Samsung SSD 840 (250GB) Review
by Kristian Vättö on October 8, 2012 12:14 PM EST- Posted in
- Storage
- SSDs
- Samsung
- TLC
- Samsung SSD 840
Lower Endurance—Why?
Below we have a diagram of a MOSFET (Metal Oxide Semiconductor Field Effect Transistor). When programming a cell, voltage is placed on the control gate, which forms an electric field that allows electrons to tunnel through the silicon oxide barrier to the floating gate. Once the tunneling process is complete, voltage to the control gate is dropped back to 0V and the silicon oxide acts as an insulator. Erasing a cell is done in a similar way but this time the voltage is placed on the silicon substrate (P-well in the picture), which again creates an electric field that allows the electrons to tunnel through the silicon oxide.
While the MOSFET is exactly the same for SLC, MLC and TLC, the difference lies in how the cell is programmed. With SLC, the cell is either programmed or it's not because it can only be "0" or "1". As MLC stores two bits in one cell, its value can either be "00", "01", "10" or "11", which means there are four different voltage states. TLC ups the voltage states to eight as there are eight different combinations of "0" and "1" when grouped in groups of three bits. Below are diagrams showing the graphical version of the voltage states:
SLC
MLC
TLC
The above diagrams show the voltages for brand new NAND—everything looks nice and neat and the only difference is that TLC has more states. However, the tunneling process that happens every time the cell is programmed or erased wears the silicon oxide out. The actual oxide is only about 10nm thick and it gets thinner every time a smaller process node is introduced, which is why endurance gets worse as we move to smaller nodes. When the silicon dioxide wears out, atomic bonds break and some electrons may get trapped inside the oxide during the tunneling process. That builds up negative charge in the silicon oxide, which in turn negates some of the control gate voltage when the cell is programmed.
The wear results in longer erase times because higher voltages need to be applied for longer times before the right voltage is found. Remember, the controller can't adjust to changes in program and erase voltages (well, some can; more on this on the next page) that come from the trapped electrons, cell leakage, and other sources. If the voltage that's supposed to work doesn't, the controller has to basically go on guess basis and simply try different voltages before the right one is found. That takes time and causes even more stress on the silicon oxide.
The difference between SLC, MLC, and TLC is pretty simple: SLC has the fewest voltage states and hence it can tolerate bigger changes in voltages. With TLC, there are eight different states and hence a lot less voltage room to play with. While the exact voltages used are unknown, you basically have to divide the same voltage into eight sections instead of four or two like the graphs above show, which means the voltages don't have room to change as much. The reason why a NAND block has to be retired is that erasing it starts to take too long, which impacts performance (and eventually a NAND block simply becomes nonfunctional, e.g. the voltage states for 010 and 011 begin to overlap).
There is also more and more ECC needed as the NAND wears out because the possibility for errors is greater. With TLC, that's once again a bigger problem because there are three bits to correct instead of one or two. While today's ECC engines are fairly powerful, at some point it will be easier to just retire the block than to keep correcting errors.
Facebook
Twitter
RSS
86 Comments
View All Comments
Kristian Vättö - Tuesday, October 9, 2012 - link
Samsung will be releasing Magician 4.0 later this year and we are definitely going to do an article about it. I tried to get a beta version of it but Samsung wasn't willing to send us one, yet. Magician 3.2.2 that comes with the 840 is the same that's available for the 830, so I didn't find it that important to cover.But you are right, we need to cover the software side more as well because it will also drive manufacturers to make their own SSD tools.
P.S. Try installing Ubuntu to a USB stick (just use UNetBootin) and then use hdparm commands (see below) to secure erase the SSD.
sudo hdparm --user-master u --security-set-pass password /dev/sdX
sudo hdparm --user-master u --security-erase password /dev/sdX
"password" is whatever you choose as the password and "X" is the drive (you can see which it is in Disk Utility)
Impulses - Monday, October 8, 2012 - link
At one point in the article you mention 830 users would at least wanna upgrade to the 840 Pro if they aren't gonna wait for next gen drives... I kinda thought the Pro qualified as next gen, with the vanilla 840 being more of a side grade from current gen drives. So what would you consider or expect from a next gen drive?KAlmquist - Tuesday, October 9, 2012 - link
One thing I've been wondering is what voltage the controller uses. I can't identify every component on the board, but I don't see anything that looks like a voltage regulator. It seems kind of crazy to run a chip with three processor cores at 5 volts, but I guess it's possible.Death666Angel - Tuesday, October 9, 2012 - link
Loved the review. I'm really looking forward to more data on the DSP usefulness inside SSDs, could be a huge advantage (though I don't worry about going through my P/E cycles at home).With SSDs, we are at a point similar to GPUs and CPUs where I don't understand upgrading from generation to generation. Unless you have a very specific usage case and can make great use of the better speed (video/picture editing or something equally I/O intensive), I don't think a 840 Pro will feel that much faster than a 830.
I'm still rocking a first gen Agility 60GB in my (ULV) notebook and a Vertex2 120GB in my desktop and they feel plenty fast, especially those times when I'm at a friends PC with only mechanical storage.
I might upgrade to a 830 or 840 (non Pro) when the prices have dropped for the 830 or stabilized for the 840 (I don't like paying for launch pricing). :D
And I finally got my brother to go to SSD storage, too! He'll either get a 512GB 830 or 2x256. Does anyone know why 256GB usually has the best price/GB instead of the 512GB drives? Considering the cost for casing, PCB, controller etc. stays the same, I would expect the 512GB to have better pricing/GB. Do they need to go with more dense packages which are more expensive?
name99 - Tuesday, October 9, 2012 - link
I agree completely.I've said before that the current emphasis on performance over everything else is as stupid as Intel's Pentium4 philosophy. Devices have to become more balanced if they want to sell better.
In particular the current atrocious write power draws for these devices mean that they will not sell in the external drive market --- they can't be fully powered by USB2 and
- no-one wants a drive that needs two USB connections because it's clumsy and it uses up limited USB slots
- saying it can be fully powered by USB3 is not good enough because, for the near-term future, people want to share drives between USB2 and USB3 computers.
Of course MOST, but not all, the external drive market is about capacity, not speed. However each time SSD capacity doubles, a larger portion of the external drive market makes sense; AND power is not only an issue for the external market. The higher you set your peak power draw, the less your flash appeals to Ultrabook manufacturers.
Just remember Pentium 4 and its power draw, and compare to Intel trying to get Haskell to run at <10W nowadays. Balance matters, and current flash manufacturers seem to be far too unbalanced compared to the requirements of most users.
(It would be fascinating, for example, if we learn that the flash in the next round of iOS devices, or even MacBooks, was designed in-house by Apple because they were not satisfied with the power vs performance choices that were being made by the major flash manufacturers. Remember --- Apple bought Anobit...)
moadip - Tuesday, October 9, 2012 - link
does the 830/840 family support AES encryption? Want to buy one of those but there is no official word on any support for this. Is anyone of you guys/girls using them in that way?Kristian Vättö - Tuesday, October 9, 2012 - link
840 and 840 Pro both support 256-bit AES encryptionmoadip - Tuesday, October 9, 2012 - link
what about the 830?Kristian Vättö - Tuesday, October 9, 2012 - link
I'm not 100% sure but as far as I know, only the OEM/enterprise version (i.e. PM830) supports encryption, the consumer 830 doesn't.MrSpadge - Tuesday, October 9, 2012 - link
What an amazingly well balanced drive! Currently it's more expensive than the 830 at similar capacities, but as soon as 21 nm yield ramps up (and prices are adjusted) this will probably be one of the best or maybe the best consumer SSD on the market.