TCCD and high voltage mystery explained

[ryanpgroovy]

Alright after a few in the lab we figured out the awe inspiring mystery of TCCDs that do well with higher voltages

It appears that samsung made some minor die revisions and that newer production chips do perform better with high voltages


I pulled about 20 current production OCZ-PC -3200 plat rev 2 and put the booster to them , from 2.8 to 3.1 volts showed no peformance improvement
from 3.1-3.6 we were able to run between 466 and 480 at 2-2-2 memtest stable and 510 and 550 2-3-3 memtest stable , with a fan running over the modules

These are current production parts and it seems that we have been shipping these chips to retailers for about 2 weeks

keep in mind we only saw a diference at 3.1 volts and above ,

I would guess most TCCD products produced in the past 3-4 weeks should respond the same .. but thats a guess based on initial testing


oh yeah , heres a quickie screenshot at 3.3 volts

:toast:

[eshbach]

sounds good,

do these sticks clock any higher at 2.5-3-3 with 3.4v? could we possibly see 320 1:1 2.5-3-3 with high voltage?

[ryanpgroovy]

I have not had time to play that in depth , but I would bet that with higher voltages they do better at all timings.

[lalPOOO]

You scared me there for a second, I thought that was 1mb super pi :p:

You guys at OCZ are too much, keep it up :D

[Nanobot]

Thank you for solving the TCCD mistery Ryan :idea: :toast:


Are the late production OCZ TCCD the ones with the new OCZ badge on the HS ?

The HS on mine are just the regular OCZ stamped units.

[ryanpgroovy]

Just trying to post up a quicky .... no time for real benchmarking in my busy day

[ryanpgroovy]

Its hard to tell when exactly samsung changed the die , I do know the chips were using to build now are the same batch we have been using for 3 weeks at least.

[Dissolved]

Anyways to get a 1gb set of this memory with the newer dies?

id gladly pump 3.4v~ to get 280(560ddr) cas 2.5 :)

[trans am]

Its hard to tell when exactly samsung changed the die , I do know the chips were using to build now are the same batch we have been using for 3 weeks at least.

What is the ram that Tony mentioned that does 2-2-2 at 260mhz? was that just a new tccd that got lucky? Or is there some serious :banana::banana::banana::banana::banana: brewing in the ocz LAb?

[saaya]

wow thx a lot for soending so much time on solcing this riddle! :toast:

id recommend you to find out what chips respond well to high vdimm and put a new sticker on that ram :D

[TEDY]

so better wait a week or two so all new ram is new samsung die :D

what ram mostly likely then ?

mushkin/ocz/gskill ?

i really want high voltage tccd cause i have 3.3vdimm available on my DFI NFII :toast:

[ryanpgroovy]

The betas tony have are not TCCD ....

hopefully people will buy lots of ddr boosters with all this new voltage lovin .....

[trans am]

[QUOTE=ryanpgroovy]The betas tony have are not TCCD ....

A new mystery ic that isn't TCCD that does 260mhz+ at 2-2-2. Everyone get your checkbooks ready!

[STEvil]

heh.. wonder if they have another "shift point" higher up? :D

[TEDY]

damn i smell a big xmas surprise from OCZ

DDR 600 2-2-2-5/11 3.3Vdimm TCCD OWNAGE :D

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
and wake up with PC2100 nanya ram LOL :D

[Dissolved]

The betas tony have are not TCCD ....

hopefully people will buy lots of ddr boosters with all this new voltage lovin .....


Have you tryed better cooling on these?

IMHO i would like to see some Heatsinks, water cooling possibly?

my EB burns me when i use 3.0v and testing when memtest last night the heater in the house came on cuz its cold in calif at night this week, and seconds after i started getting errors (it was on the 34th run of test5) so it was heat related im 99% sure.

i would like to pull the spreaders off and add some sinks or something else, but i wouldnt wanna kill the warranty if i havent allready by excideing the voltages..

[STEvil]

use a fan.

[iboomalot]

if I could get some 1:1 DDR 600 stable ram at 2-2-2-5-T1 I would spend 400.00 on a pair without blinking. I wonder what my OCZ 3700EB would be worth if this happens??

[trans am]

Have you tryed better cooling on these?

IMHO i would like to see some Heatsinks, water cooling possibly?

my EB burns me when i use 3.0v and testing when memtest last night the heater in the house came on cuz its cold in calif at night this week, and seconds after i started getting errors (it was on the 34th run of test5) so it was heat related im 99% sure.

i would like to pull the spreaders off and add some sinks or something else, but i wouldnt wanna kill the warranty if i havent allready by excideing the voltages..


Maybe OCZ should make a new heatspreader for this new High voltage ram. Maybe something with fins that dissipates heat better? or maybe a windtunnel design that surrounds the dimms with a fan at the top?

[WeStSiDePLaYa]

damn, so can we say yet this stuff is officially better that bh-5? i think so.

[trans am]

damn, so can we say yet this stuff is officially better that bh-5? i think so.

As soon as my mushkin comes, I'm putting the BH5 on Ebay before this info gets leaked onto every forum.

[Dissolved]

use a fan.


common sense yes, But even at stock speeds at 2.8v they get pretty warm in a cold room in the open air.

What im Saying is, From what ive seen, memory is getting more power hungry and Might benifit from better Cooling Support.

Problem is tho, memory Dimms are so close together heatsinks make this impossible.

Im on my DFI, i use dimms 1 & 3 so i have a 1/2" area between the memory dimms to allow me to do something. Fan is OK, but fans wont truly disspate all the heat from the memory..

[WeStSiDePLaYa]

well i know i have good airflow and i get a 80mm fan right over my dimms and my ram stays pretty chilly.

but wow, i really want to get some of this BH-5 killers in my rig.

[Ubermann]

I have been trying mine up to 3.8 Volt, after that computer crash.
From what i notice is that Cas2 will gain most of the boost from higher volt.
~3.5Volt seems to give most boost at Cas2.
Cas 2.5 runs best at 3.1volt.
Cas 3 dont like above over 3.0 volt.

I think it is this way, have only noticed after messing with a booster.

[Sauro]

What a wonderful time to be building a new comp :)

[drunkenmaster]

well have to say, though i think most of what corsair does is crap, months behind everyone else and generally at a premium dealing off their small 4-6 month period of putting out better stuff than others(hasn't been that way for a long time) their PRO range of mem with heatspreaders have the basic idea down, unless you're using a thermalright xp120(is the hyper x and asus ice cooler better than it due to copper ness?) then you ahve lots of space upwards from the mobo. You really can't fit wider spreaders with more fins, but extending the heatspreaders up will increase area for cooling a lot. If someone can actually sort out a decent tim to attach them then something that increases area would be most welcome, and will make fan cooling that much more affective.

PS< ryan, i take it this means the BS in other thread about OCZ and others admiting voltage limiting is shens?

PPS< while its nice it will do higher timings at 2,2,2,5, 3.4v or was it 3.5v in teh review to hit 257Mhz is similar to bh-5, it has more headroom above that to hit much higher fsb at lower timings, but we still don't have a bh-5 trade in type of mem here, IMHO. I'm getting some to play with cos 1gig seems to do very well, bh-5 x 1gb is difficult. but for benching, bh-5 + 3.6v is the win.

[Nanobot]

Problem is tho, memory Dimms are so close together heatsinks make this impossible.

Im on my DFI, i use dimms 1 & 3 so i have a 1/2" area between the memory dimms to allow me to do something. Fan is OK, but fans wont truly disspate all the heat from the memory..


I have found the same advantage running in 1&3 even tho Oskar recommends 1&2,for some reason 1&2 dont work out as well as 1&3 for stability in my system(heat ?).

[trans am]

I have found the same advantage running in 1&3 even tho Oskar recommends 1&2,for some reason 1&2 dont work out as well as 1&3 for stability in my system(heat ?).

Yeah I just figured 1 and 3 would be the best, simply because air can pass through dim 2. The only thing I can see that using 1 and 2 will do for you is generate a lot of heat. After reading that thread where oskar mentioned I decided to switch to 1and2. Just made everything hotter. and btw, dram drive strength 1 and 2 gives me no post. lol

[ryanpgroovy]

I dont know where all OCZ voltage limiting talk comes from , Mostly nonsense , though I did try to explain what our EVP ( what people mistake for voltage limiting ) really is. after that people said I was dodging the question , at which point I gave up trying to say anything about it.

Chip Fabs are the only ones doing voltage limiting on chip and they do it for pretty specific reasons , such as limiting the power draw , limiting heat output , ect ect ...


We can only control that on chips that are made for OCZ...

[GenTarkin]

arent you guyz afraid of killing ram at 3.8v LOL or does that not happen anymore? that voltage just seems outrageous to me.

[trans am]

arent you guyz afraid of killing ram at 3.8v LOL or does that not happen anymore? that voltage just seems outrageous to me.

My question is if you run 3.8v etc. can that damage the ATHlon 64 cpu since it has the mem controller integrated on the cpu? I've been running 3.4v 24/7 and so far no problems.

[lalPOOO]

I don't think there will be a lot of people running 3.8v. Probably more then there would have been 6 months ago, with no boosters out, but i don't think most people will go past 3.3v or so. The people that do run 3.8v generally know what they're doing :)

[WeStSiDePLaYa]

i would defiently say this ram is better than bh-5.

[eva2000]

sooo that die revision enough for OCZ to tag on rev2.1 to the label to differentiate future batches of OCZ PC3200 Plat rev2s?

[WeStSiDePLaYa]

sooo that die revision enough for OCZ to tag on rev2.1 to the label to differentiate future batches of OCZ PC3200 Plat rev2s?


yeah i would think thats a good idea. people wont like the fact that there is superclocking ram labeled the same as good but not super ram.

[morbidj]

You two make a very good point, so whats the deal OCZ are you gonna help us out with ID'ing the the newer mems so we dont have to go the Mushkin route in search of the Super TCCD?

[megahurtz-oc]

ya really i was about to buy some mushy. ill wait and see what ocz says.

[saaya]

if I could get some 1:1 DDR 600 stable ram at 2-2-2-5-T1 I would spend 400.00 on a pair without blinking. I wonder what my OCZ 3700EB would be worth if this happens??

your heard the man! :D

and hes not the only one, i can tell you :D

[Lithan]

Not many options for aircooling ram Build out and you limit how many sticks you can use, build up and you start blocking drive bays in most cases.

[SHiRaKaWa]

ALL I can say is thanks to ryanpgroovy, other ocz guys, colin and everyone who helped solving this mystery.
I'm new here in this forum.
And in these few days that I've joined: the mushkins were asked, the reviewer came, and the engineers worked on the question.
Certainly this forum is full of great people and is more than just the word "beneficial".
It is simply THE FORUM ;)

[WeStSiDePLaYa]

ALL I can say is thanks to ryanpgroovy, other ocz guys, colin and everyone who helped solving this mystery.
I'm new here in this forum.
And in these few days that I've joined: the mushkins were asked, the reviewer came, and the engineers worked on the question.
Certainly this forum is full of great people and is more than just the word "beneficial".
It is simply THE FORUM ;)


yeah you will find this is a pretty interesting place. people from all different coutnries, races, and religions brought together by one truly fun thing, overclocking! :rocker:


and lithan, there is one easy way to almost double surface area on the heatspreaders. ive mentioned this before, its to use a wave or crinkle pattern on the outer surface of the spreaders. this little diagram you can see how much more area there is compared to a flat design.

normal spreaders. ---------------

my vision /\/\/\/\/\/\/\/\/

of course the gradient doesnt have to be so drastic. ;)

[saaya]

ALL I can say is thanks to ryanpgroovy, other ocz guys, colin and everyone who helped solving this mystery.
I'm new here in this forum.
And in these few days that I've joined: the mushkins were asked, the reviewer came, and the engineers worked on the question.
Certainly this forum is full of great people and is more than just the word "beneficial".
It is simply THE FORUM ;)

welcome to XtremeSystems :toast:

yeah, thats why i love this place so much :) many many great, nice, friendly, helpfull, wise, talented and smart people :)

and now your a part of it :D welcome to the sickness! :D
make yourself a home and have fun!

if you have any questions or problems feel free to pm me or any other mod :)


great work ryan, eric, steven, tony and colin :toast: and sorry again if i sounded a bit harsh towards you colin :toast:

[corrupt]

Well I just ordered some OCZ 3200 rev 2 before I read this thread, I should have them tuesday night and I will give them some voltage and let you guys know how they do.

[saaya]

yeah you will find this is a pretty interesting place. people from all different coutnries, races, and religions brought together by one truly fun thing, overclocking! :rocker:


and lithan, there is one easy way to almost double surface area on the heatspreaders. ive mentioned this before, its to use a wave or crinkle pattern on the outer surface of the spreaders. this little diagram you can see how much more area there is compared to a flat design.

normal spreaders. ---------------

my vision /\/\/\/\/\/\/\/\/

of course the gradient doesnt have to be so drastic. ;)

:toast: :)

about the heatspreaders, keep in mind that they have to be flat on some spots to contact the ic's!

so maybe /\/\/\____/\/\/\____/\/\/\ ?

think that would look really nice as well :D

the best would still be if ocz or other manufacturers would just epoxy thermaltake copper or alu ramsinks to each ic :D but im sure that would cost quite a bit more...



Well I just ordered some OCZ 3200 rev 2 before I read this thread, I should have them tuesday night and I will give them some voltage and let you guys know how they do.

cool! make sure to post your results :)

[STEvil]

I pushed 3.9v into my bh-6 with an 80mm over top and they didnt get too hot, but with even 2.8v and no airflow they were cooking.

Same went for my tccd's too, actually..

[WeStSiDePLaYa]

what i mean saaya is just the outer part. imagine this as one solid peice of metal. the ram and another spreader. really messed up scale though. ;)

remember, if you ocz guys think this is a good idea and use it you could just send me a set of ram to repay me. ;)
/\/\/\/\/\/\/
-----------
-====-====-
------------
\/\/\/\/\/\/\

not sure about how much extra cost that would add though. and yeah if the HS were made better so that they contacted the chips better than a TIM could be used. like corsairs pro series, the chips use a TIM. alot of the thermal pads suck. the mushkin ones have a fabric. the best ones ive seen was on a pair of aftermarket heatspreaders i bought. the pads were made by 3M and they seemed to work very well.

[saaya]

i think the ocz heatspreaders also use a tape from 3m :) no?

and i think it would be really hard to get the folded metal to make good contact with the straight metal that contacts the ic's dude...

[WeStSiDePLaYa]

i think the ocz heatspreaders also use a tape from 3m :) no?

and i think it would be really hard to get the folded metal to make good contact with the straight metal that contacts the ic's dude...

yup that is true, but some heatsinks use bonded fins. also they could cut the grooves.

i think the best and easiest thing would be getting a good way to replace the pads with TIM. only probably with that is then ocz would be getting sticks back that were short circuit by some goof thinking it would be a good idea to smear them with half a tube of AS5.

[Geforce4ti4200]

looks like bh5's days are numbered! however, let me point out eshbach hit 285fsb at 2.5-3-3-7 at 2.8v, youd need 260fsb 2-2-2-5 to match his bandwith! sometimes relaxing timings for higher fsb is just as good and at much less volts too

[eshbach]

looks like bh5's days are numbered! however, let me point out eshbach hit 285fsb at 2.5-3-3-7 at 2.8v, youd need 260fsb 2-2-2-5 to match his bandwith! sometimes relaxing timings for higher fsb is just as good and at much less volts too


umm, me and every other person with TCCD on brainpower.

anyway, i'm still interested to see what the new TCCD does at 2.5-3-3 with volts :)

[WeStSiDePLaYa]

geforce, im not even going to start on how messed up that is.

and yeah. get a booster and push the volts. remember these chips are safe to 3.6vdimm. ;)

[Geforce4ti4200]

most dont respond to volts and me and eshbash know theres 25fsb gap. why not just run 2.5-3-3-7 at 291fsb like eshbach is getting now all on the low 2.8v? whats the need to use high volts when 2.8v does? you want tight timings, thats what bh5 is for, but then you most likley wont hit 291fsb

[SPiKE]

Something is strange to me... I've tested two different pairs of TCCD on my nForce2 system.

They performed different. The Corsair XL reached 250MHz 2-2-2 with 3.3v of vdimm. On the other side, the PDP Patriot XBLK reached 260MHz 2.5-3-3 (they just didn't wanted to work at CL2 after 220MHz) with 2.9v of vdimm. Over 3v on the Patriot, the system turns unstable.

:confused: :confused: :confused:



http://www.hardware4geeks.com/pics/spika/memroundup/XL3.jpg

[STEvil]

different batches of chips on each

[TEDY]

so which ram then ...OCZ/MUSHKIN ? CORSAIR XL ?

i want also 250@2-2-2-11 3,3vdimm TCCD

[WeStSiDePLaYa]

so which ram then ...OCZ/MUSHKIN ? CORSAIR XL ?

i want also 250@2-2-2-11 3,3vdimm TCCD


right now i would say mushkin is your best bet. ocz still hasnt told us if they will revise the naming of the rev2 plats so i wouldnt take the chance of getting the older crappier tccd.

[uwackme]

Anyone checked a BEFORE / AFTER to see what ID numbers are on the actual chips? Surely some outward marking would tell which is which?

I'd luv to find 4x 512M sticks that run 250Mhz @tight...together in a 939 board :cool:

[TEDY]

buuuu uwackme , you and your 4 dimm slots :P

/me jokes :)

[TEDY]

transam did you got your MUSHKIN ?

they're 286$ now.....

[WeStSiDePLaYa]

transam did you got your MUSHKIN ?

they're 286$ now.....


damn, retailers are starting to get word about the mushkin. prices in canada have jumped like 30$ :(

[eva2000]

Anyone checked a BEFORE / AFTER to see what ID numbers are on the actual chips? Surely some outward marking would tell which is which?

I'd luv to find 4x 512M sticks that run 250Mhz @tight...together in a 939 board :cool:
yup still waiting for board makers to bring out an overclockers board that can take 4 dimms populated to the max! They did it when going from 845 chipset to 865/875 for 2 sticks, now I want 4 stick oc'ing :D

[TEDY]

damn, retailers are starting to get word about the mushkin. prices in canada have jumped like 30$ :(

really 100% ?

[death metal]

Wow, thanks for solving the mystery ryan ;)...that's awesome...

[Reefa_Madness]

Anyone checked a BEFORE / AFTER to see what ID numbers are on the actual chips? Surely some outward marking would tell which is which?

:cool:

I've looked at the pic of the chips in AnandTech's initial Corsair PC3200 / Samsung PC4000 review showcasing the TCCD chips and the ones posted by Colin of PCStats which are the ones used in his review that hit 255 at 2-2-2-5. The basic Samsung coding is the same (same "F" revision), although the numbers across the bottom are different, I could not read the ones in Colin's pic. The only obvious difference is the mfg date of 431 (approx 8-1-04) on the newer, voltage-loving revision. With a program like Everest (AIDA32) I believe that you can get the mfg date displayed so that could clue you in if you have access to the ram, like if you are purchasing a previously owned set or if you are willing to buy and test, then pay the 15% restock fee if it is the "wrong" revision.

For lots of people with unmodded boards or no Booster, these new chips will not have a significant impact, however, for those with voltage at their disposal, this is big time news. I suspect that within a few months max, they will all be of the new revision if you purchase from a retailer with decent volume.

Edit/Correction:

My mistake, it is not AIDA32, but CPU-Z (ver. 1.24) that displays this date in the SPD tab. I guess I just crossed up these two tools when I posted. Luckily one of the guys caught it.

[STEvil]

431 is 8-23-04 I beleive.

[WeStSiDePLaYa]

really 100% ?

theres only like 3 places that carry this ram in canada. and its all jumped. makes me sad. :(

[TEDY]

order directly to muskin then.....around 260 $ :D

[Wingz]

are these the Muskins you guys are talkng about - Link (http://www.pc-memory-upgrade.co.uk/new-prod/mushkin-pc3200-level2-rev2.asp)

[TEDY]

yes

[Wingz]

yes

i've ordered some Gskill ram but i think i'll call them tomorrow and change to these Muskins. Need to find some reviews or test results of this ram.

[Nanobot]

With a program like Everest (AIDA32) I believe that you can get the mfg date displayed so that could clue you in if you have access to the ram.


I'm not seeing any mfg date with Everest(v1.50.187)

This is the only info when I right click SPD and make a report:

[ DIMM1: OCZ OCZ4001024DCPER2-K ]

Memory Module Properties:
Module Name OCZ OCZ4001024DCPER2-K
Serial Number None
Module Size 512 MB (2 rows, 4 banks)
Module Type Unbuffered
Memory Type DDR SDRAM
Memory Speed PC3200 (200 MHz)
Module Width 64 bit
Module Voltage SSTL 2.5
Error Detection Method None
Refresh Rate Reduced (7.8 us), Self-Refresh

Memory Timings:
@ 200 MHz 2.0-2-2-5 (CL-RCD-RP-RAS)

Is there is a way to tell what date/series the modules are?

If there is,could someone please post it,it would be nice to start a database on a seperate thead with oc+voltage results above 3.0v.

[Rabbi_NZ]

interesting

[LANjack]

Does anyone have any word on whether or not the gskills are using these yet?

[STEvil]

read the numbers and letters off the top of the ram chip, that will give you production year/week.

[twist]

ive got two 512 sticks of corsair 3200XL coming in the post, retailer informed me theyve just got them in today... ill see how they do.

[PnoT]

damn, retailers are starting to get word about the mushkin. prices in canada have jumped like 30$ :(


Haven't seen it around but which mushkin are you guys talking about?
(edit: seen the link above after i realized a 2nd page was here duh)

Is it the new TCCD high volt stuff or just tccd?

Think this is it as well on newegg Link (http://www.newegg.com/app/ViewProductDesc.asp?description=20-146-355&depa=0)

[sparkie34]

Haven't seen it around but which mushkin are you guys talking about?
(edit: seen the link above after i realized a 2nd page was here duh)

Is it the new TCCD high volt stuff or just tccd?

Think this is it as well on newegg Link (http://www.newegg.com/app/ViewProductDesc.asp?description=20-146-355&depa=0)


The newest batch of tccd

[Reefa_Madness]

I'm not seeing any mfg date with Everest(v1.50.187)

This is the only info when I right click SPD and make a report:

[ DIMM1: OCZ OCZ4001024DCPER2-K ]

Memory Module Properties:
Module Name OCZ OCZ4001024DCPER2-K
Serial Number None
Module Size 512 MB (2 rows, 4 banks)
Module Type Unbuffered
Memory Type DDR SDRAM
Memory Speed PC3200 (200 MHz)
Module Width 64 bit
Module Voltage SSTL 2.5
Error Detection Method None
Refresh Rate Reduced (7.8 us), Self-Refresh

Memory Timings:
@ 200 MHz 2.0-2-2-5 (CL-RCD-RP-RAS)

Is there is a way to tell what date/series the modules are?

If there is,could someone please post it,it would be nice to start a database on a seperate thead with oc+voltage results above 3.0v.

My mistake. I thought it was AIDA32 (working from memory...bad memory at that), but it is actually available in the SPD tab of CPU-Z (ver 1.24). I just got the two tools crossed up in my head when I posted. I just went back and verified both that it is in CPU-Z and not in AIDA32.

Sorry for the goof.

[Reefa_Madness]

431 is 8-23-04 I beleive.


I just approximated as follows:

13 wks per quarter x 2 qtrs = 26 weeks (June 30th)
Add 5 more weeks for July = 31 weeks, resulting in the approximate August 1st date in my post. I was not looking at a calendar and still haven't.

[trans am]

Guys wtf? I ordered this from newegg last week N82E16820146355. Now when you enter it into newegg search nothing comes up. I must've ordered the last of the old stuff on accident. also I noticed it's costs only $286.00 now. I paid $291 just 2 days ago. http://www.newegg.com/app/ViewProductDesc.asp?description=20-146-355&depa=0 I think I will test these then I might have to rma or just wait for the new OCZ to come out.

(edit nevermind, it comes up, but it's still $5 cheaper today.

[georgesod]

So any news from testers yet? :slobber:

[LikwidKool]

Man I have spent to much money lately, but I need some tccd. This just makes me more hungry for them. The pc3700 rev 3 I have do good at 2.5-3-3 but don't do the low latencies.

[Rabbi_NZ]

Welcome to XS Reefa_Madness, it's great to see you here dude... your help will be very much appreciated round these parts!

[Reefa_Madness]

Thanks for the welcome. As for my being able to help out...only if I can keep my facts (and my programs) straight, huh?

[Rabbi_NZ]

Thanks for the welcome. As for my being able to help out...only if I can keep my facts (and my programs) straight, huh?
Better you help and make the occasional blunder than sit back and contribute nothing...
A bit off topic, but I am getting a DFI LanParty UT 250GB soon Reefa, I'll let ya know how the Buffalo 5B-C does with 3.3v up its keeeshta on a A64 platform :)

[Ubermann]

My OCZ rev 2 i bought is over 1 month old and does simliar results with high volts.

[Reefa_Madness]

Better you help and make the occasional blunder than sit back and contribute nothing...
A bit off topic, but I am getting a DFI LanParty UT 250GB soon Reefa, I'll let ya know how the Buffalo 5B-C does with 3.3v up its keeeshta on a A64 platform :)

How's this for aggravating. Before posting the CPU-Z "correction", I checked it out on an XP rig, running Win 2K and 2 x 256 of Samsung PC2700. It read/displayed the ram's mfg date in the SPD tab. I just received an OCZ RMA a couple of days ago (PC3200 Plat Rev 2) and figured I would just check these guys out to see if I got lucky. Swapped it out and that darn box that displays the info in the SPD tab is just "gray'ed out". Nothing displayed. Is it because the ram is new and not included in CPU-Z's database?

So much for my easy way to check dates, huh? Rabbi, I'm off to a great start...two strikes against me now.

I too decided to play the A64 game so I'm currently waiting for an Socket 939 3000/MSI K8N Neo Plat. I've been reading up, including the thread here at XS on that board, just trying to get up to speed. About 10 days ago I placed an order for the DFI and a Sempron 3100, but it was on a whim and I cancelled it the next day. I went back to my original plans, which was to just get an entry level socket 939 rig to play with. The 3000 fits the bill nicely. Wishing us both luck and high overclocks.

[trans am]

My OCZ rev 2 i bought is over 1 month old and does simliar results with high volts.

lets see a cpuz screenshot please :)

[TEDY]

transam how long newegg needs to deliver mushkin of yours ?

has been like 5 days so far, right ?

[Ubermann]

Its in the ocz rev 2 thread.
It max 220 at 2-2-2 at 2.9 volt.
And 232 at 3.5 volt.
So the higher volt does something.
I always run it 3.1 nowadays.

[freshp]

@Ubermann What is with a higher fsb? 270? 280? Timings? Volts?

[mad mikee]

For those that are willing please list the 4 digit code for your TCCD
mine is 0431 (PQI)

NEXT! :D

[Ubermann]

http://www.xtremesystems.org/forums/showthread.php?p=574370#post574370

Check, those was my last tries before booster died.
Posted 4 pics at the end of the tread.
But as i can see it, its only Cas 2 that like volt over 3.1.

[trans am]

transam how long newegg needs to deliver mushkin of yours ?

has been like 5 days so far, right ?

I thought it was going to come today, but it wasn't on the truck. I am guessing tomorrow for sure. I'll keep you posted.

[idontknow]

.............. I think I'm a lucky one........ 2 gigs of this stuff sittin infront of me :D

[WeStSiDePLaYa]

.............. I think I'm a lucky one........ 2 gigs of this stuff sittin infront of me :D

to bad 4x512 sticks oc like crap

[idontknow]

thats why im testing for the best 2

[trans am]

thats why im testing for the best 2

so what happened?

[idontknow]

seems that they may not be the higher voltage ones........ I'm not 100%

best stable voltage to get higher mem scores = 3.1-3.4vdimm with this stuff

[WiCKeD]

Alright after a few in the lab we figured out the awe inspiring mystery of TCCDs that do well with higher voltages

It appears that samsung made some minor die revisions and that newer production chips do perform better with high voltages
So uh, ryan... what should we be looking for? :confused: :stick: "431"... "Rev.F" or is something that's not even listed?

mad mikee, what revision are your chips?

[Jesus]

A bit OT but there was talk bout the design of the Hetaspreaders my idea is following on to the stock HS you apply Heatsinks, copper heatsinks, or to be precise REVOLTEC RAM Heatsinks ultra thin finns and there bout 2,5-3cm high (1-1 3/4inches), tohose on to teh psotiton of teh chips and a fan ove rthem. Killer!ASAP i finish the project of applying my old Alpha Heatsink onto the NB i'll try this trick out

[STEvil]

431 or better would be my guess. Does samsung only have one factory that makes TCCD? Maybe one is making better chips than another if they have two or more?

[trans am]

431 or better would be my guess. Does samsung only have one factory that makes TCCD? Maybe one is making better chips than another if they have two or more?

Ok guys, I got it. But I've been posting in a different thread. Instead of taking up 2 threads, I'd like you to just click here and read:
http://www.xtremesystems.org/forums/showthread.php?t=44527&page=7&pp=25


SO far no good! and it's the same date from 431

http://www.xtremesystems.org/forums/attachment.php?attachmentid=18818

I tried all the tests at 1T and got really bad results. Is tccd doing high speeds at only 2T or should this ram do it all at 1T?

**update, this ram won't even run it's advertised timings of 2-2-2-5 or 2-2-2-6 1T. at 2.8v. it will run these timings if I run it at 166mhz only. WTF! I either killed it with the volts, or it was DOA to begin with.

[Revv23]

yeah i dunno i think ill hang onto my BH-5 a bit longer, at 3.1vdimm these little KHX3000's run 240 2-2-2-5...

i need an A64 rig and a ram booster man, im tired of not being able to push these things...

(note - i cant break 240 period, not even with 3.3v, not sure whats holding me back, either proc, mobo or my PSU... who knows...)

[saaya]

ask the ocz guys :)

they tested a bunch fo tccd and should be able to tell you what tccd seems to like volts more than other :)

[trans am]

ask the ocz guys :)

they tested a bunch fo tccd and should be able to tell you what tccd seems to like volts more than other :)

There is still something wrong here.
I looked at some gskill 3200 2-2-2 tccd and noticed it has the same 431 date and K4H560838F code. I really think there is more to this than just the date and the ID code. The answer must lie in the last set of small code on the bottom of the ic. The ones that nobody can seem to read.

here are 4 different ram pictures showing TCCD from 431. the 2nd to last one is the one that Colin reviewed. What do they all have in common?
What are their differences? Did anyone notice that the last code on mine they are all different? But notice the kingston(2nd one) all the numbers match.

http://www.xs4all.nl/~famchow/Memory/Gskill%20PC4400/Gskill_pc4400.jpg


http://www.hardwarezone.com/img/data/articles/2004/1230/ram-chips.jpg

http://members.lycos.co.uk/csmo/PCStats/Mushkin/Mushkin_2.jpg

http://www.xtremesystems.org/forums/attachment.php?attachmentid=18818

[TEDY]

damn transam that sux :(

i might not order mushkin then...

[trans am]

damn transam that sux :(

i might not order mushkin then...

It could be that I just got bad ram from the beginning. It wouldn't even work at ddr400 @2-2-2-6 So i'm thinking it was just DOA. Or I fried it by running 3.3volts into it. But I thought TCCD will take up to 3.6v, so I'm leaning towards the DOA assumption. I'm going to call newegg and get another set. I noticed the prices dropped on them. In past experiences this ususally means they got a new shipment and the supply is great, so the prices are lower from the distributors.

[HKPolice]

Uhhh.. BH5 is rated for upto 3.6v, not TCCD.

[georgesod]

So isn't anybody out there with good results @2.2.2.5?

[trans am]

Uhhh.. BH5 is rated for upto 3.6v, not TCCD.
Westsideplaya posted a chart from Samsung earlier that said TCCD also is supported up to 3.6v. Uhhh..

(quote from westsideplaya)
some more info to help decode some product stepping. not sure how usefull this is.
http://www.samsung.com/Products/Sem...fo/DDR_code.pdf

EDIT: some more really giberish stuff. mentions slew rates. hopefully someone can put this stuff to use.
http://www.samsung.com/Products/Sem.../k4h56xx38f.ibs



also good news. says max voltage is 3.6v

[TEDY]

link doesnt works :D

[berk]

Are you taking into consideration the PCB design with these chips too?
There was a post about the Brainpower PCB design maybe a month ago,OCZ rev.2s had a
great advantage over other kits tested with the same TCCD chips.

http://www.legitreviews.com/article.php?aid=53&pid=3

[trans am]

I just got off the phone with mushkin. They are going to see if they can find the latest batch of ic's they got form Samung and do a test. I told them about my problem and we talked about the batch numbers(last row of numbers) it looks like kingston does some very labor intensive picking if all the batch numbers are identical. I wonder if OCZ has same batch numbers on the ics or if they are different like the mushkin/

[TEDY]

what do you mean about kingston...cause that deal 265 - 45$ is really great...tempting hehe

[trans am]

what do you mean about kingston...cause that deal 265 - 45$ is really great...tempting hehe

http://www.hardwarezone.com/img/data/articles/2004/1230/ram-chips.jpg


look at the bottom row of codes. This is the batch number. notice they are all identical. now look at the batch numbers on the other TCCD. they are all different. Kingston doesn't use brainpower pcb, but at that price with the rebate, I think it's worth it.

[TEDY]

yep kingston has all the same thanks for pointing it out...

[quicksilverXP]

Does all the same batch code really affect performance?

[trans am]

Does all the same batch code really affect performance?

that's what we are trying to figure out. I think the point is that picking the same batch number is very labor intensive and if the same batch number didn't matter, then why go through the hassle of sorting through them?

[duane26]

Hello to all, I to have currently been looking to get some new memory and was curious about getting the new TCCD chips so I emailed OCZ and this is there reply" Both the 3200 and 3700 use TCCD yes…they are speed binned differently that’s all.We have been using the high voltage loving TCCD for a few weeks now….;-)Tony" So all we have to do is find some 3200/3700 that was shipped in the last week, right? Probobly easier said then done but its worth a try :D

[WeStSiDePLaYa]

Uhhh.. BH5 is rated for upto 3.6v, not TCCD.
yes it is, please do your research. or at least read the whole thread as i posted this info.

and yeah maybe the chips with the new die revision dont have any different markings? tccd sure does seem to be a mystery.

[duane26]

Sorry, I have read the entire thread just not today. I have been reading so much stuff on the web it sometimes makes my head spin :rolleyes: Anyway, this is a mystery and the powers that be, that know, aren't telling so in the end it seems to be a "crapshoot" at best. I do think there is a better chance in getting the stuff in ocz then say munchkin.

[JDizzle]

I think someone stated in the other thread that Mushkin uses "new" brainpower PCB, could that be it?

[berk]

Thats what i wondered earlier on inthis thread,no one seems to have taken up on it,or it's moot.
I'm still in the dark,maybe its the pcb's, maybe its a certain kind of ICs.
Maybe both? :)

[trans am]

Thats what i wondered earlier on inthis thread,no one seems to have taken up on it,or it's moot.
I'm still in the dark,maybe its the pcb's, maybe its a certain kind of ICs.
Maybe both? :)

Well I talked to ryan at ocz today and he said they had some dates from 440 and 431. They were both a mixed bag as far as the volts go. He mentioned something about testing the high volt ones and setting them aside so they can relabel them, maybe a pc3200LL rev. 3 is on the way?

[WeStSiDePLaYa]

man this is really messed up. maybe i should just ask colin to buy his sticks. :p: so far mushkin is the only stuff we have seen put up 250 2-2-2.

man this makes my head hurt. looks bad though when ocz and others use tccd from different batchs on the same stick. the stick will always be limited by the slowest IC on it.

[JDizzle]

Colin also said he had Geil that ran high FSB with tight timings alsol

[esoteradactyl]

damn transam that sux :(

i might not order mushkin then...

i got some yesterday also and have similar results as trans am. stuff really sucks.

[trans am]

i got some yesterday also and have similar results as trans am. stuff really sucks.

yep. I rmaed mine. I told newegg it wasn't compatible with my mobo and they gave me a refund. THis is what I ended up getting:

http://www.mwave.com/mwave/viewspec.hmx?scriteria=BA20672

$267 plus $45 mail in rebate!!!

http://www.xtremesystems.org/forums/showthread.php?t=43613

read more here:
http://www.xtremesystems.org/forums/showthread.php?t=44816

[esoteradactyl]

yep. I rmaed mine. I told newegg it wasn't compatible with my mobo and they gave me a refund. THis is what I ended up getting:

http://www.mwave.com/mwave/viewspec.hmx?scriteria=BA20672

$267 plus $45 mail in rebate!!!

http://www.xtremesystems.org/forums/showthread.php?t=43613

exactly what i did. im even going to go grab some kingston tomorrow from mwave :p:

[trans am]

exactly what i did. im even going to go grab some kingston tomorrow from mwave :p:

i just sold my Mushkin BH5 on ebay. the auction just ended! got $360!! I can't believe it! This payed for my new ram and 939 motherboard!

I would rather support ocz, but I need something until that new voltage loving stuff comes out. Hurry up Ryan!

[WeStSiDePLaYa]

i just sold my Mushkin BH5 on ebay. the auction just ended! got $360!! I can't believe it! This payed for my new ram and 939 motherboard!

I would rather support ocz, but I need something until that new voltage loving stuff comes out. Hurry up Ryan!



what!!!!!! and i though mt 2x512mb mushkin bh-5 for 2x512 3500EB and 130$ was a good deal for me.

[TEDY]

lol some nuts people on ebay :D

[Hybrid]

Hey guys i havent been here since i sold the prommy but i have been following this thread. Now im not 100% sure but most of the OCZ sticks just have the logo stamped on the heat spreaders? I ordered some OCZ 3200 R2 plat from newegg on monday and got them today. They have the OCZ badge on them like their website shows. Does this mean i have the good stuff? Thank in advance!

http://img.photobucket.com/albums/v126/hybridhb/ocz_r2.jpg

[trans am]

[QUOTE=Hybrid]Hey guys i havent been here since i sold the prommy but i have been following this thread. Now im not 100% sure but most of the OCZ sticks just have the logo stamped on the heat spreaders? I ordered some OCZ 3200 R2 plat from newegg on monday and got them today. They have the OCZ badge on them like their website shows. Does this mean i have the good stuff? Thank in advance!
try to run it at 2-2-2-6 1t @ 240mhz. if memtest86 pasees test 5 with 0 errors, you have the good stuff.

[Hybrid]

At how many volts though? Im limited to 2.85 since i have no booster. Im on a neo2 and 3200 90nm setup (air cooled with xp-90 right now). So far its doing stock timings at 200mhz at 2.5v, something my corsair XL couldnt do.

[TEDY]

2.85 is not enough for 2-2-2-x @ 240 hehe

[Hybrid]

Yea i didnt think so, i guess a booster is next on my list :)

[trans am]

Yea i didnt think so, i guess a booster is next on my list :)

Don't worry about doing those tight timings yet. Try something like 2.5-3-3-7 1T @ 280mhz+ and use your 2.8vdimm.

[Hybrid]

Well as of now they are doing 2.5-3-3-10 at 2.85v 240 HTT. Dont know if thats stable yet, been running prime for about an hour in "Large FFT's". As far as i know the voltage loving sticks perform the same at lower volts right? Only the newer ones respond to high voltage?

[Revv23]

^thats what ive heard.

[burnhead]

What a wonderful time to be building a new comp :)hehe...me too....

how do you all cool your RAM´s? Do you put this Thermaltake Heatsink with included fan on it? Or do you just place a fan in front of them?!?!?!?! :stick:

man, :banana::banana::banana::banana:, I MUST HAVE DDR 600 2-2-2-5-1T!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!

[krampak]

I've just tried ocz booster to my Geil Ultra X tccd, nothing, they does the same at 2'7V than 3'5V, at least with 2'5-3-3 timings.

[rick_fx]

Ok, I just received my gskill pc4400 1GBle today. Since my mobo hasn't arrived yet, I was wondering if I have good sticks or if they need to be RMA'ed.

Date is 437 and the code is K4H560838F. I ordered from
As for the small markings on the bottom of the memory chips, all 16 have the exact same markings: H56 8 BMG099PX.

Trans am, ryan, any idea? Thanks

[jinu117]

Mine does 285mhz 2.5-3-3-7 individually. or 281 2.5-3-3-7 in dual channel. 2.6v. Haven't bothered higher voltage, etc as I am waiting on another new CPU (woot, I think I am jumping in FX stuff finally)

[Brum Man]

So I am I right in thinking, after reading the whole of this thread, that there is still no certainty to when Samsung made this minor die revision, and how we can tell by the heatspreaders or some other method?

That's a bit annoying. I'd like to get some also but if I can't be sure I think that would put me off for a while. I would really like to get the OCZ stuff. Would you rate the 3200 Plat.Rev2 above the OCZ 3700EB (what I currently have), for reaching high FSB (I have seen 270-300 floating about), and high bandwidth?

Also, as asked above, how do all mount your fans above your RAM?

Cheers (Ooooh look, my first post :))

[krampak]

I've just ordered an FX55, wanna see if I get the same FSB at 1:1 than with my winchie (to see if every chip gets a diferent fsb).

[STEvil]

I use zip-ties to hold the fan above the ram if its in a tower. If its just on a bench I just set the fan on top.

Welcome to XS btw, Brum Man.

[Susquehannock]

So I am I right in thinking, after reading the whole of this thread, that there is still no certainty to when Samsung made this minor die revision, and how we can tell by the heatspreaders or some other method?

That's a bit annoying. I'd like to get some also but if I can't be sure I think that would put me off for a while. I would really like to get the OCZ stuff. >>>>>>>

Ditto!

Looking to buy some new RAM. ... 1Gb (512x2) to replace the pc3700
256mb modules I currently have.
Guess it would be best to wait until there is some certainty as whether
the new TCCD chips are truely better.

[Ubermann]

I just superglued a 80mm fan on the side of PSU.

[Susquehannock]

Concerning the production week & voltage ...

Found some interesting stuff in the GSKILL forums:

http://www.gskill.com/forum/index.php?showtopic=105

'As I do a test in our TCCD.
ALL cycle:
TCCD 0425
TCCD 0431
TCCD 0437
These are all do not like high voltage.
If I want to overclocking 275MHZ or 300mhz or above.
To increase high voltage is no useful.
But I have seen in the XS,But I do not the chip.
Maybe I just guess that it is newest cycle(later 0437) or it is not the TCCD CHIP.

ex:0431 is 2004/31th weekend.
so 7x25/30~6(It is maybe manufacture in 5 or 6 month)

TCCD and high voltage mystery explained
http://www.xtremesystems.org/forums/showthread.php?t=44647


Like all of you see the extrmemperformance .
They are ddr550 2.5-3-3-7.
These is almost 0431 or 0437.

They can run ddr600 2.5 3 3 7 or even more.

As the newest edition,I must wait for samsung.
Ans I can tell end-user.

2-2-2-5 has 5 type Chip.
BH-5
TCC5
TCCD
IINFERION
MICRON-G-Gie.

....... ........... ..................... < goes on >


How do you like the acknowledgement of this thread? They lynk to XS
often over there. :)

Above thread also mentions voltage limiter on said modules.

Now we just sole in the TAIWAN ,USA and europe (0425 0431 0437).
ALL these tccd just suit 2.65~2.85v.
add greater voltage 2.85,the system will unstable.

As the 3.2~3.5~3.7~4.0.
because I have no "ocz bootster" to test it.

this is the "private guess"
I guess that maybe when it add 3.2 above ,it maybe boot.
but ismaybe guess.

or Hynix-:-dt"D5,micron ,They like high voltage as BH-5.

BEST REGARDS

This would explain why some are disappointed in their modules
with "431' chips don't you think?

[Ubermann]

Cas 2 is the only one that like high volts, was it that he tried ?

[Susquehannock]

Good question. The same member mentions trying cas 2 in other threads though. Lots of posts
with bencies around 300mhz over there. Seems these chips really open up
at cas 2.0 with some juice.
Also looking at some Chinese OCing forums about this subject. One was
up to ddr660 2.5 3-3-7 @ 2.75v.

Better yet, how about 350fsb?

DFI - LAN Party
3700+ ClawHammer
G.SKILL 256mb TCCD 0425

http://www.coolaler.com/ipb/index.php?showtopic=25354 :worship:

The cpuZ SPD page with mfg. date is there. (someone was mentioning that earlier)

[esdee]

try to run it at 2-2-2-6 1t @ 240mhz. if memtest86 pasees test 5 with 0 errors, you have the good stuff.

TransAm i have to get a little oftopic but why test 5 and not 6? with my Bh-5 i can get no errors at test 5 and hundreds in test 6, whats so special about test 5? :off: <---sorry

[Ubermann]

I though oscar was the one getting 350fsb.
Maybe i read wrong somewhere.

[Zeus]

I though oscar was the one getting 350fsb.
Maybe i read wrong somewhere.

He was, as well as Coolaler.
Both with just one stick of 256mb though. ;)

[burnhead]

:banana:

[easypanic]

look at my A-Data vitesta DDR566 :D

[Joe Camel]

Posted this in the mem with TCCD thread, but it might "help" more here...


got my G Skill (from eXcaliberPC (http://www.excaliberpc.com/product_info.php?cPath=163_520_789&products_id=4242) ) today :)

chips are ALL (that i can see):

437
K4H560838F


but the (hard to read #s) are NOT all the same:

456 8 BMG 159 P1
...............160 PX
...............109 XX
...............197 BB
...............147 P2
...............175 P2
...............099 PX
...............138 B4

dont have my CPU or HSF yet, so i cant tell ya how they clock (all will be here friday along with some Mushkin 222 v2 :banana: )

[Susquehannock]

He was, as well as Coolaler.
Both with just one stick of 256mb though. ;)

Still, 350fsb is not to shabby.
Would be interested in how the 512mb perform as well. Also how they
do in DC mode. And speeds at Cas 2.0 would be great too since that
would be what I'd try in my nf2 setup.. :D


look at my A-Data vitesta DDR566 :D

Nice one! :toast:

[Susquehannock]

Another review of the TCCD (this time in 'F' flavor) reaching 350fsb ...

http://www.x86-secret.com/?option=newsd&nid=771 (French page)

More benchies and PCB info w/image (and another reference to this thread) ...

http://www.hwtweakers.net/postp62826.html&highlight= (Italian page)

[burnhead]

good evening..... which of these RAMs is this "high voltage liking" OCZ EL 3200? the one with 2-2-2-5 or the one with 2-3-3-5 or the one with 2-3-3-6 or even the one with 2,5-4-4-7?? I have only a German website, but you´ll understand the product names...they´re the same as in USA... *grin* here´s the link: http://www.cssc.de/shop/advanced_search_result.php?keywords=OCZ

[LANjack]

Another review of the TCCD (this time in 'F' flavor) reaching 350fsb ...

http://www.x86-secret.com/?option=newsd&nid=771 (French page)


Was only one 256 chip, still impressive though. He said it was stable at 340. Revision F.

edited for dumb redundancy dumbness.

[aNabolic]

Ok, Just bought 2x512MB OCZ 3200 Plat Rev 2 from www.over-clock.co.uk. They state "Usually Samsung TCCD on Brainpower PCB". I have asked them to let me know when these were stocked so I have an Idea if they are the new ones or not.

I bought these to replace my old OCZ4200 PS - My current 3.4EE doesnt hit the FSB my old P43C used to so the extra room isnt needed. Ill be running 230-255 FSB so Im hoping with my Booster I should get much tighter timings with the new 3200. The 3200 will be teamed with my modded P4C800-E DLX.

Ryan, would you know if I can expect the good'uns or not ?? Would you know the restock of UK suppliers ??

Thanks for any help you can provide, you OCZ guys do amazing work @ xtremesystems. Amazing forums.

Thank you,

aNaBolic

[Joe Camel]

been playing with my Mushkin 222 V2 and G-Skills @ 3.3V 3,3,3,11,16
(3500 winnie, DDR Booster, Abit AV8, XP-120 w/ NidecBetaV)


Mushkin maxed out @ 235 but the G-Skill hit 275! (using 275x7 1:1)

will try to tighten timings after I get a "real" multy (could only get 275x8 stable :down: )

testing 10x265...

o BTW, if you plan on running Vs like this you BETTER have a fan blowing on the RAM!!

[burnhead]

@joecamel: so abit av8 is good for overclocking?

[Rodzilla]

Wouldn't you know it! I bought ram and wasn't even looking for TCCC and found some! :lol:

Anyway... It's Patriot PC4000 512MB kit (2x256MB) CL3-4-4-8 Model#PDC5124000ELK
http://www.newegg.com/app/viewProductDesc.asp?description=20-220-004&depa=0


It goes up to 250MHz but I can't boot if it's not set at 3-4-4-8 (Crappy I know... not my system or RAM ;) ) I'm stuck at 2.8V on an IC7-G for now.

Chips are:
SAMSUNG 443
K4h560838F - TCCC
H56 8 CFI590HBC (same on all 16 chips)

Just thought I would give you guys the heads up!