FREEZING - OCZ memory tricks

[DSE]

Hmm, I don't own OCZ sticks, nor have any experience with them, so I can't comment about personal experience with them. The issue of ethics and practices seems fairly straight forward. Binning chips have been a practice among companies for a while; Gainward, with their Golden Sample, sold cards with handpicked chips to overclock better; Nvidia doesn't, and I'd assume ATI doesn't produce their own GPU, but we have no problems seeing their names printed on them.

As long as OCZ had Samsung's permission to blast the logo, it's not fraud or illegal.

I suppose the combination of business and marketing practice could be considered shady, I know of a local shop here that sells 2000+ as 2Ghz system. Yes, they put both numbers side by side on their TV commercials. That puts a bad taste in my mouth, either the people there want to lie, or don't know the difference. Either way, not a shop I want to deal with. Same with OCZ, it seems. I don't think you can call their practice illegal or fraudulent, but the fact that they want to go to such length as burning off Samsung's logo to -hide- what should be an open fact, does not make them a company I want to deal with.

 

[jaeger66]

Originally posted by: DSE
I don't think you can call their practice illegal or fraudulent, but the fact that they want to go to such length as burning off Samsung's logo to -hide- what should be an open fact, does not make them a company I want to deal with.

I still want to know what they gain by remarking the chips. The blast process may or may not be valid, but why the new rating? I can't think of any reason other than deception. It isn't like CPU speed binning where the chips are sorted and THEN marked.

 

[wicktron]

Originally posted by: BingBongWongFooey
Haha, wixt0r, do you realize the kind of company you're defending? It's been known for YEARS that ocz is a total con job. Google around, or search archives here, I'm sure you can dig up the relevant stuff.

Originally posted by: Budman
stevejst and wixt0r stop arguing & look at this Picture

I am not arguing whatsoever. I have been contributing to this thread sharing my opinion without throwing any flames. It was I who was told I was stupid and drunk. It was I who was told I needed to have my hand held to add 2+2.

Look, I am completely aware of OCZ's shady past. I was burnt by them before. It was only recently with the release of their Gold series that I bought their RAM again. It was such a great experience that I now believe in their company once again.

Originally posted by: DSE
but the fact that they want to go to such length as burning off Samsung's logo to -hide- what should be an open fact

They do make this an open fact, the OCZ representative at the ABXZone and Xtremesystems forums has made this fact known for a long time.

This has been blown way out of proportion. The people at ABXZone and Xtremesystems do not see this as a big deal, and actually commend OCZ for improving product.

 

[Acanthus]

because a 4.3ns chip can run at 4ns, and a company tests it to make sure it works, and rebadges it, its fraud?

 

[bigtoe33]

You guys must have seen the reviews that shows just how good the OCZ gold is. Also OCZ is NOT the company it once was, its MUCH better.Their customer support is top notch..much better than corsairs at the moment.

I would say it best to buy ram from them and judge them on their product and support...NOT old news that needs to be left in the past.

 

[wicktron]

Bigtoe from ABX?

 

[rickn]

OCZ is crooks, they've been crooks for years. Buy at your own risk, but you wouldn't catch me dead buying anything from them

 

[orion7144]

Originally posted by: rickn
OCZ is crooks, they've been crooks for years. Buy at your own risk, but you wouldn't catch me dead buying anything from them

Your calling them crooks because some real crooks rebaged cheap ram as OCZ and sold it at a premium. It has been proven that they had nothing to do with the rebadging and people are blaming them for not honoring thier warranty on products that they did not manufacture.

 

[orion7144]

You can also check out OCZ's website and read all about the rebadging and how you can tell the real stuff from the fake. They also have info from a couple of years ago when this all happened.

 

[rickn]

Originally posted by: orion7144

Originally posted by: rickn
OCZ is crooks, they've been crooks for years. Buy at your own risk, but you wouldn't catch me dead buying anything from them

Your calling them crooks because some real crooks rebaged cheap ram as OCZ and sold it at a premium. It has been proven that they had nothing to do with the rebadging and people are blaming them for not honoring thier warranty on products that they did not manufacture.

This isn't the first time they've mispresented themselves. they must have ego's the size of new york. I remember their so called "manufacturing facilities" pictures they had on their website, which were totally bogus and ripped off some other website. Hand picked memory for reviews, the stuff end users would get would be total crap, then when they would get bad mouthed on big message boards, like AT, they'd start posting their own message from fictiticious users about how great OCZ ram was. You get the picture?

I would also add, OCZ is under new management from those days, but the bird never falls far from the nest.

 

[jaeger66]

Originally posted by: orion7144
You can also check out OCZ's website and read all about the rebadging and how you can tell the real stuff from the fake. They also have info from a couple of years ago when this all happened.

None of which has any bearing on the Overclockerz Store disasters. The RAM straight from the source was still crap.

 

[jaeger66]

Originally posted by: wixt0r
I understand that, but just the fact that the product description says "branded" rather than "produced" opens up a myriad of possibilities of what it can really mean. One could assume for days whether or not they use the IC's listed on their Semiconductors page, but the fact remains that through the wording of the description alone means that it can or can not be true.

http://www.ocztechnology.com/aboutocz/press/2002/25

Even if you were correct in this case, how is this any different from AMD/Intel binning chips and labelling differently because of yields. Just because an IC was originally labelled -5ns or -6ns does not mean that a process cannot necessarily make them a -4.3ns part. Is there actually any governing body that says there is some sort of direct correlation between original manufacturing of IC's at a particular -ns and processing the IC to run at another -ns rating? If there is, and there is proof that one cannot modify an IC to run at a different -ns rating, then and only then will I believe that OCZ is wrong. Otherwise, they are just improving product using their "EL DDR process," which is undoubtedly engineering ingenious on their part.

Speed binning is not what is going on here. And it is not the process I take issue with, it is the relabeling. NOTHING OCZ can do will ever change the fact that they are overclocking the RAM, and making it quite difficult to tell what you are buying before you buy it. There are some companies I would trust in this case(maybe), but OCZ is certainly not one of them. With a little more longevity that may change.

 

[wicktron]

Read the threads from Xtreme and ABX, they are far more informative than anything I can say. Ryan Peterson from OCZ is on both boards and contributed to both threads, and MS, who did the review for Lost Circuits regarding the OCZ Gold PC3700 contributes on the ABX thread.

 

[MS]

In all the turmoil, there are a number of issues that are easily overlooked. For once, I don't believe the story about Overclockerz store being a victim of third parties that plagiarized their brand name.

Second, there is no benefit of removing the top layer of the encapsulate. We tried it, sanding it down to a few µm above the actual die and bondwires (see XtremeSystems thread) and there was no benefit whatsoever.

OCZ claims that this is because of the differences in the process used but somehow I fail to see the rationale here.

None of the above should distract from the fact that the DIMMs we tested were actually working just like they should, even though the issue of improper documentation and lacking disclosure remains.

On the other hand, Corsair has been using -6 parts for their 3200 CL-2 parts without remarking them but whoever would be able to see that through removing the heatspreaders, would void his or her warranty...

In other words, the differences are more on a logistical level than based on overall "Quality of documentation" and that includes Kingston as well as KingMAX with their repackaged repaired chips from Actram and pretty much everybody else (TwinMOS and Winbond CH)

 

[SNN47]

While there have been a lot of quotes from the "internal" document
I can't recall any direct answers to the various detailed question that have been asked
http://www.xtremesystems.org/forums/showthread.php?s=&threadid=15875

I made a summary that I quote

While there have been a lot of replys restating always the same information, I can't recall having seen a direct answer to the question how this sanding will effect performance.

So far there was no reply to my sumary in here http://www.xtremesystems.org/forums/showthread.php?s=&threadid=15875, which I quote below
Quote

I am trying to summarize what I understood of the claims so far everything in Everything in ?? is quoted from the OCX fans and this thread:

The idea behind is to provide a ?Thermal modification of standard DDR DRAM TSOP packages , for increased thermal dissipation and speed reclassification.?

A tsop chip case ?is not even 7 millimeters deep?
Of each tsop case 7mil or 0,1778mm were taken of the top (1mil = 1/1000inch an 1 inch = 25,4mm, 1mil= 0,0254mm) which leaves the original markings readable.

An additional outer plastic film protects the ? semi-porous layer of heattransfer encapsulant that houses the DRAM die?.

This protection layer was removed in the process which is why they ?cover the chips with the heatspreader which keeps the safe from moisture?.

I hope I didn?t miss any fact .

Unless device is tested before and after modification, how can you compare values.
IMVHO whenever you claim a change you have to provide the reference information to the condition of the device before and after the modification.

e.g. if this process brings a gain of 23MHz to what was the previous reference clock-frequency this 23 MHz refer to?

Another problem is that we don?t have knowledge of the thermal properties of the heattransfer encapsulant and the outer plastic film, as well as the physical dimensions of such samsungs tsop-cases.

So in absense of reliable facts we have to accept that no one can prove nor disprove whenever an esotheric claim is made.


The cooler you operate a semiconductor below the max. operational temperature

-The more voltage/current they can handle =the faster they may be clocked within certain limits that depends not only on the design of the die but also their mounting, connection layout, of casing and PCB
-the more you increase the live expectancy and decrease aging.

Whether you use a liquid or a gas (air, liquid nitrogen) for cooling you need to circulate (= forced cooling) the coolant across the heated surface. The coolant used should be many degrees cooler then the devices surface, to give you any heatreduction at all.

The benefit of any heatsink (which includes a flat piece of heatconducting material here called heatspreader)is through the increase of the surface area which allows better heat dissipation.
Without forced cooling the efficiency of a heatsink depends on how much natural convection can be generted the way the heatsink is mounted.

Based on these facts I dare to say that adding a small fan that forces air across the surface of the chips will reduce the surfce temperature of a ram-chip more then any heatspreader or reduction of case thickness can bring under even the best conduction when only natural convection is available.

 

[LED]

Originally posted by: SNN47
There have been no direct answers to the question that have been asked
e.g. on on how sanding can physically improve performance.

From Doc himself!

Second, there is no benefit of removing the top layer of the encapsulate. We tried it, sanding it down to a few µm above the actual die and bondwires (see XtremeSystems thread) and there was no benefit whatsoever.

 

[SNN47]

Of course doc is always courteous and anwers when he is asked.

I meant a reply from the OCZ-rep that reposts always the same quotes at xtremesystems but doesn't answer directly or provides solid information to questions .



Offtopic
Since I can't reply in the forum anymore using my snn47@compuserve.de I had to use another provider, and tested it with this short test sentence.

Since I didn't get the previous "you are not allowed to post or edit" I added the rest I wanted to write, only you were faster in replying then I could edit.

 

[LED]

K...SNN47...just wanted to Raz you anyways as I knew what you meant

 

[SNN47]

I know, just wanted to make shure everyone else knew what I meant.

 

[Tab]

Damn what a bunch of fags!

 

[Smilin]

My $0.02:

If they sell PC37000 memory and it runs at PC3700 speeds there's no problem right?

Ok, what if they are selling 6ns memory at 4.3ns rated speeds? As long as it works there's no problem right?

Wrong.

4.3ns memory and 6ns memory clocked at 4.3ns are NOT the same thing - one is running beyond spec and will fail, one won't. There apparently have been some people having success with this but without access to OCZ's RMA records who's to say how well their memory is really performing.

OCZ doesn't appear to be doing anything illegal, but they certainly are doing something wrong... I don't think the FTC should start investigating them but I'm certainly not going to take my chances on buying their product. Why not just get the real deal and buy memory that's actually rated at it's intended speed rather than some that's pulling at the seams trying to reach such speed?

 

[Miramonti]

In summary, BUY SAMSUNG!

They obviously have much higher spec. standards for their ram than OCZ or they would be selling it as 4.3ns instead of 6ns!

 

[MS]

Nope, not true, I got a few Samsungs, same die, same specs and they won't go beyond 195 MHz, no matter what. Sanding them down did not help either. However, Samsung has a history of launching great product and then have their quality deteriorate, at least those are my personal experiences with them over a period of almost 25 years.

 

[wicktron]

Oh my God guys, Corsair is at it too

"Looking at the Corsair module you will notice this is in fact standard Samsung TCCC4 PC3200. Samsung even time these IC''s on their own modules at 3-4-4-8 for PC3200 operation although most will operate at much tighter timings."

ABXZone

Oh my gosh, they use PC3200 TCC Samsung chips! Oh no! (Sarcasm)

 

[jaeger66]

Originally posted by: wixt0r
Oh my God guys, Corsair is at it too

"Looking at the Corsair module you will notice this is in fact standard Samsung TCCC4 PC3200. Samsung even time these IC''s on their own modules at 3-4-4-8 for PC3200 operation although most will operate at much tighter timings."

ABXZone

Oh my gosh, they use PC3200 TCC Samsung chips! Oh no! (Sarcasm)

You still don't get it. How very, very sad. Corsair doesn't mark the chips as "Corsair" with make believe timings, so you should probably be quiet.