How is 3D TLC NAND more reliable than MLC?

[c627627]

Example 1: ADATA Ultimate SU800 3D TLC NAND SSD
http://www.anandtech.com/show/10601...800-ssd-smi-controller-3d-nand-sata-interface
says
Thanks to higher endurance of 3D TLC NAND compared to planar flash memory made using thin process technologies, ADATA declares 2 million hours MTBF..." (Mean time between failures).

Example 2: Toshiba Q300 Pro MLC SSD
http://us.toshiba.com/storage/internal/ssd
1.5 million hours MTBF (Mean time between failures).

So in real life what does a claim of
3D TLC NAND being 2 million hours MBTF
vs
MLC NAND being 1.5 million hours MBTF
mean?

TLC is by definition less reliable than MLC.
3D TLC NAND is better than TLC but is the claim than 3D TLC is more reliable than MLC exaggerated in the example of those two drives...

How does ADATA's 3D TLC NAND compare to Toshiba's MLC NAND in real life?

 

[Elixer]

In a nutshell, it is a different tech, planar uses floating gate transistors, and 3D uses Charge Trap Flash.
Before (planar), as the nodes got smaller, it became less reliable (which is why planar TLC was worse than MLC), now with 3D, they are actually bigger, giving more reliability.

If you want more information, then http://www.samsung.com/us/dell/pdfs/EE-Times-article-final.pdf is a good paper to read.

 

[Valantar]

In a nutshell, it is a different tech, planar uses floating gate transistors, and 3D uses Charge Trap Flash.
Before (planar), as the nodes got smaller, it became less reliable (which is why planar TLC was worse than MLC), now with 3D, they are actually bigger, giving more reliability.

If you want more information, then http://www.samsung.com/us/dell/pdfs/EE-Times-article-final.pdf is a good paper to read.

Yep. Just as 32nm TLC is more than likely more endurant than 15nm MLC. It's all about the amount of trapped electrons and barrier thickness/permeability.

 

[c627627]

Thank you for that educational paper. I feel good about replacing my 128GB Toshiba Q300 Pro MLC SSD with 256GB ADATA Ultimate SU800 3D TLC NAND SSD.


I do have another question for you, let's take it a step further:
How does a 500GB Samsung 850 EVO non Pro compare to 512GB ADATA Ultimate SU800 3D TLC.
Controller, performance and most importantly, reliability.

 

[Billy Tallis]

The most important thing to realize is that MTBF is just a meaningless fantasy. You'll never derive useful information from comparing those ratings across different vendors. Look at actual write endurance ratings to get an idea of which drive will last longer.

In a nutshell, it is a different tech, planar uses floating gate transistors, and 3D uses Charge Trap Flash.
Before (planar), as the nodes got smaller, it became less reliable (which is why planar TLC was worse than MLC), now with 3D, they are actually bigger, giving more reliability.

The 3D NAND in question is IMFT's, which actually still is a floating gate design rather than charge trap. And switching to 3D NAND was a one-time boon to write endurance; as layers are added, the cell sizes are being held roughly constant, not growing.

 

[c627627]

Yes, when doing research, I come across a lot of posts that say two things (just like yours):
Warnings not to look at MTBF and explanations of what 3D NAND is.

But what would help specifically is to look at the two example drives in post#1, one clearly using MLC and the other 3D TLC and to post how they (specifically) would differ when it comes to their controllers, performance and most importantly, reliability.

It would also help if you or someone else could similarly compare 3D TLC of ADATA to Samsung 850 EVO non Pro.

Then reading your opinions on comparisons would tell us more than MTBF could, I agree...

 

[Elixer]

The 3D NAND in question is IMFT's, which actually still is a floating gate design rather than charge trap. And switching to 3D NAND was a one-time boon to write endurance; as layers are added, the cell sizes are being held roughly constant, not growing.

Thanks for the clarification...

A planar floating-gate NAND technology has previously realized a 0.87Gb/mm2 memory density using 3b/cell [1] and achieved a minimum feature size for 16nm [2]. However, the development of planar NAND flash is expected to reach the scaling limit in a few technology generations. To break though this limit, a significant shift to 3D NAND flash has begun and several types of 3D memory cell structures have been proposed and discussed [3-5]. Recently a 3D V-NAND technology achieved 1.86Gb/mm2 using charge-trap cells and 3b/cell [6]. This paper presents a 3b/cell NAND flash memory utilizing a 3D floating gate (FG) technology that achieves 4.29Gb/mm2.

Too bad you need to be a IEEE member to read it.

 

[Schlitzy]

In a nutshell, it is a different tech, planar uses floating gate transistors, and 3D uses Charge Trap Flash.
Before (planar), as the nodes got smaller, it became less reliable (which is why planar TLC was worse than MLC), now with 3D, they are actually bigger, giving more reliability.

If you want more information, then http://www.samsung.com/us/dell/pdfs/EE-Times-article-final.pdf is a good paper to read.

Hey there Elixer, considering this document you cited is approaching 2 years old this October, how true is all of this still today? I'm just getting caught up here on what's currently available and looking to make a few purchases.

Thank you and thanks everyone.

Schlitzy