnt be able to hit speeds like that with those timmings without something else majorly different.

also from what ive read in the specs of the tccd(the ones i posted) there is no mention of internal voltage regulators.

EDIT: no flaming! pm sent
-saaya

---

there is no way that they modified the ic's! no way!
maybe they integrated something in the package of the ic's, but thats the most they could do, and i really doubt this, it would cost them a LOT of money and work, and afaik all ic's are sold already in their ceramic packages!

the only way you can tweak memory chips is by manipulating the package or pcb.

i dont know anything about ic designs, but im sure its possible and easy to integrate a resistance into the ic's wich protects them from too high vdd, and from what ryan said those chips have such an protection inside the ic! the onyl way to overcome it would be to increase vdimm on the pcb, or manipulate the vss to be below 0V so that 3.3v effectively becomes 3.4v+.
dont know if this is possible and doable from a price point of view and how well the memory would like this.

so once and for all, do ocz sticks have something on their pcb that reduces and limits vdimm?

if not, then why do the mushkin chips oc higher? or maybe drunkenmaster is right and its a vanilla hand picked chip?


westsideplaya, why do you keep defending mushkin all the time? :stick:

i would have to say ocz defeintly have a regulator due to the fact you ocz guys seem to be avoiding the question. ;)

yeah im not saying the ic modding is true its just what a crucial rep said.

but like i said i found no mention of a internal voltage regulator. look thru the spec sheets more but i havent found any mention. and they said the can handle 3.6vdimm according to spec sheets.

still colin you may have tested an found the new bh-5! :banana:

Quote:

---

Originally Posted by WeStSiDePLaYa

i would have to say ocz defeintly have a regulator due to the fact you ocz guys seem to be avoiding the question. ;)

yeah im not saying the ic modding is true its just what a crucial rep said.

but like i said i found no mention of a internal voltage regulator. look thru the spec sheets more but i havent found any mention. and they said the can handle 3.6vdimm according to spec sheets.

still colin you may have tested an found the new bh-5! :banana:

---

come on ryan tony or steve :D speak up!


crucial=micron so that should clear up what they meant when they said "we" manipulated the ic ...

and i doubt you will find any infos about the ic internals and any voltage buffer or overvolt protection they might have built in as its probably secret.

Ok.. well.. Overvolting won't be possible for me. I lack the skills to do a vDimm volt mod, and I do not have the OCZ DDR Booster.

I'm running an IC7Max3. My vDimm max is 3.2, but from what I heard, setting vDimm to anythign over 3.0 is not stable, because of the limit of the 3.3v+ rail. Is this correct?

I do have an OCZ Powerstream, would it be possible to crank up the 3.3 rails a bit to help stability on the vdimm?

how high should I be expecting to hit at 2-2-2?

I'll report my findings this weekend.

SAD

Quote:

---

Originally Posted by SuAsDu

Ok.. well.. Overvolting won't be possible for me. I lack the skills to do a vDimm volt mod, and I do not have the OCZ DDR Booster.

I'm running an IC7Max3. My vDimm max is 3.2, but from what I heard, setting vDimm to anythign over 3.0 is not stable, because of the limit of the 3.3v+ rail. Is this correct?

I do have an OCZ Powerstream, would it be possible to crank up the 3.3 rails a bit to help stability on the vdimm?

how high should I be expecting to hit at 2-2-2?

I'll report my findings this weekend.

SAD

---

thats exactly what we are trying to find out with this ram. but yeah crank all your rails on that psu and test at max vdimm in bios and see how far you get with 2-2-2

suasdu your our saviour! :D

yes, increasing the3.3v rail to 3.4 or 3.5v will stabilize and increase your vdimm slightly and let you run 3.2v+, whatever your board maxes out at :)

i would recommend you to put heatsinks on all hot chips on your motherboard if you are planning to run high vdimm and higher rails though. for some tests its fine, but if you plan to run higher rails and high vdimm 24/7 you should cool all the parts :)

pleae post the results asap!

if you can get to 240mhz+ with 2-2-2 timings then its proof enough that its possible to run 2-2-2 at high speeds with high vdimm on tccd memory :)

Wow.. talk about a little pressure now :P

Ok.. i'm going to crank up the 3.3v+ rails to 3.4-3.5, and set vDimm to 3.2 in BIOS.

So how should I test for stabilty? I mean is considered stable here memtest86, superPi, or DualPrime? I know that different forums and group judge stability differnetly.

All I am trying to achieve is the best possible 2-2-2 @ 1:1 timing to complement my 245 ~ 250 OC on the CPU.

Hopefully I will have some results tonight. Dang.. this is going to cut into my pumpkin carving time.

SAD

Quote:

---

Originally Posted by SuAsDu

Wow.. talk about a little pressure now :P

Ok.. i'm going to crank up the 3.3v+ rails to 3.4-3.5, and set vDimm to 3.2 in BIOS.

So how should I test for stabilty? I mean is considered stable here memtest86, superPi, or DualPrime? I know that different forums and group judge stability differnetly.

All I am trying to achieve is the best possible 2-2-2 @ 1:1 timing to complement my 245 ~ 250 OC on the CPU.

Hopefully I will have some results tonight. Dang.. this is going to cut into my pumpkin carving time.

SAD

---

use memtest86 for mem. prime 95 torture test is nice too.

PS, just PMed Bigtoe from OCZ.

Great thread. I asked the Gskill guy what happens above 2.9v and he said he starts getting reboots etc. maybe extra voltage on the tccd means only a bit higher 2-2-2 but no higher 2.5-3-3-7 which really wouldn't help us much.

Collin was running a 120mm fan on the Mushkin,maybe that's what hellped them.

My EL3200 rev2 wont do above 220 with 2-2-2-5 @3.v,I haven't tried higher voltage; dont want to fry my warranty.

Quote:

---

Originally Posted by Nanobot

Collin was running a 120mm fan on the Mushkin,maybe that's what hellped them.

---

Obviously it helped them, but most users that overvolt ram 3.2 or higher tend to put a fan over their dimms. This is expected.

I'd love to see a pic of that Mushkin review stick with the heatspreaders off. :)

Maybe there was a die change :censored:

Quote:

---

Originally Posted by trans am

Obviously it helped them, but most users that overvolt ram 3.2 or higher tend to put a fan over their dimms. This is expected.

---

I aware of that,but the 50+ cfm that a 120mm fan gives must really help

Quote:

---

Originally Posted by SteveOCZ

Maybe there was a die change :censored:

---

Ie :OCZ Voltmaster ? ;)

Quote:

---

Originally Posted by Nanobot

I aware of that,but the 50+ cfm that a 120mm fan gives must really help

---

Honestly, how much could a fan help? I mean yes it cools the ram down, but 2-2-2-5 at 255mhz just because of a stupid fan? Cmon man!

**update *** I just checked and newegg shipped it. Should be here monday since it was shipped from the warehouse in NJ. It's ON!

Quote:

---

Originally Posted by Nanobot

Ie :OCZ Voltmaster ? ;)

---

Nope....you'll never get me to tell..... :p:

Quote:

---

Originally Posted by Nanobot

I aware of that,but the 50+ cfm that a 120mm fan gives must really help

---

lol, i wish memory would oc 10mhz higher per 10 more cfm of airflow you put over it :P

come on dude, thats :bs:

(direct email reply from Mushkin) note this was sent before Colin's response that confirmed it was TCCD)

Hello Alvin,

It is indeed TCCD. It is stated in the review. " PCstats removed the heatspreader on the Mushkin PC3200 LII V2 for a closer look at the DRAM. Below, we found that the PC3200 LII V2 is using Samsung K4H560838F-TCCD DRAM." When using the kind of setup discussed in the forums we cannot make a blanket statement which forces everyone to acheive the same results. If after you receive your modules, you find you are having difficulties obtaining advertised results then please contact us and we'll be happy to walk through it with you. Thanks so much.

Ian


------------------- Original Message -------------------
Sent: 10/28/2004
Subject: BLK lev 2 ver2 TCCD or not? PC stats review

I just ordered this new memory from newegg.com and
haven't recieved it yet.
http://www.newegg.com/app/viewproduc...=System+Memory


After reading this review, It raised a lot of
questions. Maybe you could carify some things.
First, if this indeed is TCCD, then how come it
crapped out at only 2.5-3-3 at only 232mhz on 2.8v?
if it was tccd wouldn't it at least go to 250-260mhz
at those timings? We are discussing this at
xtremesystems.org. Maybe you would like to stop by.
http://www.xtremesystems.org/forums/...7&page=1&pp=25

here's the review I'm reffering to:
http://www.pcstats.com/articleview.c...id=1668&page=1

warm regards,
alvin hayek (trans am)

We are not avoiding the question at all...i have run 233+ 3.5V with our 3200Plat rev2 as Has Eric at the lab...we are looking further into whether all TCCD does 250+ or just the cherry ones ;)

As soon as I know more i will let you all know.

Quote:

---

Originally Posted by bigtoe

We are not avoiding the question at all...i have run 233+ 3.5V with our 3200Plat rev2 as Has Eric at the lab...we are looking further into whether all TCCD does 250+ or just the cherry ones ;)

---

Well you better get your ass back to work because if all Mushkin Blk lev2 ver. 2 works like this, you guys are screwed!

Just kidding Tony. :)

Does anyone have a high res pic of a chip on PCB from some of this mushkin vs. some GSKILL?

Quote:

---

Originally Posted by trans am

Well you better get your ass back to work because if all Mushkin Blk lev2 ver. 2 works like this, you guys are screwed!

---

OH....screwed...how so? We have ram that does 260+ 2-2-2 now.

Quote:

---

Originally Posted by bigtoe

OH....screwed...how so? We have ram that does 260+ 2-2-2 now.

---

Figures, just my luck. How soon will it be out? any more details? And that previous remark wasn't serious.:)

Quote:

---

Originally Posted by STEvil

Does anyone have a high res pic of a chip on PCB from some of this mushkin vs. some GSKILL?

---

I'm sure Colin has one for sure. i'll have mine on monday.

This Thread is so Whacked ..


Number one .. no one avoided any questions ,

Number two .. Mushkin are solid , nice guys and deserve better that the assumption they have done something wrong...

Number Three... OCZ is never screwed , Our engineers are the best :cool:

Finally, The answer to all your questions regarding this oddity are now answered

http://xtremesystems.org/forums/show...d=1#post579587


:banana:

Quote:

---

Originally Posted by ryanpgroovy

This Thread is so Whacked ..


Number one .. no one avoided any questions ,

Number two .. Mushkin are solid , nice guys and deserve better that the assumption they have done something wrong...

Number Three... OCZ is never screwed , Our engineers are the best :cool:

Finally, The answer to all your questions regarding this oddity are now answered

http://xtremesystems.org/forums/show...d=1#post579587


:banana:

---

I knew that remark was going to get me in trouble lol. Thanks for clearing this up.

ok, dont need high res pics then :D

2-2-2 @ 255MHz

vs.

2.5-3-3 @ 300MHz

Why doesn't anyone compare ?

This low-latency is nice and everything, but if people are breaking 8K in Sandra with only having to put 2.8v on TCCD ... who cares ?

Enlighten me :)

You guys can trust Colin he is a good reviewer and a stand up guy as well :) (Colin you can pay me the money later ;))

Steve

Hey there guys. :)

Please forgive the late reply, was exceptionally busy at work today. Anyway here are the pictures as promised...

http://members.lycos.co.uk/csmo/PCSt.../Mushkin_1.jpg

http://members.lycos.co.uk/csmo/PCSt.../Mushkin_2.jpg

http://members.lycos.co.uk/csmo/PCSt...shkin/DRAM.jpg

Going to run high latency tests later, right now I'm a bit busy with dinner...

If you'd liek a scan of something specific, just post here (although I might miss it, not on purpose though) or PM me.

Anyway if I do not respond to questions please understand that I have a few things to do this wknd and will try my best to address any issues.

Hehe, Steve I can't pay if you keep increasing your... fees. ;)

yup, so colin was right all along. he never steered me wrong before, always straight cut and truthfull.

Quote:

---

Originally Posted by WeStSiDePLaYa

yup, so colin was right all along. he never steered me wrong before, always straight cut and truthfull.

---

He's a Canadian! Canadians are honest imo. :)

Quote:

---

Originally Posted by trans am

He's a Canadian! Canadians are honest imo. :)

---

thats because us canadians are nice people. :D

we are just people like everyone else.

Well anyway thanks to the guys who supported me throughout the thing, I'm glad Ryan at OCZ addressed the issue (for those who missed it).

Haha I actually thought I went crazy there for a few minutes! ;)

Quote:

---

Originally Posted by LilGator

2-2-2 @ 255MHz

vs.

2.5-3-3 @ 300MHz

Why doesn't anyone compare ?

---

OCZ 3700EB----3-3-2-7-2-2-2-4-11-13=49 mem. clocks

GEIL Ultra(old)--2-2-2-5-2-2-1-3-10-12=41 mem. clocks

My GEIL PC3500 ran 2-2-2-5 at 217mhz and the OCZ matched it around 250 mhz in PiFast testing

250(fsb)/49 (clocks) = 5.10204 mhz/clock vs 217/49 =4.428 =SLOW
217(fsb)/41 (clocks) = 5.29268 mhz/clock

Sandra was 7000 (ocz) vs 6600 (geil) which equalized the advantage the lower timings had.

How fast your computer is IMO is how much info/bandwidth you can push and also how quick the system can push that info/bandwidth.

Main reason for keeping my OCZ was due to it being able to handle 260+ mhz and I figured once I get a Mobo that can handle 3v+ the OCZ should rocket up to 290 or more and keep decent timings.
-------------------------------------------------------------------------
Example 1----2.5-3-3-6-2-2-2-4-11-13=48.5 mem. clocks

Example 2-------2-2-2-5-2-2-1-3-10-12=41 mem. clocks

300/48.5 =6.1855 vs 255/41=6.219 thus very equal


I know some changes like RAS to CAS effects the results more just giving a Generic example. The experts can shed more light on this

Guys latency vs high fsb. I have seen 2.5-3-3-10 at 300 and 285 3-2-2-0 all 1T...which do you think is faster?

Buffered sandra only tells you the fsb is high and not much else, unbuffered is better but im now leaning to everest and memtest86 as a guide.

If you get high write speeds on everest your going to see faster/higher benches (3d pi etc)...I have hit 3K writes with DDR2 at 733 ish 4-2-2-12 and 2800 with DDR1 2-2-2-5 on the max3...both were fast on PI and real good with 3dmark.

We should start an Everest data base so we can all compare, we are all used to sandra but for me it doesn't tell the whole story.

Quote:

---

Originally Posted by bigtoe

Guys latency vs high fsb. I have seen 2.5-3-3-10 at 300 and 285 3-2-2-0 all 1T...which do you think is faster?

Buffered sandra only tells you the fsb is high and not much else, unbuffered is better but im now leaning to everest and memtest86 as a guide.

If you get high write speeds on everest your going to see faster/higher benches (3d pi etc)...I have hit 3K writes with DDR2 at 733 ish 4-2-2-12 and 2800 with DDR1 2-2-2-5 on the max3...both were fast on PI and real good with 3dmark.

We should start an Everest data base so we can all compare, we are all used to sandra but for me it doesn't tell the whole story.

---

Isnt that waht the unbuffered test is for ? :D

Quote:

---

Originally Posted by WeStSiDePLaYa

yup, so colin was right all along. he never steered me wrong before, always straight cut and truthfull.

---

who said he was untruthfull?

sorry for soundsing a bit harsh in an earlier post colin :D :toast:

but tbh it looked exactly like it was bh5/bh6, with the bad 2.5-3-3 results and the high vdimm 2-2-2 results :D

cool, tccd is getting better and better with 2-2-2 and high volts... thats great! :D

yes it is to a point.

Simply put someone with both types of mem available need to just post some 265-280Mhz cas 2,2,2,5 (or nf3 whatever works best for you) vs max some of the best tccd can do 300Mhz or whatever, post everest, sisoft buff/unbuff, 3dmark, superpi, pifast and we will all pretty much get a rough idea. Sisoft on its own is no good, its a decent bench sometimes people knock it, seems to be the bench of choice every 3-4 months then everyone bashes it for a bit as useless. Buff tends to mean smeg all, unbuff gives you a better idea of what will happen with other benchmarks, but we'll see.

Heck, if my tccd(coming monday/tuesday) does ok even comparing say 250Mhz bh-5 to 270-275 tccd should give a rough clue as to whats up.

just another statement of fact, this is all high end stupid volts available situations. I think 90% of us without doubt could agree for your average user looking for best clock friendly ram on non modded boards tccd is the mem to recommend. it may or may not be the official best mem available, but for your average user at low volts it rocks.

Quote:

---

Originally Posted by drunkenmaster

just another statement of fact, this is all high end stupid volts available situations. I think 90% of us without doubt could agree for your average user looking for best clock friendly ram on non modded boards tccd is the mem to recommend. it may or may not be the official best mem available, but for your average user at low volts it rocks.

---

uhhhh..... if your looking for what is good for the average user your in the wrong forum. ;)

and 3.4vdimm isnt crazy volts at all. lots of people would and do use that much.

Quote:

---

Originally Posted by bigtoe

Guys latency vs high fsb. I have seen 2.5-3-3-10 at 300 and 285 3-2-2-0 all 1T...which do you think is faster?

Buffered sandra only tells you the fsb is high and not much else, unbuffered is better but im now leaning to everest and memtest86 as a guide.

If you get high write speeds on everest your going to see faster/higher benches (3d pi etc)...I have hit 3K writes with DDR2 at 733 ish 4-2-2-12 and 2800 with DDR1 2-2-2-5 on the max3...both were fast on PI and real good with 3dmark.

We should start an Everest data base so we can all compare, we are all used to sandra but for me it doesn't tell the whole story.

---

sounds like a good idea, an everest database :D

Quote:

---

Originally Posted by µnrealneo²

285 is a tad faster.

275 x-2-2-x is matched by 300 x-3-3-x approximately in 3D Mark 01. (OnePageBook's test)

---

275 2-2-2 is matched by 300 2.5-3-3 ;)

275 3-2-2 is a slightly different story, just like 3-3-3...

and i dont know if opb used the a64 tweaker...

Quote:

---

Originally Posted by saaya

275 2-2-2 is matched by 300 2.5-3-3

---

So why are we getting hoping that TCCD can do 2-2-2-5 @ 255MHz if you get lucky ... when it can already smash that at 2.3-3-3- @ 280-300MHz ?

:confused:

Quote:

---

Originally Posted by LilGator

So why are we getting hoping that TCCD can do 2-2-2-5 @ 255MHz if you get lucky ... when it can already smash that at 2.3-3-3- @ 280-300MHz ?

:confused:

---

Its just so nice to have ram that is so flexible. You could run it at 2-2-2-5 at 255mhz in an axp system or you could take that very same ram and through it into an intel setup doing 2.5-3-3-7 at 300mhz!

Its flexible,,,,,and that's the name of the game.

Quote:

---

Originally Posted by LilGator

So why are we getting hoping that TCCD can do 2-2-2-5 @ 255MHz if you get lucky ... when it can already smash that at 2.3-3-3- @ 280-300MHz ?

:confused:

---

Not all the people have and amd system. (the open multiplier)
Average overclocks of a p4 3.0gh for example is just around 240fsb
Here I would benefit from the tight timings more the high fsb.

colin you got these nice 255 2-2-2-5 3.3v on INTEL

what about DFI NFII ULTRA INFINITY ? any idea...i have 3.3vdimm tho ...

http://www.overclockercafe.com/Revie...kin_PC3200_V2/

also 0431 TCCD batch

http://www.systemcooling.com/mushkin_v2_pc3200-01.html

same

seems batch 0431 does better then old 0416 ? :confused:

For the record....

There are NO voltage regulators in DDR chips or on sticks of DDR memory.

Not on/in CH-5 and not on/in TCCD. Whoever/whereever that stuff comes from is mis-informed, period.

Looks like the latest revisions of the TCCD chips are showing the expected signs of manufacturing improvements that always follows the excellent ongoing work of high quality manufacturers, ever improving thier yeilds and efficiencies. Samsung has always been such a company.

The chips....just like BH-5....have an absolute maximum rating of 3.6V. It would seem these newer chips remain stable with higher operating voltage than older TCCD's and we get to enjoy the speed benefits of running the chips at higher voltages. Older versions probably got toohot at some point inside at higher voltages and instability prevented reliable operation, the newer chips have a higher STABLE maximum performance point.

I cant read the labels well enough in the pictures to see the date-code, but there is most likely a date, after which, you get the higher max performance available.

I for one would luv to see the best someone can get out of 4x 512M of these newer versions in a 939 board ...aka 2 GIGs... as this is my goal for a 939 future, 2 GIGs. 250Mhz x11 with 4x 512M TCCD @tight on a 3200+ 90um 939chip would be really cool.

To repeat.... there are NO VOLTAGE REGULATORS in DDR chips. None, period.

I dont give a :banana::banana::banana::banana: who told you, or what engineer buddy said what, etc... they are ALL full of it. The chips are made by the BAZILLIONS in Samsung's factory and shipped out to everyone. There are NO custom anythings being made, and I really hate to burst your bubbles but Mushkin/OCZ/etc/etc amount to SQUAT for volume as customers of Samsung. They ship MILLIONS of these chips per month and the "specialty" ram suppliers (the folks like Mushkin that solder them onto PCboards) dont make a DENT in Samsungs production.

Anyone of YOU can go into the DDR stick making business at the drop of a hat, all it takes is a few $K to get started. There are a ton of Taiwanese contractors that make generic "reference" DDR pcboards, there are zillions of contract manufacturing houses that will take your components (chips, resistors, caps, pcboard, spd-rom) and solder them together for you. You then test and ship and support.

It is NOT rocket science. Nobody is building in special super-duper circuits for mad-speed boosts, etc. What the Mushkins/OCZ's/ etc do is put in the extra effort to SUPPORT us OCr's and our efforts, and they do a great job, and we all thank them very much. Its not a big margin business and they take alot of crap in the process, so thanks guys.

The "EDGE" you can get is in quality components....resister packs and caps with good tolerence's meeting the DDR spec correctly, good quality pcboard materials, and like in the case of the "brainpower" boards a little better layout of components to increase high FSB performance. Thats about it. Hell, we dont even see any of the OEMs bothering to employ "better" serious cooling to differentiate themselves from competitors. As has been discussed in this thread, a heatspredder with more BULK using perhaps copper and mated to the DDR chips with a really GOOD heat transfer medium would be a major point of differentiation....assuming everyone uses the same TCCD chips.

Just keep in mind how many units of DDR all the specialty makers ship in a month VERSUS one months production at...DELL? HP? etc. It pales in comparison.

we were mostly talking about pcb voltage limiters. not inside the actual IC.

my TCCD does 250fsb 7-3-3-2.5 @ 2.9vdimm on an LGA system

Quote:

---

Originally Posted by SuAsDu

I'm running an IC7Max3. My vDimm max is 3.2, but from what I heard, setting vDimm to anythign over 3.0 is not stable, because of the limit of the 3.3v+ rail. Is this correct?

---

The main problem with the Max3 is the lack of vtt tracking over 2.8vdimm

New chips... guys look at the "Rev. F" marking on the chips...

Checkout the different revisions here:
http://store.yahoo.com/digi4me/gskddrexseeo.html

Hey wait... look at the PCB on the Mushkin LVLII V2s Colin has. They're the BrainPower PCBs. That's new...

Quote:

---

Originally Posted by Colin

Hey there guys. :)

Please forgive the late reply, was exceptionally busy at work today. Anyway here are the pictures as promised...

http://members.lycos.co.uk/csmo/PCSt.../Mushkin_1.jpg

http://members.lycos.co.uk/csmo/PCSt.../Mushkin_2.jpg

http://members.lycos.co.uk/csmo/PCSt...shkin/DRAM.jpg

Going to run high latency tests later, right now I'm a bit busy with dinner...

If you'd liek a scan of something specific, just post here (although I might miss it, not on purpose though) or PM me.

Anyway if I do not respond to questions please understand that I have a few things to do this wknd and will try my best to address any issues.

Hehe, Steve I can't pay if you keep increasing your... fees. ;)

---

seems new PCB yep

hmm i heard that intel 865/875 doesnt really like TCCD. is this true?

nope, with the right bios, it works great :
http://users.pandora.be/gamer/Gskill/32.JPG

what timeings is that with and what about Abit?

2.5-3-3-5.
don't know about Abit, but I think they are aware of the problem.

OK, guys, the ram came just now. I popped the spreader off on 1 of the sticks carefully and it looks like I have the same thing that Colin posted. take a look. I'm going to take out my BH5 and try these suckers in and see if I get the same results. Wish me luck. :)

transam you ordered from newegg 291$ not from mushkin directly 255-264$ ?

can you see which PCB they're using now ?

how does it do?

/me steals MUSHKIN TCCD from transam :D

You rock, trans ;) Way to void that warranty for the cause! The Kingston looks like it's using the new ICs too, but it's definitely not using the BP PCB. The Mushkin on the other hand... well let us know how it does on standard board volts (2.9v)

Quote:

---

Originally Posted by WiCKeD

You rock, trans ;) Way to void that warranty for the cause! The Kingston looks like it's using the new ICs too, but it's definitely not using the BP PCB. The Mushkin on the other hand... well let us know how it does on standard board volts (2.9v)

---

Actually, the 3rd screenie I posted is after I removed the heatspreader. Notice I did it "carefully". lol

The ram looks exactly how it did before I removed the heatspreader. I paid $291 from newegg. After I ordered, I noticed you could buy it direct from Mushkin for less, but it wasn't much less. I think it was only $10 bucks. I did this just in case the ram turned out to be crap. Which it looks like it's starting to head in that direction. I tried it at 10*240 1:1 2-2-2-6 1T drive strength at 4. vdimm at 3.3v system posted fine started to load windows, then I get the dreaded file missing error black screen message. WTF! I kicked the volts down to 2.8v and tried running at 2.5-3-3-7 1T @ 282mhz. 282*9 1:1 dram drive strength was still at 4. posted and then rebooted before windows started to load. Then I tried a fail safe 240*10 2.5-3-3-7 1T 1:1 and booted again. Windows loads and then I get windows is going to shut down popup. I went back turned my drive strength to auto and ran memtest86 test 5 and got 0 errors. Maybe I was having problems running with the drive strength at 4, but I thought 4 was the best option for stability, at least this was the case with my BH5. I am going to read some more and try Oscar wu's tips for running TCCD at high speeds. I'll try to run it at Colin's setting again with the dram DRive strength on auto and see what happens. I'm in windows now at 240 2.5-3-3-7 1T, but this is nothing to get excited about. any other tips?

is the "brainpower" better than your standard PCB?

EDIT: woah, posted right before me, def check out oskar's thread - work from there

Quote:

---

Originally Posted by Playful_Buffalo

is the "brainpower" better than your standard PCB?

---

yes. it seems to oc about 10% better

275fsb @ 2.5-3-3-X 1T & 2.9v and I will be happy. :)

Quote:

---

Originally Posted by WiCKeD

275fsb @ 2.5-3-3-X 1T & 2.9v and I will be happy. :)

---

dudes, I got bad news. This ram plain sucks! I don't know if I blew it right from the get go with the volts or what, but I tried a tried and true setting for TCCD. 270*9 2.5-3-3-7 1T and 2.5-4-3-7 1T @2.8v and 2.9v and ran memtest86 test 5. I got over 100,000+ errors and it was still going but I decided to turn it off because I just couldn't look at it anymore. :( Either I got old TCCD and killed it by running too many volts, or I just got a bad set. I'm not even going to bother running single dimms, because I ordered a dual set and it should work no matter what. I am going to wait for OCZ to release their latest and greatest top secret memory (should be this week or early next week) and try to exchange this crap at newegg for the new ocz or something else. Your tips are welcome, but I doubt anything will work. PS, I tried to boot into windows again at 2.5-3-3-7 1T at 240mhz and it kept looping post. I running 240mhz at 3-3-3-7 1T as I'm typing this.

Quote:

---

Originally Posted by trans am

dudes, I got bad news. This ram plain sucks! I don't know if I blew it right from the get go with the volts or what, but I tried a tried and true setting for TCCD. 270*9 2.5-3-3-7 1T and 2.5-4-3-7 1T @2.8v and 2.9v and ran memtest86 test 5. I got over 100,000+ errors and it was still going but I decided to turn it off because I just couldn't look at it anymore. :( Either I got old TCCD and killed it by running too many volts, or I just got a bad set. I'm not even going to bother running single dimms, because I ordered a dual set and it should work no matter what. I am going to wait for OCZ to release their latest and greatest top secret memory (should be this week or early next week) and try to exchange this crap at newegg for the new ocz or something else. Your tips are welcome, but I doubt anything will work. PS, I tried to boot into windows again at 2.5-3-3-7 1T at 240mhz and it kept looping post. I running 240mhz at 3-3-3-7 1T as I'm typing this.

---

thats that ram out the window then ...

anyone else actually tried this ram yet ?

Quote:

---

Originally Posted by Gamer

nope, with the right bios, it works great :
http://users.pandora.be/gamer/Gskill/32.JPG

---

Here is my Super PI with FX55 overclocked to 2.8ghz , no fsb added just multi.

Quote:

---

Originally Posted by The Virus

Here is my Super PI with FX55 overclocked to 2.8ghz , no fsb added just multi.

---

And your point is? Dude, how is this in any way relevant to this thread? Can you please delete it?

Quote:

---

Originally Posted by trans am

I just love seeing any 2x512 go higher at 2-2-2 235mhz 1T than my BH5. I just hope it all works out. :)

---

Just reading through this thread, thought I might drop this in, been running 12 hours so far at 3.3v, wish me luck. :)

(2x512 OCZ BH-5)

Quote:

---

Originally Posted by st0nedpenguin

Just reading through this thread, thought I might drop this in, been running 12 hours so far at 3.3v, wish me luck. :)

(2x512 OCZ BH-5)

---

Ya, OK man try that at 2-2-2-6 1T and I'll be impressed.

Quote:

---

Originally Posted by st0nedpenguin

Just reading through this thread, thought I might drop this in, been running 12 hours so far at 3.3v, wish me luck. :)

(2x512 OCZ BH-5)

---

Why would you run BH-5 at 2-3-3 .... ?

Quote:

---

Originally Posted by trans am

Ya, OK man try that at 2-2-2-6 1T and I'll be impressed.

---

Pfft, 12 hours to go, then it's timings time.

I should just add, this RAM has never ran more than 200MHz stable in the year or so since I got it, in all honesty, I'm amazed it even posts...

Quote:

---

Originally Posted by st0nedpenguin

Pfft, 12 hours to go, then it's timings time.

---

12 hours to go? Dude it's stable!!! at this rate it's going to take you 1 year to get those at 2-2-2 1T and i'm guessing they crap out at 240mhz or so.

Quote:

---

Originally Posted by st0nedpenguin

I should just add, this RAM has never ran more than 200MHz stable in the year or so since I got it, in all honesty, I'm amazed it even posts...

---

What amazes me is that after 3k posts on this forum, you still double post.

I am really confuse,

You guy said Sumsung 431, K4H560838F - TCCD is a new revision chip, how can this be, because i have a pair of corsair twinx cmx512 3200c2 ver4.1 about a month ago and it have the same chips on these module.

Quote:

---

Originally Posted by trans am

And your point is? Dude, how is this in any way relevant to this thread? Can you please delete it?

---

The point is that stock FX55 gets 32sec when the P$ 4.2ghz with 280fsb gets the same. What? You don't think 4.2ghz vs 2.6ghz makes P4 look like $:banana::banana::banana::banana:?

Quote:

---

Originally Posted by The Virus

The point is that my stock FX55 gets 32sec when your P$ 4.2ghz with 280fsb gets the same. What? You don't think 4.2ghz vs 2.6ghz makes P4 look like $:banana::banana::banana::banana:?

---

First of all I have a 3200+ 62w mobile. So I don't get where your getting your info that says I have a P4. Read my sig, it's pretty self explanatory. And why would you post something in here that has no relation to Mushkin memory or new TCCD? Get a clue.

i cant get these mushkins to run for :banana: :banana: :banana: :banana: in my neo2 :( spd setting are about the only speed they will run at. everything either results in no post or random reb00ts.

i've read a couple of reviews on this ram and it seems 250+ is difficult to reach even with loose timings.

http://www.gruntville.com/reviews/me..._pc3200ddr222/

Quote:

---

Originally Posted by trans am

First of all I have a 3200+ 62w mobile. So I don't get where your getting your info that says I have a P4. Read my sig, it's pretty self explanatory. And why would you post something in here that has no relation to Mushkin memory or new TCCD? Get a clue.

---

It does actually, running 200fsb at 2-2-2-6 timings is way better then having ram running with crappy timings like 3-4-4 at 250-270fsb in real time performance.

Quote:

---

Originally Posted by esoteradactyl

i cant get these mushkins to run for :banana: :banana: :banana: :banana: in my neo2 :( spd setting are about the only speed they will run at. everything either results in no post or random reb00ts.

---

Is that BH5 your talking about??

Quote:

---

Originally Posted by The Virus

It does actually, running 200fsb at 2-2-2-6 timings is way better then having ram running with crappy timings like 3-4-4 at 250-270fsb in real time performance.

---

Yeah, duh. but I test new products often and I don't run 3-4-4 timings on anything. Yeah I changed my BH5 just now so I could play with this new ram, but you still seem to miss the point. Could you please go away. Or would you like me to get a mod to do it. :)

Quote:

---

Originally Posted by Wingz

i've read a couple of reviews on this ram and it seems 250+ is difficult to reach even with loose timings.

http://www.gruntville.com/reviews/me..._pc3200ddr222/

---

I just read the review. This looks identical to how it's performing on my system. It's RMA time!

Quote:

---

Originally Posted by trans am

Yeah, duh. but I test new products often and I don't run 3-4-4 timings on anything. Yeah I changed my BH5 just now so I could play with this new ram, but you still seem to miss the point. Could you please go away. Or would you like me to get a mod to do it. :)

---

Yeah, im glad you agree with me. Why did you get rid of bh5 ram? They come alive above 3.3v.

Quote:

---

Originally Posted by The Virus

Is that BH5 your talking about??

---

this is a TCCD thread isnt it ;)

Quote:

---

Originally Posted by The Virus

Yeah, im glad you agree with me. Why did you get rid of bh5 ram? They come alive above 3.3v.

---

OK virus if you're so smart. You run your ram at 2-2-2-5 as high as it will go. Someone with good tccd run it at 2.5-3-3-7 @280mhz+ and lets see who scores higher.

Now I know why your name is virus. Getting rid of you is like trying to get rid of HIV.

Quote:

---

Originally Posted by trans am

OK virus if you're so smart. You run your ram at 2-2-2-5 as high as it will go. Someone with good tccd run it at 2.5-3-3-7 @280mhz+ and lets see who scores higher.

Now I know why your name is virus. Getting rid of you is like trying to get rid of HIV.

---

Dont believe everything you read in review sites . Almost every new ram that is reviewed OC's like sick but when its out for retail it bearly does default settings.

Quote:

---

Now I know why your name is virus. Getting rid of you is like trying to get rid of HIV.

---

Oh no he didn't?!?! :eek: :lol: :ROTF: :rotf: :flame:

Quote:

---

Originally Posted by sparkie34

Oh no he didn't?!?! :eek: :lol: :ROTF: :rotf: :flame:

---

He maybe needs some of this.

http://prozac.com/index.jsp

:D

Quote:

---

Originally Posted by The Virus

He maybe needs some of this.

http://prozac.com/index.jsp

:D

---

(quotes from The Virus)
Your lucky man, I can't even do 217fsb on my FX55 and MSI NEO2, it boots but crashes when i try to run 2001mark. Maybe sata 3/4 are dont have pci lockage.

Kakaroto, what should the HT be set at for 12x217fsb?? This should give me exactly 2.6ghz which is stock speed for fx55. I got corsair pc3500c2 bh5 2x512 .

My experiance with FX55 and MSI neo2 is nothing but great performance since i got it 4 days ago. When comes to overclocking, I can't even run 210fsb with 2x512 pc3500c2 bh5 ram stable. Everytime i try to run 2001 3d mark it pauses then get a graphical glitch. I'm starting to assume this board doesnt even have its AGP locked. I know that its not my ram because i tested it with Memtest and passed even at 220fsb 2-2-2-6 timings. I'm extremelly dissapointed with this board. Everything stock runs Fine , try to oc even by 10fsb it freaks out, last time i crashed and rebooted thewindows reported an error coming from my 6800gt. AGP is probably running out of spec. Very dissapointed. Whats even more shocking to me is even at everything stock 13x200 , if i just raise my cpu voltage to say 1.55 or 1.6v and if i try to run 2001 mark it would freeze. Something is really wrong with this mobo when overclocked. Reading review sites i see they getting like 2.8-3.0ghz on Air and i can't even get 2.7ghz running. This plain sucks.


Quote:
Originally Posted by saaya
HEY! hahahah funny, i remember your first posts on xs, you said the a64 sucks in games compared to an athlonxp at the same clockspeed, linking to an hardocp article hehehe, do you remember?

hehe, im glad you found out the a64s are indeed very nice cpus

sweeeet rig you got there man! im glad you calmed down and stopped flaiming people and became a nice member of xs!

check out the neo2 tweaking guide in the amd section, theres a stick with many pages all about the neo2


(the virus)
I never said A64 sucks in games ,what i said was there was not that much of a difference in Doom3 maybe because its heavily video dependent. I always knew that A64 is king when comes to gaming. Plus , what happend in that thread really wasn't flamming, i was just stating my opinion.

Anyway, in bios i cant find the cpu/mem ratios like 1:1 , or am i that noobish that i cant find it.







I think your last question sums it up.

Quote:

---

Originally Posted by trans am

(quotes from The Virus)
Your lucky man, I can't even do 217fsb on my FX55 and MSI NEO2, it boots but crashes when i try to run 2001mark. Maybe sata 3/4 are dont have pci lockage.

Kakaroto, what should the HT be set at for 12x217fsb?? This should give me exactly 2.6ghz which is stock speed for fx55. I got corsair pc3500c2 bh5 2x512 .

My experiance with FX55 and MSI neo2 is nothing but great performance since i got it 4 days ago. When comes to overclocking, I can't even run 210fsb with 2x512 pc3500c2 bh5 ram stable. Everytime i try to run 2001 3d mark it pauses then get a graphical glitch. I'm starting to assume this board doesnt even have its AGP locked. I know that its not my ram because i tested it with Memtest and passed even at 220fsb 2-2-2-6 timings. I'm extremelly dissapointed with this board. Everything stock runs Fine , try to oc even by 10fsb it freaks out, last time i crashed and rebooted thewindows reported an error coming from my 6800gt. AGP is probably running out of spec. Very dissapointed. Whats even more shocking to me is even at everything stock 13x200 , if i just raise my cpu voltage to say 1.55 or 1.6v and if i try to run 2001 mark it would freeze. Something is really wrong with this mobo when overclocked. Reading review sites i see they getting like 2.8-3.0ghz on Air and i can't even get 2.7ghz running. This plain sucks.


Quote:
Originally Posted by saaya
HEY! hahahah funny, i remember your first posts on xs, you said the a64 sucks in games compared to an athlonxp at the same clockspeed, linking to an hardocp article hehehe, do you remember?

hehe, im glad you found out the a64s are indeed very nice cpus

sweeeet rig you got there man! im glad you calmed down and stopped flaiming people and became a nice member of xs!

check out the neo2 tweaking guide in the amd section, theres a stick with many pages all about the neo2


(the virus)
I never said A64 sucks in games ,what i said was there was not that much of a difference in Doom3 maybe because its heavily video dependent. I always knew that A64 is king when comes to gaming. Plus , what happend in that thread really wasn't flamming, i was just stating my opinion.

Anyway, in bios i cant find the cpu/mem ratios like 1:1 , or am i that noobish that i cant find it.







I think your last question sums it up.

---

And your point is?? how is this in any way relevant to this thread

Please keep this thread on track and about the main subject. If you guys want to flame each other do it in private. Consider this a warning.

Quote:

---

Originally Posted by Shroomalistic

Please keep this thread on track and about the main subject. If you guys want to flame each other do it in private. Consider this a warning.

---

No flaming here bud. Read all my posts, not a sign of flame, this guy is harrasing me and telling me that im HIV virus. Not very nice.

Quote:

---

Originally Posted by trans am

What amazes me is that after 3k posts on this forum, you still double post.

---

What amazes me, is that after not posting for a few months, I find jumped up little forum nazis like you verbally abusing people left right and centre, you ran out of riddalin or something, sheesh?

And Shroom, yeah this is a flame, but sorry man, this guy is a complete :banana::banana::banana::banana::banana:.

Quote:

---

Originally Posted by st0nedpenguin

What amazes me, is that after not posting for a few months, I find jumped up little forum nazis like you verbally abusing people left right and centre, you ran out of riddalin or something, sheesh?

And Shroom, yeah this is a flame, but sorry man, this guy is a complete :banana::banana::banana::banana::banana:.

---

Ok, I admit I wasn't very nice to you but the Virus was a different story altogether. Talking about some Pentium 4 and making absolutley no sense in regards to the thread. I have zero tolerance for that. I appologize for being rude to everyone and I'm sorry to shroomalistic. This is an amazing forum and It's a privilege to be a part of this outstanding community. I was really pissed off about that Mushkin not working as expected. After waiting almost a week to get it, and it's pretty much DOA. so I was a bit flustered. Anyone that has had new computer parts not work knows the feeling and I hope you understand. I took it out on you guys and I am ashamed of my actions. Thanks. Lets keep it under control. Could a mod delete all the garbage from here so we can start fresh?
or maybe just keep future posts in this one since it's about high voltage TCCD:
http://www.xtremesystems.org/forums/...d=1#post587199

Meh, not a problem mate, I know the pains of dud hardware, the only option is to suck it up, deal with it, and sacrifice another sheep. :D