Just what is TCC5?

[Reefa_Madness]

Everyone seems interested in TCC5 all of the sudden (including me), especially now that it is being used more and more as a replacement for the EOL TCCD.

How about you post any info you may have on these chips and ram using these chips here so that guys can get an idea of what to expect?

I'll start with this link to a review done by Anandtech when Samsung first released these chips and then lets take it from there.

http://www.anandtech.com/memory/showdoc.aspx?i=1998&p=2

edit:
It is my understanding that the "F" revision is better than the "E" pictured below. Any info along these lines would also be welcomed and helpful.

[gundamit]

Wow. That article is 14 months old and yet it seems like ancient history. I'm surprised TCC5 has been around that long. I thought it was a replacement rather than a precursor to TCCD. Does this mean Samsung is not firing up the TCC5 machines for new production but is instead just picking through older memory that has been sitting in inventory?

[(sin)morpheus]

Maybe so, I was wondering what this TCC5 was too. I don't remember seeing it in the past, maybe I just wasn't watching carefully enough. :D

[saaya]

search button? :P


from my understanding samsung was binning the same memory chips either as tccc tcc5 or tccd depending on what timings and speeds they would run, right?
samsung is slowly reducing their tccd production, but this only means they no longer sell chips labeled as tccd. the only thing that is changing is that samsung does less extensive speed binning and sells most chips as tcc5 wich arent tested for as high speeds and as tight timings as tccd was binned at. so most chips that are now sold as tcc5 would have been rated and sold as tccd just some weeks ago.

now its ocz and all the memory manufacturers who speed bin the chips they buy wich probably means theres a wider range in clockspeeds, and that the low end sticks dont oc as well as they used to do when they were using tccd.

so afaik thats why ocz and no other memory manufacturer said something about this "different" chip they use. it isnt really a different chip... :)

im not 100% sure though, i asked the ocz guys to make a thread explaining what tcc5 is and i think they will make a thread about it soon :)

[nugzo]

I have 2x512 supertalent memory tcc5 chips. running at 275 2.5-3-3-3 or at 290 3-4-4-7 3.1 volts. i like them better than my twinmos bh-5 i just recieved. I picked up from a local parts distributor for 50 bucks each. had no idea what chips they were till i pried off a heat spreader (and a chip along with it) I'm real happy with em. A friend ordered some mushkin that was suposed to be tccd, instead it was tcc5, he called mushkin to complain and they told him it was just as good as tccd.

[OnDborder]

What does it mean when you speed bin the memory at 466 instead of 500?

[metro.cl]

do you know that there is also tcc4, im not sure but if

tcc5 means 5ns mems
tcc4 means 4ns mems

tcc4 should be way faster or oc better, can anyone comment on this couse im not sure about my theory about ns

[alphaalien]

do you know that there is also tcc4, im not sure but if

tcc5 means 5ns mems
tcc4 means 4ns mems

tcc4 should be way faster or oc better, can anyone comment on this couse im not sure about my theory about ns

That is a bad assumption to make, especially considering you can find the information on samsungs website: www.samsungsemi.com :

X-TCXX

The first X designates the Die revision, samsung has used B, C, E, F as their primary revisions, the second "-TCXX" designates speed grade/timings.

-A0 100MHz/200Mbps@CL2 2-2-2

-B0 133MHz/266Mbps@CL2.5 2.5-3-3

-A2 133MHz/266Mbps@CL2 2-3-3

-AA 133MHz/266Mbps@CL2 2-2-2

-B3 166MHz/333Mbps@CL2.5 2.5-3-3

-C4 200MHz/400Mbps@CL3 3-4-4

-CC 200MHz/400Mbps@CL3 3-3-3

Samsung doesn't even post about "C5, CD" as you can see.

[5-Clicks]

so would TCCD be just a tad better than TCC5 because of the samsung speed bin testing along with the OCZ, G.Skill, Corsair or whoever? more stress tests = better chips type of thing?

[[timko]]

What does it mean when you speed bin the memory at 466 instead of 500?

Speed binning is the process of find if a the memory can perform at a certain speed. Therefore, speed binning of 466 means testing to see if the RAM can perform stable at 233Mhz and binning at 500 means looking for 250Mhz stability. Simple :)

[OnDborder]

So the tccd were held at a higher standard than tcc5.. Or passed the higher standard. To speed bin the tccd and the tcc5's at 233 both will pass.. But speed bin at 250 will both pass? Obviously the tccd will pass
So then in turn, the tccd are a better chip than tcc5's. Does this make sense?

Also, when did these companies start using the tcc5 instead of tccd's?

[5-Clicks]

Also, when did these companies start using the tcc5 instead of tccd's?
when Samsung started making them apparently.

[poiuy223]

Also, when did these companies start using the tcc5 instead of tccd's?
when tccd was announced eol

[metro.cl]

That is a bad assumption to make, especially considering you can find the information on samsungs website: www.samsungsemi.com :

X-TCXX

The first X designates the Die revision, samsung has used B, C, E, F as their primary revisions, the second "-TCXX" designates speed grade/timings.

-A0 100MHz/200Mbps@CL2 2-2-2

-B0 133MHz/266Mbps@CL2.5 2.5-3-3

-A2 133MHz/266Mbps@CL2 2-3-3

-AA 133MHz/266Mbps@CL2 2-2-2

-B3 166MHz/333Mbps@CL2.5 2.5-3-3

-C4 200MHz/400Mbps@CL3 3-4-4

-CC 200MHz/400Mbps@CL3 3-3-3

Samsung doesn't even post about "C5, CD" as you can see.

thanks for the answer, me and my asumption
:slapass:

[Reefa_Madness]

Wow. That article is 14 months old and yet it seems like ancient history. I'm surprised TCC5 has been around that long. I thought it was a replacement rather than a precursor to TCCD. Does this mean Samsung is not firing up the TCC5 machines for new production but is instead just picking through older memory that has been sitting in inventory?

When I first read this review right after it came out I tried to find some ram with these chips and had no luck. I forgot all about it until one of the OCZ reps posted over at OCF that the TCC5 were DDR466 and then I recalled the article and sure enough...TCC5.

The thing is that the chips in this article are "E" revisions and the stuff being used in the newer ram is the "F" revision, which is supposed to clock better. Clearly it is being used in some stuff that is running 300, so when binned, it works good. In my mind it is kinda like using BH5 in PC3500 rated ram. Do we turn our noses up at it because its only rated at 200?

My early impressions are that these chips, when well binned, are probably just as good as TCCD, although at the low end of the speed spectrum they may or may not be as good as TCCD. Only time and more posts from users will answer those questions.

By the way guys, thanks for posting. Keep it going!

[dippyskoodlez]

When I first read this review right after it came out I tried to find some ram with these chips and had no luck. I forgot all about it until one of the OCZ reps posted over at OCF that the TCC5 were DDR466 and then I recalled the article and sure enough...TCC5.

The thing is that the chips in this article are "E" revisions and the stuff being used in the newer ram is the "F" revision, which is supposed to clock better. Clearly it is being used in some stuff that is running 300, so when binned, it works good. In my mind it is kinda like using BH5 in PC3500 rated ram. Do we turn our noses up at it because its only rated at 200?

My early impressions are that these chips, when well binned, are probably just as good as TCCD, although at the low end of the speed spectrum they may or may not be as good as TCCD. Only time and more posts from users will answer those questions.

By the way guys, thanks for posting. Keep it going!

I havent looked under my heatspreaders, but if what Im reading is true, then the mushkin XP 2-2-2 I have may be TCC5. It behaves a lot like my friends TCCD, but 3v helps. Does ddr600 @ 3v pretty easily. 2.5-3-3-7, but will NOT run 2-2-2 no matter what.
:stick:

[Reefa_Madness]

I havent looked under my heatspreaders, but if what Im reading is true, then the mushkin XP 2-2-2 I have may be TCC5. It behaves a lot like my friends TCCD, but 3v helps. Does ddr600 @ 3v pretty easily. 2.5-3-3-7, but will NOT run 2-2-2 no matter what.
:stick:

According to the OCZ reps that had been posting in the "Vent" thread that was closed, they were claiming that TCC5 actually runs tighter timings better than TCCD.

I would think being able to do DDR600 at 2.5-3-3-7 would be of more value than being capable of 2-2-2 at 200, or at least in my book.

[andyOCZ]

According to the OCZ reps that had been posting in the "Vent" thread that was closed, they were claiming that TCC5 actually runs tighter timings better than TCCD.

I would think being able to do DDR600 at 2.5-3-3-7 would be of more value than being capable of 2-2-2 at 200, or at least in my book.

Hi, I don't think we said tighter timings were had with TCC5, but that most of the TCCD we were getting did not do 2-2-2 at 200mhz reliably. The TCC5 was yielding well at 2-2-2, so it was being used. It also "bins" extremely well and our fastest module are using TCC5. We are using the "F" revision.

[uscfan]

Seems like TCC5 does just about as good as TCCD, anyone have anymore results with TCC5? OCZ likes it ;)

[dippyskoodlez]

I would think being able to do DDR600 at 2.5-3-3-7 would be of more value than being capable of 2-2-2 at 200, or at least in my book.

Thats why I still have it, and resisted RMA ;)

[Reefa_Madness]

Hi, I don't think we said tighter timings were had with TCC5, but that most of the TCCD we were getting did not do 2-2-2 at 200mhz reliably. The TCC5 was yielding well at 2-2-2, so it was being used. It also "bins" extremely well and our fastest module are using TCC5. We are using the "F" revision.

My bad...I think that I really meant to say what you just said, but didn't say it exactly how you said it, even if that was what I meant to say.

I tried to make this post confusing...did I succeed??? :)

And thanks for letting us know which revision you use. Is there truth to the "F" being better than the "E" (it appears that way), or is it that the newer stuff is being tested on A64s and just performing better than when the earlier stuff was tested on Intel rigs? It just seems that the newer A64 rigs will get more out of ram than an Intel rig.

[nugzo]

310 HTT (DDR620) 2.5-4-3-8 memtest 5 stable for hours but will not run SuperPi
275 HTT (DDR550) 2.5-3-3-6 Completely stable (Best Overall Benchies)
290 HTT (DDR580) 2.5-4-3-8 Completely stable
200 HTT (DDR400) 2-2-2-2 Completely stable

3.0-3.1v

Again this is Stuff called SuperTalent. It is rated PC3700 3-4-4-8. i had never heard of them untill i picked these up. Super Talent Link (http://www.supertalentmemory.com/desktop.php)

[g0dM@n]

Anyone know about the PDP Patriot XBL... is that definitely TCCD, or is it maybe TCC5? I never ripped off the heatsinks, and I'm actually now thinking about doing it.

I was told that the TCC5 does tighter timings than TCCD at lower fsb, but needs looser timings at higher fsb than TCCD. I guess it's kinda like how BH UTT does tighter timings at lower voltages than CH UTT, but at higher voltages CH UTT wins... just a rumor. I have had better yields with CH UTT over BH UTT, but in comparison to sammy chips I don't know if my patriot is TCCD or TCC5, so I cannot compare them directly. It seems to do tighter timings at high fsb when I up the voltage up to 2.9-3.0v. Gskill and other TCCD don't seem to really do much better with voltages about around 2.7-2.8v. Maybe all this time that people have been doing well on the so-called TCCD at 3.0-3.1v actually have TCC5???

[LikwidKool]

Here's what I heard. I took some parts out as they had to do with certain modules, but here is the gist about tcc5 and tccd.:


There's absolutely zero difference between tested TCC5 and tested TCCD.

Anyway, overall it doesn't matter what chips they are, as they are exactly the same chips. Samsung just relabels them and doesn't test them as high on the chip level tester. It's the same process, same die, same package as far as we know.

Hope this helps to clear some things up.

[Reefa_Madness]

After spending the last couple of days running down all the different threads and discussions on this topic...that is my understanding as well. Same thing, just not TESTED by Samsung to DDR500 anymore.

All indications are though, that you still want the "F" revision over the "E".

[saaya]

i stickied this thread to make sure everybody gets what this tcc5 discussion is about :D

[andyOCZ]

After spending the last couple of days running down all the different threads and discussions on this topic...that is my understanding as well. Same thing, just not TESTED by Samsung to DDR500 anymore.

All indications are though, that you still want the "F" revision over the "E".

OCZ, for example uses only "F", which is "the good stuff". :)

[[timko]]

Good idea Saaya. With TCC5 becoming more prolific it's gonna become a commonly asked question :)

[Reefa_Madness]

OCZ, for example uses only "F", which is "the good stuff". :)

Good to get that confirmation from you, AndyOCZ. Several guys have asked specifically what you used. They have also asked when you started using the TCC5 chips in the PC3200, for example.

My question is more general...do you know when Samsung switched/started production of the "F" revision of the TCC5 chip? If TCCD and TCC5 are the same then it should have been around late April or early May of 2004 (aren't the first TCCD 0417 chips).

I ask because maybe it would also be a good idea to track the various batches of TCC5 to see if patterns develop as to their overclocking potentials, similar to what has been done with other chips.

If you guys post any results with mfg. dates, I can try to organize them in the first post. That is the problem with starting a thread...you've got to be the record keeper. :)

[LikwidKool]

Good to get that confirmation from you, AndyOCZ. Several guys have asked specifically what you used. They have also asked when you started using the TCC5 chips in the PC3200, for example.



From what I have been told OCZ started using TCC5 back in March. You can tell what your sticks have by the rev number in the lower right hand corner of the sticker. v1.0 is TCCD and v1.1 is TCC5.

[Reefa_Madness]

Do the TCCD sticks even have the version # on them. My sets don't.

I was under the impression that only the TCC5 were labeled with a version number.

[[timko]]

My TCCD sticks do not have the extra rev number on them.

It is ONLY the TCC5 sticks that will say something like --> "Rev.2 rev1.1" on the white sticker.

[Reefa_Madness]

That is my understanding, too. Only the TCC5 have a version number. No version # = TCCD.

[hitmanx2]

so, is TCCD generally a little bit better than TCC5 :confused: like BH5 vs BH6?

[dippyskoodlez]

so, is TCCD generally a little bit better than TCC5 :confused: like BH5 vs BH6?

I think its more like a TCCD Vs TCC5 as BH-5/CH-5 is UTT ?

The stuff i have which I believe is tcc5, does ddr600, so you definatly cant say its always worse.

[g0dM@n]

You can do ddr600 at what timings? And is that 2x512 or 2x256?

[dippyskoodlez]

You can do ddr600 at what timings? And is that 2x512 or 2x256?

2.5-3-3-7 1x 512mb. I dont have another stick yet. BP pcb. its strange.. I bought black level II rev 2 from newegg, got this.. it says level II 2-2-2, green PCB, and clocks a hellova lot better than my friends Black level II. Im guessing its XP 2-2-2, and if it is, its some nice stuff.

[Frank the Tank]

I have some OCZ PC4800EL with a rev 1.1 (TCC5), which clocks way better than my Corsair 3200XL (TCCD) pro series. I can reach 310+Mhz 1T, with 2.5, 5, 4, 7. My Winchester mem controller is junk (2500 max.) and I can still get over 300+ @ 1T, the only value that gives me trouble is that TRCD fails SuperPI with anything lower than 5, but passes Memtest Dos fine?? I'm trying to work that bug out, but running at 5 doesn't seem to affect the bandwidth by much at all. A 3700+ SD core is on order, so I'll see if that works even better. These are the ICs used in the new PC5000 DFI Special Edition modules also, just binned a little different.

Here is a quote from Andy @ OCZ:

It seems you are having great results despite the odd TRCD of 5. The CPU you have most likely does not work at the usual timings above 270mhz. It is fairly common for Winchester cores to cause problems. Please see the following timings as they may help with a new starting point. You can tighten up from these after you establish your TRCD setting.

http://www.bleedinedge.com/forum/showpost.php?p=89870&postcount=11

There is other interesting info in that thread also.

The Rev 1.1 would designate TCC5, and as you observed it’s akin to a red Corvette vs. a Yellow one; same thing under the paint. ;) The fastest memory I have here in my lab is TCC5, hands down.

Please let me know your results.

Thanks,

Andy

Online Media Coordinator

OCZ Technology, USA

[bugeyes]

Well it seems Samsung is still making TCCDs, it's just not supplying it to all companys!

I wan'ted to buy some for when I uprade to A64. so I decided to track down what was available locally (Australia) and the only ones I could find were G.skill 4400LE @ $478 AUD and Corsair 4400C25 @ $328 AUD.... I was a little sceptical about wether the Corsair's were infact TCCDs as there price seemed very low compared to the G.Skill. I went over to The House of Help and asked Corsair's representitive and here's his reply.

As I said before all XMS4400C25 and XMS3200XL modules use Samsung TCCD IC's. Corsair uses TCC5 RAMs only on other speed grades, such as the XMS3200C2 and XMS4000. Note that Samsung did not stop producing TCCD IC's, they just stopped shipping them to non direct suppliers. Or in other words we still get a large quantity every month and will do so for some time to come.

LINK ---> http://www.houseofhelp.com/forums/showpost.php?p=193555&postcount=20

[Eldonko]

Just found out that 4200EL is now TCC5. Will post my results in a bit.

http://members.shaw.ca/Eldonko/v11.jpg

EDIT: 284Mhz @2.5-3-3-5 and 2.8v, 289Mhz @2.5-4-3-5 and 3.0v are the limits on this RAM

[Reefa_Madness]

Well it seems Samsung is still making TCCDs, it's just not supplying it to all companys!

I wan'ted to buy some for when I uprade to A64. so I decided to track down what was available locally (Australia) and the only ones I could find were G.skill 4400LE @ $478 AUD and Corsair 4400C25 @ $328 AUD.... I was a little sceptical about wether the Corsair's were infact TCCDs as there price seemed very low compared to the G.Skill. I went over to The House of Help and asked Corsair's representitive and here's his reply.



LINK ---> http://www.houseofhelp.com/forums/showpost.php?p=193555&postcount=20


That information would be consistent with a post made by Duonger, the Mushkin rep, over at OCF. He stated that TCCD was still available, but you had to have contacts and buy a boatload.

I guess we now know who's got the contacts....

[LBJGH]

... see sig... 283mhz 1:1 2.5,4,7,3

Seems like TCC5 does just about as good as TCCD, anyone have anymore results with TCC5? OCZ likes it ;)

[Reefa_Madness]

I've just received two set of OCZ PC3200 v1.1 so I will be able to at least try two different sets on the same rig (MSI Neo2/3200 winney combo). That will help in determining consistency in the OCZ product. I'll play with them over the weekend and let everyone know how it went.

[bugeyes]

I would love to see a comparison of the OCZ 3200R2 v1.1 against Corsair 3200XL... It should show hows the boss TCCD or TCC5 :stick:

[Rick_Hunter]

hi guys
my corsair c2pt 4.2 came with tcc5 and I'm very happy....
this modules are very close to popular and expensive tccd in my opinion
the max fsb I can reach is 325mhz 3-4-3-7 but 2T :( ...the best configuration was 300mhz 3-4-3-7 1T :) (ras to cas work with 3 but show some inestability issues after heavy load)

http://img126.echo.cx/img126/9387/249oh.th.jpg (http://img126.echo.cx/my.php?image=249oh.jpg)
http://img296.echo.cx/img296/8937/bw8gg.th.jpg (http://img296.echo.cx/my.php?image=bw8gg.jpg)
:toast:

[Reefa_Madness]

Self quote:

"I've just received two set of OCZ PC3200 v1.1 so I will be able to at least try two different sets on the same rig (MSI Neo2/3200 winney combo). That will help in determining consistency in the OCZ product. I'll play with them over the weekend and let everyone know how it went."

Not going to be able to do this two set testing as I've gone and sold a set NIB.

The set that I'm testing is probably in line with my other TCCD, which is to say 260 at 2.5-3-3-7 (1T) (2.7v), 282 at 2.5-4-3-8 (1T) (2.85v) and topping out at 292 with 3-4-4-8 (2T). I have removed the Booster from my board so these sticks have not been tested with higher voltges.

I'm still using an early BIOS and my board doesn't play as well with TCCD as others. The point that I'm trying to make is that my results were similar, whether it was TCCD or TCC5 that I was testing.

[high5]

since i see there has been much talking about tcc5 issue on various forums, i'll post the information i have:

tccd & tcc5 (ones used in newer production ocz platinum el rev2) are exactly the same ic's, the only difference between them is what they have been tested for. 'tccd' and 'tcc5' stuff is not ic type, its their speed rating, what they've been binned for. tccd means that it has been tested & qualified by samsung as ddr500 part, tcc5 means its ddr466. they're both k4h560838f ic, that 'k4h560838f' thing is the ic type, samsung's premium ic. samsung tcc5 reviewed at anadtech that reefa posted has nothing to do with tcc5 used on ocz platinum el rev2, for example, since ones reviewed on anand are k4h560838e ic, notice the 'e' instead of 'f' at the end. that is the different ic and can't be compared to f-tcc5 we're talking about here. some people call that rev.f or rev.e. calling that thing revision is wrong b/c most people are getting impression that its revision f of the tccd ic (for example), while in fact its the other way around: its ic type f, speed grade tccd.

now the speculation part: why did samsung stopped (or just reduced) testing their ic's for ddr500 parts? well, this is my guess; winbond sells their ic's completely untested, and that means they're very cheap b/c testing ic's is expensive. and the more demanding testing is, the more it costs. therefore the same ic tested for ddr466 operation instead of ddr500 would cost less and would be more competitive on the market b/c price difference between utt and tccd is huge.

[amrgb]

...and the more demanding testing is, the more it costs.

I can´t see how can Samsung save in direct costs of testing when testing for DDR466 instead for DDR500.

However they indeed save costs because less chips will fail the test procedure (those that can run between DDR466 and DDR500 but not above DDR500).

Additionally, since they are still producing and selling TCCD, unlike many have claimed, Samsung will relabel those chips that failed the TCCD testing.

Concluding, a batch of TCC5 chips will have inferior quality compared to a batch of TCCD.

Of course that one can find TCC5 chips performing the same or even better than some TCCD chips. But with TCC5 it is possible to get a chip clocking between DDR466 and DDR500, whereas with TCCD this would not happen (provided that the testing procedure is accurate).

In the past, when buying TCCD based memory rated at 2-2-2-5 DDR400 we had two thing for granted: DDR400 and DDR500 performance. Nowadays we only have DDR400 and DDR466 for granted. And the probability of not achieving DDR500 performance is not negligible (in the same conditions as Samsung has tested the chips, since there is little evidence of TCC5 not achieving DDR500 in real world use, although there are some cases out there).

So, it is not correct to state that TCC5 is the same as TCCD. On average TCC5 is worse. This justifies the fact that it is cheaper than TCCD.

[high5]

I can´t see how can Samsung save in direct costs of testing when testing for DDR466 instead for DDR500.
as i stated in previous post, the more testing is demanding, the more it costs. testing ic's is not a simple thing and is certainly not being done in one step. since i believe many reps hanging around here, there would be no problems confirming this i hope.

Additionally, since they are still producing and selling TCCD, unlike many have claimed, Samsung will relabel those chips that failed the TCCD testing.
this is exactly why i posted my initial post, tccd is not being produced, k4h560838f is what is produced and tccd/tcc5 is what its being tested for. there is no such thing as relabeling tccd to tcc5, simply b/c after the production, they're all just k4h560838f ic's before testing part. and for testing, i believe the ic's are being tested at the lower speed first, and then if they pass the test they go to higher grade testing and so on and so on... however, i'm not 100% certain about that but i think its the way it is.

So, it is not correct to state that TCC5 is the same as TCCD. On average TCC5 is worse. This justifies the fact that it is cheaper than TCCD.
i stated that physically tccd and tcc5 are completely identical b/c they are the same ic's of the same manufacturing process. please, re-read my post with some understanding.

[BioPC]

physically tccd and tcc5 are completely identical b/c they are the same ic's of the same manufacturing process.

This is what i got testing my OCZ EL PC3200 Rev.2 TCC5: 320 2,5-4-3-7 1T @2.9v :D (x8 multiplier)
Sorry for the bad foto quality.

[high5]

nice stuff. i thought i saw that (or something similar) in another thread. was it you? i myself consider getting a gig of that stuff, i currently have 512mb corsair c2 v4.1 (tccd based), but it craps out at 270ish due to not having brainpower pcb.

[amrgb]

this is exactly why i posted my initial post, tccd is not being produced, k4h560838f is what is produced and tccd/tcc5 is what its being tested for. there is no such thing as relabeling tccd to tcc5, simply b/c after the production, they're all just k4h560838f ic's before testing part. and for testing, i believe the ic's are being tested at the lower speed first, and then if they pass the test they go to higher grade testing and so on and so on... however, i'm not 100% certain about that but i think its the way it is.

In the essence you're saying the same I said before. What's the difference between relabelling TCCD to TCC5 or labeling k4h560838f ic's as either TCCD or TCC5. I can't see much of a difference. All chips that do not pass the TCCD tests are labeled as TCC5. To meet the TCC5 demand some chips that could be sold as TCCD will actually be sold as TCC5. It's the same sort of speed binning that cpu manufacturers do. A A64 3000 cpu may overclock to more than a 3500. But on average 3500 ones are better clockers.


i stated that physically tccd and tcc5 are completely identical b/c they are the same ic's of the same manufacturing process. please, re-read my post with some understanding.

I've understood your post clearly. But still one could not say that TCC5 is the same as TCCD (and I do not mean that's your opinion). To think that TCC5=TCCD is the same to think that an A64 3000=A64 3500. The core is the same, but the 3500 is better (on average), just like TCCD is better than TCC5 (on average).

[[XC] Jaco]

the way I see it :
"Samsung makes k4h560838f ic's. these are basicly DDR 500 chips
Before they were binned and sold as TCCD
Now , they're not speed binning anymore to cut costs. they have a good IC so they are selling it untested DDR466. TCC5

A good move from Samsung , now the market is flooded with dirtcheap winbond UTT ."

don't bash me for anything written between the quotes.
this is all speculation on my part , so could be complete BS :D

[high5]

amrgb: okay, glad we sorted that out. :up:

the intention of my post was to make clear that tccd & tcc5 are both products of the same manufacturing process just rated differently since i see many ppl are not aware of that from what i've read over the past days/weeks. and some are overreacting b/c some manufacturers switched to tcc5. hope you agree with me here.

To meet the TCC5 demand some chips that could be sold as TCCD will actually be sold as TCC5.
this was my point also, its just that i didn't put it into the words so simple as you did. :) only i didn't mean 'some' but rather 'a lot' of them, i stated why in my initial post (cutting costs).

[amrgb]

I have to thank you for the important information about the genesis of TCC5/TCCD.

But overreaction is a normal thing after the uncover of hidden facts, especially when there is little information. TCC5 may not be so bad compared to TCCD as many (including me) thought at the beginning, but you must agree that on average its worse than TCCD. If the prices go down accordingly, fine. We are even.

The problem is that retailers took profit of the lack of information about this fact to sell TCC5 sticks at TCCD prices. Memory manufacturer like OCZ claim that since the introduction of TCC5 they have lowered their prices, but that they cannot control the prices chaged by the retailers. I have no reasons to doubt OCZ words, and its true that they cannot control the prices charged by their retailers. Unless they had disclosed this information. Then, people would be awarre of the greedy retailers and stop buying overpriced sticks until they charged fair prices.

That's how I see this question know. And I still think that memory manufacturers should have protected their customers interests from opportunist retailers.

PS: at the present, OCZ Plat rev2 v1.1 costs 235 EUR and GSkill 4400 LE 265 EUR here in Portugal. Two weeks ago it was 250-265. Aren't the Plat rev2's overpriced? I think so, although their price is slowly decreasing.

[high5]

But overreaction is a normal thing after the uncover of hidden facts, especially when there is little information. TCC5 may not be so bad compared to TCCD as many (including me) thought at the beginning, but you must agree that on average its worse than TCCD. If the prices go down accordingly, fine. We are even.
agreed, however, i feel some ppl jumped right at ocz without really checking the real difference, and that certainly isn't fair. as for tcc5 being worst than tccd on average, well, theoretically yes, but you must have in mind that vast majority (if not all) of tccd did better than its rated speed (ddr500). and only time will tell for tcc5, but i expect the trend will continue and the difference won't be so obvious. i may be wrong... we'll see.

The problem is that retailers took profit of the lack of information about this fact to sell TCC5 sticks at TCCD prices. Memory manufacturer like OCZ claim that since the introduction of TCC5 they have lowered their prices, but that they cannot control the prices chaged by the retailers. I have no reasons to doubt OCZ words, and its true that they cannot control the prices charged by their retailers. Unless they had disclosed this information. Then, people would be awarre of the greedy retailers and stop buying overpriced sticks until they charged fair prices

That's how I see this question know. And I still think that memory manufacturers should have protected their customers interests from opportunist retailers.
i agree with everything you stated above 100%. especially the retailers part. but unfortunately, there's not much we can do about it.

at the present, OCZ Plat rev2 v1.1 costs 235 EUR and GSkill 4400 LE 265 EUR here in Portugal. Two weeks ago it was 250-265. Aren't the Plat rev2's overpriced? I think so, although their price is slowly decreasing.
well, here in croatia price dropped 22% on ocz last week. it was €295, and now it stands at €230. we don't have any gskill available over here. :(

[Eldonko]

Did a bunch of tweaking since my last post, updated results for 4200EL TCC5:

http://members.shaw.ca/Eldonko/307-3.JPG

:woot:

[high5]

eldonko, thank you for your input. however, could you post your results again in ocz tcc5 overclocking results thread (http://www.xtremesystems.org/forums/showthread.php?t=64501) so we can have all the users experiences in one place.

anyone else running tcc5, please post your results in thread mentioned above so we could all get better impression of tcc5 in global.

[Eldonko]

Ok, done.

[Bloody_Sorcerer]

I'm really impressed by some of your guys' TCC5 clocks/results. My current stuff will only do around 245 2.5-3-3-10 at 2.9v. maybe just a voltage issue. *considers getting a DDR booster*
this stuff can do tighter timings too; 2.0-3-2-10 at 2.9v, but only up to around 210 mhz, after which it produces all sorts of interesting errors.
My TCC5 is rev. F 0431 (i think) on XMS3202v4.2s (2x512)

[amrgb]

That's the problem with TCC5. Some good results mixed with some very bad results. This behavior was not so common with TCCD.

You can try 3v or even 3.1v, but with these volts you definitely need active cooling and most probably you will be working beyond your module's warranty.

[Bloody_Sorcerer]

i'm already out of warranty and got active cooling, so thats no problem :)
just did some more tests; it looks like something doesn't like a 1T command rate.
2.5-3-3-10-1T 250mhz: errored during windows boot
2.5-3-3-10-2T: stable to around 260mhz
3-3-3-10-2T: stable to around 275mhz
Perhaps it's the mobo (K8N Neo Platinum) that doesn't like the 1T at high clocks.
Didn't really try anything above 275mhz because chipset voltage was too low (based on previous experimentations).
Even so, 3-3-3-10-2T @ DDR550 netted me an extra 1000mb/s in everest memread.

[high5]

i have similar experience with my tccd on s754, 1t goes only upto 235mhz, when i switch to 2t i'm hitting 275-280 with no problem. friend of mine had similar issues with corsair xl v1.1. so it has nothing to do with your ram being tcc5.

try testing your ram in another system. my ram does easily 280 w/ 1t in nf4 mobo. from what i've seen, tccd/5 without brainpower has problems reaching high clocks using 1t in s754 systems.

[LikwidKool]

the OCZ 4200EL TCC5 seems to be good stuff. So far I am at 290 3-3-3-10-1t 2.8v

[high5]

what about 2.5-3-3?

[g0dM@n]

I miss my PDP Patriot XBL that I sold... sold it for 160, too... dunno if that was such a bright idea... been stuck with utt twinmos since.

How is the OCZ 4800 stuff doing?

[hovo73]

2x512 PC3200, OCZ Platinum Rev 2 v.1.1 (TCC5)
DFI NF4 Ultra-D, 510-2 BIOS
Winchester 3000+
ATI X800XL
Enermax Noisetaker 485
1.325V+123%, 1.3V, 1.7v, 2.9V
CPU: 9x280, RAM: 200
2.5-4-3-7-1T-8-16-3-3-2-3-4708-A-E
A-0-L6-L2-8.0-F-A-256-D-16x-7x-D

These sticks were refusing to work at 280 no matter what timings, and voltages until I relaxed Max Asyncrh Latency to 8.0ns. As you can see, now they are rock stable. They will go up to about 290 at these settings, but my CPU won't go over much over 2520MHz (9x280).

http://www.xtremesystems.org/forums/attachment.php?attachmentid=33852

This results are equal or better than the ones I got from PDP Patriot XBLK TCCD which I bought about 6 month ago. I still own them and since I learned a thing or two about DFI boards and ram timings, I am going to try those Patriots again soon.

[dqniel]

Is it possible to get TCCD new anywhere anymore? Or has it all gone to TCC5? I want to buy whatever is the cheapest available 2x512 TCCD Brainpower kit.

For example, is this still tccd? http://www.newegg.com/Product/Product.asp?Item=N82E16820220033

[dippyskoodlez]

Is it possible to get TCCD new anywhere anymore? Or has it all gone to TCC5? I want to buy whatever is the cheapest available 2x512 TCCD Brainpower kit.

For example, is this still tccd? http://www.newegg.com/Product/Product.asp?Item=N82E16820220033

Corsair looks to be the only place with a fresh supply of tccd, seeing as thats who samsung is still making it for.

[dqniel]

Corsair looks to be the only place with a fresh supply of tccd, seeing as thats who samsung is still making it for.

Thanks for the info. Looks like the only "fresh" ram being produced with TCCD is 3200xl and 4400C25PT now :(. Guess I'll look to buy some used or something cause 185 for 2x512 is a bit steep. Either that or I'll go with some UTT chips.

[Anarki]

Is anyone using TCCD/TCC5 with a P4/Celeron/P4-M/Celeron-M if so what timings are you getting and on what board?

Thanks :toast:

[hansbroken]

What about the G.Skill's at New Egg? These are advertised as TCCD, but is that not the case anymore? For example, I noticed that New Egg no longer lists the Patriot XBLK's as being TCCD. And since the price on the Patriot have gone down to the mid $160's it would make sense that they switched to TCC5. Whereas the G.Skill's seem to be holding steady in price -- more or less. Can anyone confirm this about G.Skill?

Either way, I hope that since G.Skill is still advertised as TCCD -- and the price seems to be in-line with that assumption -- that they are in fact TCCD. If so, which one to get (if I’m not mistaken, LA is their low end designation??): LD, FR, LC (getting pricey going further)...

Also, what about Muskin? Aren't they like Corsair in that they are getting the TCCD Chips?

Lastly, since -- for example -- companies like OCZ seem to be putting some real effort into producing quality memory using these newer rev 'f' TCC5 ic's in their 3200 rev 2, then is this TCCD / TCC5 issue more dependent on how the manufacturer of the RAM implements them (such as testing etc..) rather than the ic's alone? In other words: if we are now 'stuck' (so to speak) with TCC5, then wouldn't it make sense that a company like OCZ will make it more probably to get a set that OC's like the TCCD brainpower RAM of olde?

BTW: Out of curiosity, does OCZ use Brainpower PCB?

P.S. 'Anarki' -- that is one lucky kitten if you ask me, lol :D

Thanks y'all

[dippyskoodlez]

What about the G.Skill's at New Egg? These are advertised as TCCD, but is that not the case anymore? For example, I noticed that New Egg no longer lists the Patriot XBLK's as being TCCD. And since the price on the Patriot have gone down to the mid $160's it would make sense that they switched to TCC5. Whereas the G.Skill's seem to be holding steady in price -- more or less. Can anyone confirm this about G.Skill?

if its marked tccd, its probably tccd. there are still available sources of tccD, but its more expenbsive than tcc5. corsair has a cheap, *fresh* supply of tccd, samsung is making it for --Them. doesnt mean other companys cant get it ;) corsair is probably the one keeping it in production. i dont know the details, but im sure other companys can get it, just ahve to pay a bit more.

and yea OCZ uses the bp pcb most of the time last i checked.

[andyOCZ]

Corsair looks to be the only place with a fresh supply of tccd, seeing as thats who samsung is still making it for.

OCZ has plenty of "fresh" TCCD, back by popular demand. We just announced our PC4800 Platinum Elite TCCD.

http://www.ocztechnology.com/aboutocz/press/2005/139

Check it out. It should be available this week.

[4rory]

Theres a typo in that press release, Where it says all ddr2 is hand tested, this ain't ddr2.

[hansbroken]

if its marked tccd, its probably tccd. there are still available sources of tccD, but its more expenbsive than tcc5. corsair has a cheap, *fresh* supply of tccd, samsung is making it for --Them. doesnt mean other companys cant get it ;) corsair is probably the one keeping it in production. i dont know the details, but im sure other companys can get it, just ahve to pay a bit more.

and yea OCZ uses the bp pcb most of the time last i checked.

Thanks for the info :toast:

[ray99mond]

http://lp.pcmoddingmy.com/albums/userpics/10003/gskillfc23.JPG
http://valid.x86-secret.com/show_oc?id=28155
These is my Gskill FC modules overclocking result.
this is the highest stable setting (passed memtest86, super PI 32M, gaming and etc) for this Gskill FC PC4400

BIOS 414-3
2.8Vdimm
Samsung TCC5, 0522

[Super strokey]

Well i just got my abit an8 sli and my ocz rev 2 3200 (which i assume is tcc5) and its running 256 mhz at 2.5,3,3,7 with 2.8 vdimm. My only real question is should i feed it more volts for more clocks or is tcc5 the same as tccd and more volts actually reduces stability?

[SubX]

Corsair PC3200C2 Rev4.2 (TCC5)

1.5-2-2-5 T1 @ 450Mhz
2-2-2-5 T1 @ 466Mhz
2.5-3-3-5 T1 @ 610Mhz
2.5-4-3-5 T1 @ 680Mhz
2.5-4-4-5 T1 @ 690Mhz

[BioPC]

Corsair PC3200C2 Rev4.2 (TCC5)

1.5-2-2-5 T1 @ 450Mhz
2-2-2-5 T1 @ 466Mhz
2.5-3-3-5 T1 @ 610Mhz
2.5-4-3-5 T1 @ 680Mhz
2.5-4-4-5 T1 @ 690Mhz

Could you show some SuperPi 16M-32M screens? :fact:

[rick_fx]

Any pics :rolleyes:

[Bloody_Sorcerer]

Corsair PC3200C2 Rev4.2 (TCC5)

1.5-2-2-5 T1 @ 450Mhz
2-2-2-5 T1 @ 466Mhz
2.5-3-3-5 T1 @ 610Mhz
2.5-4-3-5 T1 @ 680Mhz
2.5-4-4-5 T1 @ 690Mhz
holy crap, nice 4.2s you've got... Mine'll only do up to 2.5-3-3-10 T2 @ 450. I blame socket 754.

[Rick_Hunter]

Corsair PC3200C2 Rev4.2 (TCC5)

1.5-2-2-5 T1 @ 450Mhz
2-2-2-5 T1 @ 466Mhz
2.5-3-3-5 T1 @ 610Mhz
2.5-4-3-5 T1 @ 680Mhz
2.5-4-4-5 T1 @ 690Mhz
cas 1.5 on tcc5? :confused:

pics please :rolleyes:

[largon]

According to AMD's datasheets A64 doesn't actually use CL1.5. Even though some boards have that option it isn't *really* CL1.5.
If the RAM can do CL2, it should be able to do this pseudo-CL1.5.

[Rick_Hunter]

According to AMD's datasheets A64 doesn't actually use CL1.5. Even though some boards have that option it isn't *really* CL1.5.
If the RAM can do CL2, it should be able to do this pseudo-CL1.5.
so cas 2 on tcc5? :confused:

[JuanFlaiter]

According to AMD's datasheets A64 doesn't actually use CL1.5. Even though some boards have that option it isn't *really* CL1.5.
If the RAM can do CL2, it should be able to do this pseudo-CL1.5.

Although some CPUs prefer CL1.5 than CL2.

I can stabilize my BH5 at CL 1.5 280mhz and CL2 278mhz.

Greets!

[saaya]

are you sure?
i thougt the cas1.5 setting is actually beeing ignored by the system...

[SubX]

450Mhz 1.5-2-2-5 T1 (http://valid.x86-secret.com/show_oc?id=30074)

690Mhz 2.5-4-3-0 T1 (Sandra BW) (http://img370.imageshack.us/img370/1674/san77515hk.png)

:toast:

[kiwi]

SubX:

What about SP32M or prime? Sandra bandwidth is really low for 690Mhz

[wakdog]

also that second shot of sandra at DDR690 is actually at 2T not 1T like you said...

[Vapor]

Sandra always says that I think....the BW is WAYYYYY too low for those clocks though. Me thinks the 4x multi was used.

[SubX]

also that second shot of sandra at DDR690 is actually at 2T not 1T like you said...

Sandra allways shows 2cmd , check it your self ;)

The BW is low because the CPU multiplier was x7 (crap Venice 0524)

[S!1v3rB@cK_Dk]

Just got the new Mushkin extreme PC3200 Level 2, I think the part number is 991357.
http://www.mushkin.com/doc/products/memory_detail.asp?ID=15

I think they carry the TCC5 chips, but im too afraid to take of the heatspreaders to have a peek.
Atm they are doing 270MHz 2.5-3-3-7 with only 2.5v!! And this speed is where my winnies memory controller maxes out, so I cant get em higher. But overall this is real good mem, and they perform like memory that are way more expensive.
I bought em even though a bunch of people recommended me not to, but now I will recommend them to anyone who are looking for some good TCCD like mem! :toast:

[Reefa_Madness]

According to a thread started by a Mushkin rep (Zebbo) over at OCF, Mushkin has returned to using TCCD, instead of TCC5, in their XP3200 and XP4400.

http://www.ocforums.com/showthread.php?t=419290

Just thought that was worth sharing.

I wonder if perhaps they've done the same with the PC3200 Level 2 V2?

[RyderOCZ]

Sandra allways shows 2cmd , check it your self ;)

The BW is low because the CPU multiplier was x7 (crap Venice 0524)Its also low because of Single Channel :)

[qwerty!]

yep.

[Tamille]

Just found out that 4200EL is now TCC5. Will post my results in a bit.

h***://members.shaw.ca/Eldonko/v11.jpg

EDIT: 284Mhz @2.5-3-3-5 and 2.8v, 289Mhz @2.5-4-3-5 and 3.0v are the limits on this RAMI have the same ram as you. If i go under 280mhz the ram :banana::banana::banana::banana:s its self. i only just bought it 5 days ago. so far running at 300mhz 1T 2.5-4-3-7 2.8v 1:1 LDT 3x

[smolk]

I think TCC5 should be a bit better:

http://files.upl.silentwhisper.net/upload2/TCCD-TCC5.PNG

As you can see, TCCD is rated 4ns (250MHz) and TCC5 is rated 3,75ns (266,67MHz) at the same timings. But of course, it can be some Samsung gibberish, even though I don't think it is.

Complete .pdf: http://files.upl.silentwhisper.net/upload2/TCCD-TCC5.PNG

[Vapor]

Ugh, we don't more debate on this....and where is that image from? What's with this CF stuff that can do very well at 1.9V? Pffff....we KNOW this: TCC5 is just a lower binning (at Samsung) of the same die as TCCD.

[ozzimark]

Its also low because of Single Channel :)
if you did the math, ddr690's theoretical bandwidth is 5520mb/s, way under the results he posted ;)

or i could be stupid and missing something key :p:

[smolk]

Ugh, we don't more debate on this....and where is that image from? What's with this CF stuff that can do very well at 1.9V? Pffff....we KNOW this: TCC5 is just a lower binning (at Samsung) of the same die as TCCD.

Sorry, didn't reccon that. :slapass: The source is below in my earlier post. I think that the CF, DH and a couple of others are DDR2 ;)

[g0dM@n]

I remember when I was told that all OCZ Plat Rev2 3200 that said "v1.1" on the PCB was TCC5, and all PCBs that said "v1.0" were TCCD.

Well, I bought my OCZ Plat Rev2 3200 2x512 from ebay for $130 (-$15 ebay coupon at the time = $115). I got the set in and asked around again about how to check if it's TCC5 or TCCD, and was told the same thing. My PCB read 1.1... I said screw it, and I took one of the clips off that weren't by the sticker, and bent the heatspreader JUST enough to take my peak. Lo and behold it's "TCCD"!

I did the same with my OCZ EL Plat 3700 2x512. Got this sucker off ebay for $104 shipped (coupons were expired by this time). I stretched over the heatspreader and baBOOM TCCD! And I was told it was TCC5.

There ya have it... I got myself two TCCD sticks when they were said to be TCC5. They both do 280fsb 2.5-4-3-7 for sure... 2.5-3-3-7 at 280fsb is still tough to do. I did get 300fsb on the 3200 kit memtest stable at 2.5-4-3-7, but I dunno yet about windows.

What's the farthest someone has taken TCC5?

[high5]

v1.0/v1.1 marking isn't on PCB, its on sticker on HS.

http://2crowded.com/vedran/ocz_rev.JPG

[g0dM@n]

v1.0/v1.1 marking isn't on PCB, its on sticker on HS.

http://2crowded.com/vedran/ocz_rev.JPG
But there is also a v1.1 marking on my PCB. I thought I looked at the sticker... will have to check again... thx. :D

[high5]

marking on PCB would be PCB revision. :)