Hello guys, I just bought the AIR value ram which using Samsung TCCC chipset on it, i wonder how much juice can we push on this baby??? At the moment i only manage to push it till 250mhz which = DDR500 with 3-4-4-8 timing. Is there any where to push it futher like....275mhz????

It's good that you could reach that speed, some who bought those rams only managed 225MHz.

Anyway, I think most TCCCs maxes out around DDR500

go for lower timings and 1T, sometimes faster than ddr500 at 3448, mine run benchmark for 2336 1T vs 250Mhz 3448 2T, the 2336 200Mhz win in the benchmark score using sisoft sandra...

I'm curious bout the BGA TCCC..anyone gav them a push?

ur result is good enough for value ram lorr..wanna push higher..burn it more

wut voltage are u running it at?

ye.. i thought TCCC similar like BT-D43... can mild OC and cannot push higher..

erm...

i think getting a tccc running at 250mhz is good enough....

can it go further ??

i read somewhere that the tccc max out at 260 mhz.

anyone tried ??

Thanks for all the replys,

lichyetan,

Actually i am running at 1T with 3448 at 250mhz, the default timing is 2.5-4-4-8 but i didnt really run it with try other timing at 250mhz. I thing i am sure is that it can be boot up using 2.5 n only can with 3. From your rig, i wonder wat setting u put on your cpu? My 3200+ only manage to get 2500mhz with 250x10 with 1.5v... the voltage is quite high for me!!! How can i get it stable for 2500mhz with lower voltage???


kuchingfight,

Mine is BGA TCCC!!!



GKWong,

I using the default voltage at 2.65v because i heard Samsung dont like high voltage so i didnt push it higher. Wat do u recomment???

Wow, BGA No wonder you're getting higher clocks than those TSOP ones

1. burn-in time
2. time to get workin on A64tweaker..
refer stickied guide to get to khnow all the timings
this one really needs the patience, as need to find best combo
[loose timings do help reducing the burden of mem controller]

memtest86 n blowing fans comes in handy
u need to try it urself to check whether ur set likes extra juice, or craps out with juice

This post has been edited by e-jump: Mar 31 2006, 02:19 PM

which do u think is better to cool down ram, using the copper heatspreader or ramsinks,

i heard that the CM Copper HEATSPREADER doesnt stick to ur ram, instead the tape that comes with the spreader is just a heat transfer element.. ??

is there any other heatspreader that does not stick to the ram ??

just in case of RMA ?

Ya i want to know wat kind of cooling method is better, cause at the moment i am using 2 80mm fan to blow at the RAM but i think it is noisy n i want a better solution. Like zahri say, is the heat speader functional or no use, i remember few years back got somebody say it is better cool without the ram sink or the heat speader. Is this true??? can somebody that using those cooler give their comment here???

i'd suggest go with fans, and only fans
IMO, heatspreaders r juz for looks.. its doesnt do the job well alone
say, if u only use heatspreaders w/o fans, the heat will dessipate to the heatspreader as its supposed to do, but it still need air to cool the spreaders

well, im using one of those coolers sold by allngap in bulk section.. n i asked for voltage regulator to control the ram speed

well, im using uccc, also by samsung.. n yeah, pumping up the voltage do make em damn hot.. but with proper cooling, i guess its ok.. but i gain a little with 2.8v, so il juz stick with 2.6v for 24/7 usage =__=

so ur saying,

its better to stay away from the heatspreader and just get a fan ?

hehehe....

maybe wud get one of those spreader's they do look nice, but performace wise, maybe they dont

blur2 situation here......

bare naked RAM is better, and adding a fan to blow around it will definitely much better

somehow i am still sceptical about the heat transfer provided by thermal tapes. i don't see how well it does the job.... doesn't look like it can absorb the amount of heat and transfer it to the heatspreader effectively.

heatspreader will be very efficient if there weren't any thermal tape used, and the heatspeader touches the chips directly... just like a heatsink on a cpu.

i also dunno, i succesfully oc to i it with the 278*9 without adding vcore, but mine is venice, maybe is the batch, but never manage to get 2.6 even 1.5 volts... currently running 2.4ghz wif 200mhz for ram, my ram never manage to go over 210mhz at 1T, so i make it to run at 2-3-3-6 at 1T... ur ram is very impressive for a value ram, cause my corsair pro series wif samsung -5 chipset is expensive and couldnt match ur ram i think, 250mhz@1T for a value ram is super worth bro...

is it possible to use thermal compund ?? AS 5 ? between the ramchips and heatspreader ??

just a thought

Thanks for your complement lichyetan , my cpu is also venice core but may be i'm not lucky enough to clock it high with default heatsink. I am thinking of getting a good heatsink like ninja n see how it perform. As for the RAM, i think it is good enough for me, but like zahri say, is like possible for use to use thermal paste on the heat speader n removing the default thermal tape??? Is any one out there try that before???

erm...

i think u shud change ur heatsink, maybe u wud be able to clock it higher....

well, as for the heatsink, maybe it wud be better off without the spreader...

just bare rams and fan over it to cool it off

i just can think 1 solution if u guys going to use heatspreader for the rams. Use non-conductive aka mx-1 thermal grease.

It's sure look goods, but, heatspreader are meant for high vdimm one aka bh5 / ch5. Ive been running my bt-d43 253@2-2-1-11 1T 2.9v back while. it didnt get hot ( as hot as need some cooling ).

Just my 2 cents.

Erm.. Actually i agree with satanhead....

i've been pumping 2.9v to my DT and BT-d43 and they dont feel hot..

WArm yeah .....

thats why i decided to just get a fan blowing at it, even though i dunt think i need it...
coz my top blow hole is already aligned nicely on top of the ram....

ohh well...

err.what the voltage limit for hynix?

Looking to the D43 datasheet, 2.7v is the limit. but, oh well. with 2.7v managed to get around 230-235 onli. n 2.9v 253.... i choose the later, LT warranty summore.

hehehe..just xtra 0.2 v to gain20 Mhz more..btw till now i never heard someone burn their hynix ram when apply 2.9 v voltage

who burned their hynix ??


i'm using mine at 2.9v also

230 - 235 is not bad for 2.7v



need to push mine more

He said "till now i never heard someone burn their hynix ram"

i never did burn my 1GB sitcks of d-43 eventhough they were at 2.9 volts during my NF7-S days

and no matter how much voltage you give d-43s, they will never budge from max of 240FSB for my pair.....sheeeeeeesh =.="

my samsung tccc runs at 2.8v vdimm before, the temp is warm oni...
btw...wut is the best timing the ram can go with stock 200mhz?

Sorry for the misscomunication....

Tersalah tafsir.... Something wrong with me...... maybe... LACK OF SLEEP

u overclock urself then din run prime test to test ur stability...

That's a nice explanation

Bro...sorry to dissapoint you but i dont think you got a Samsung Chip on your rams.

Btw have you tried reading it with Everest or CPUZ?

Antonio_zth,

It is a Samsung TCCC cause it is a value ram and you can see the chip!!! Although in cpuz you cant see it but physically it is a Samsung TCCC!!! If not, what chip do u think???

hihihi..you are too young bro....the details on top of the chip can be made/remarked...

try viewing first in softwares such as lavalys Everest (www.lavalyst.com) CPUZ (Cpuid.com)...then u understand what i mean...

about heatspreaders.

i actually tried once, removing the heatspeader, the tim between the heatspreader, lapped it, then clip it back to my ram. it's a corsair xms heatspreader, the one that uses the thin strands of metal for clipping, somthing like paper clip, so i managed to get most contacts between the chips on the ram and the heatspreader. used some arctic ceramic (silver is conductive, i tried to stay out of troubles..) for extra contact.

to test it out, i used memtest for burn in. i cant remember the results coz i forgot to note it back then but.. roughly...

without heatspreaders
2-3-3-5, 2.8v, reached 240fsb stable
at 245, got few errors every several loops of memtest, roughly 1 or 2 errors per 4/5 loops.

with direct fan blowing on bare rams,
2-3-3-5 2.8v 250fsb stable, 0 errors after 8hrs, removing the fan instantly produce errors..

with heatspreaders.
2-3-3-5 2.8v 240fsb stable.
2-3-3-5 2.8v 245fsb, no errors for 10pass, then loads of errors.
- the heatspreaders managed to remove the heat fast enough when the ram modules arent really hot. but after heated up, seems like it cant dissipate the heat fast enough to cool down the module, ends up heating the ram.

heatspreaders with fans.
no problem running 2-3-3-5 2.8v 250fsb for 8 hrs.. takes sometime to produce errors after removing the fans, but generates more errors per loop compared with bare heatsinks.
i cant seem to push it higher with the same timings, possibly the limits of the ram at that timing, or bcoz pat was enabled.. didnt bother to disable and test it bcoz of the performance difference is too much with pat disabled.

so, imho i'd say that even with proper contact, heatspreaders dont have enough surface area, or heatsinks to help dissipate enough heat, especially without airflow. even with airflow, it is not really efficient. just sticks with bare rams + fan.

as for the difference between the heatspreaders with their default tim and arctic ceramic, i'd never actually did any comparisons, but i can bet that with arctic ceramic the heat can transfer to the heatspreaders faster since it heats up a lot faster, compared when using thermal tape, or the original tim.

tim = thermal interface material.

ermm...

so its better to run the rams bare with a fan then with heatspreader....

yeap..indeed u are rite.
actually AIR OEM ram with samsung tccc chips are Mosel Vitelic 5ns ram chips that are re-labeled.

read this,
picked from atomicmpc.com

Hi antonio_zth,

i will try the everest tonight, i didnt know that ram can be relabeled. Thanks for your information. But i dont think it is a big deal cause it can reach 250mhz then i am happy with it since it is a value ram.


GKwong.

Since you are a AIR owner, how about the chip on your ram, is it true that it is a fake samsung??? R u using the everest to test it?

i use cpu-z and tested it is a Mosel chip. but actually many high-end ram also uses this chip, and sometimes they re-label it.

erm..

so the AIR Ram uses MOSEL Chips....

what the best timing and FSB u got GKWONG ?

Air Rams are considered 1 step a bit higher of OEM ram. Actually it is an OEM but package for sale purpose...Thats why you wont get what it says on top of the chip.

Its kinda fraud also...but what to do...Air rams are cheap maaa....and usually an 'average joe' wont complain as long as it works...

Antonio_zth & GKWong,

After i tested with everest, it cant get any info on the chip, as for the cpuz, it show all zeros under the manufacturer column. So wat does this mean? By the way GKWong, wat type of chip is your ram, is it a BGA???

huh...urs wan is not same with mine wan. mine wan under manufacturer wrote MOSEL, Serial number FFFFFFFF, Week 255 Year 255. it is a tsop.

GKWong,

So my RAM is a true samsung TCCC or a fake one?


Does anyone have any idea???

hihihihi.......welcome to Air!....

woah...

no wonder in CPU-Z it says nothing for the ram info...

most propably most of the chips are MOSEL chips relabeled to Samsung TCCC chips ?

Isn't it cheating to relable the chips? I mean it's okay if they relabled it to AIR but to relable it to Samsung TCCC?

At least they didn't use TCCD

if tCCD kantoi gille...

ah, Air, the brand that once upon a time, a long time ago, kena kantoi for selling fake/remarked AMD processors.

its amazing though, that these dudes are still in business.

And yet they are still doing the same thing up till today Wonder if Samsung will sue them

yea...lets complain this to samsung about other company faking their chips! let samsung sue them!. wtf!

Air??i try it b4,BrainPower PCB 808...Samsung TCCC Chip...
using 1 week to buring..2.7v 260@3-3-3-8...

no like UCCC best,bt using value price can get that performance is ok lah,anyware, got seling 20~30pcs,no any problem,but oc i juz try 6 stick...
4 stick can do 260@3-3-3-8,2 stick max at 245@3-3-3-8..

n air also got the BGA chip of DDR-400...
(but i no order,n hve see it)

260MHz, 3-3-3-8? Looks to be TCCC le Maybe some are real TCCC while some are relabeled

dun't know,i juz push up Mhz,no push the latancy
,(coz lazy to do setting...)

Looks like air to me....

260 is nice for a value ram....

but i didnt really hav the chance to oc my ram yet, coz my proc ady max out. maybe after got a barton oni can test it.

i just realized that my processor is athlon xp 2200+

hehe.... i tought mine was 1800++

gkwong, ur prox max out at 200 ?

coz if i raise the fsb more, prime will fail.

yup...is better than DT-D43....

n at DFI NF4 also can stable to using...no like Hynix will crash with DFI Mobo...

Are you sure it's the proc's fault or your mobo? Just guessing since I've never used an Athlon XP before.

I guess the AXP dont like it. From what i seen. AthlonXP got alot of problem when u wan to go beyond 240mhz even on a Good UCCC or UTT BH5 with 3.2v. I guess the prob come from different memory controller from the A64.

just order it 2 sticks...

Below r some pictures of my RAM:





Do you guys think this is a fake chip???

its look like fake samsung chip...

my chip not same with ur wan...

abit nf7 shud not hav problem on fsb higher than 200mhz i think...

is not a fake chip just is a tiny bga chip..

is a bga chip, but is not from samsung....