CMT4GX3M2A2133C9 Withdrawn?

that saved ya the trouble of trying to mix 2 kits unless it was for a different board..

Wat? :confused:

There was nothing wrong with it just the memory IC we used to make that module is no longer available. And we have faster modules available since that was released. Like CMXGTX3 for example.

Thanks for clearing that up.

Well, I finally found why the memory was pulled after a lot of hunting around. In order to run this memory at 2133mhz, you have to have the voltage at 1.65. Problem is that the Sandy Bridge spec has a memory voltage limit of 1.5. Motherboard vendors have gotten the board to overclock to 1.65 but that is the absolute top limit.

The Corsair 4gb kit will work at 1.65 volts, but what happens if you try and add another 4gb? The machine will not work because it requires more power for the other two modules.

I ran into this problem when I decided I wanted to buy another 4gb kit to run 8gb of this memory. Corsair pulled the memory as a result. You can buy memory that runs at 2000mhz but 2133mhz doesn't cut it for more than 4gb. Unless you buy memory that is much more expensive running at 1.5 volts.

I emailed Corsair several times on this issue and was unable to get any kind of response. It was a mistake to release this memory due to a fundamental problem which must have been overlooked when the product was released.

Well, I finally found why the memory was pulled after a lot of hunting around. In order to run this memory at 2133mhz, you have to have the voltage at 1.65.

There was nothing wrong with it just the memory IC we used to make that module is no longer available.

I emailed Corsair several times on this issue and was unable to get any kind of response. It was a mistake to release this memory due to a fundamental problem which must have been overlooked when the product was released.

Sorry but you are wrong. Let's have a look at the datasheet directly from Intel

Have a look at page 11 - Section 1.2.1 for System Memory Support

DDR3 I/O Voltage of 1.5

Then look at page 80: Processor System Agent I/O Buffer Supply DC Voltage and Current Specifications

VDDQ - Processor I/O supply voltage for DDR3

Minimum: 1.425 V

Typical: 1.5 V

Maximum: 1.575 V

As you can see, 1.65 volts is not part of the Sandy Bridge memory controller spec.

The point is that there are many people who are posting on this forum because they are having difficulty with this memory. I seem to be one of very few who actually has an answer as to why this memory is problematic. The poor guy who started this thread has other posts where is cannot figure out if he has bad mem, bad mobo, or bad proc. He doesn't realize that 2133 at 1.65 is too much. That is why I believe the mem was pulled. Look on the site now under Dual Channel Intel Memory Upgrades. There are no modules running at 2133, although there are modules running at 2000 with 1.65.

Corsair might have been too ambitious, so it makes sense to pull the memory. After all 1.65 is above the Intel spec, as I identified from their spec sheet. But it's all a give and take, when you are doing an OC. So they dropped the rated speed to 2000 and kept the 1.65 on some modules.

My specific issue is that I cannot run 8gb of this mem. I am lucky enough to have a stable machine running 4gb at 2133. But in order to run 8gb, I have to purchase a different kit. So it was a bit of a waste of money and there could have been a disclaimer. Or at the very least, an official explanation as too why it was pulled to satisfy those who had already purchased it and were looking for more.

No, I'm right. I quoted the Corsair memory specs, while you quoted Intel's memory specs and the memory controller specs. Two different items, two different specs. Overclocking often needs more voltage than the basic JEDEC specs, but that doesn't mean more voltage will blow anything up either.

People post about having memory issues here because that's the purpose of the forum.

As for no modules running at or above 2133 MHz, they do indeed exist:

CMT16GX3M4X2133C9 - 2133 MHz @ 9-11-10-27 @ 1.5v.

CMGTX8 - 2400 MHz @ 10-12-10-27 @ 1.5v

As for your issue, mixing memory is never recommended because the two kits simply haven't been tested / guaranteed together. As for an "official explanation", that was posted in this very thread 8 months ago!

Uhm, did you notice that the memory your are using as an example is running at 1.5 volts? That is the central point of my argument. You cannot run 2133 at 1.65 volts. Corsair realized this and pulled the mem.

The original box and online ad for this memory stated it was 100% stable and tested as an Intel memory upgrade. Clearly it is not. If I wanted to make a big deal about it, I could. The product cannot perform as advertised. You should recognize this.

I understand but the the Intel Sandy Bridge spec sheet states the following:

Minimum: 1.425 V

Typical: 1.5 V

Maximum: 1.575 V

How can you argue that there is not a problem? That is why the memory was pulled and is no longer being sold by Corsair. It has been substituted with memory that runs at 2133 with 1.5 volts.

Sharkfoot, i have an I-7 920 (first generation )that i have been using 1.65v memory(Dominators) with for the past two years with out any problems or frying any components. XMP is Intels technology that automatically overclocks the RAM.

When you enable XMP on those systems it sets the voltage to 1.65v for overclocking purposes and also boosts memory controller voltage from default of 1.2v to 1.35v too.(mostly x58 chipset)

When the first generation I series CPU's were released there was no 1.5v memory available. So they recanted their statement and said that 1.65v memory would run fine.

Now with the second generation (Sandybridge P68/z68 chipset) they are pretty adamant that you use 1.5v memory. But there are still people that use 1.65v memory(Dominators) on those platforms too without any issues. Corsair also offers RAM (Vengeance series) that will overclock to higher frequencies and still maintain 1.5v.

There are also certain sets of Domminators that will overclock at 1.5v too.