The choice comes down to how you will use the memory
BY JEFFREY BAUSCH
NAND flash memory has experienced tremendous growth in the last decade. In fact, market revenue has increased from $370 million in 2000 to over $20 billion in 2010. ( www.reportlinker.com/p0571408-summary/Low-Density-SLC-NAND-Flash-Memory-in-Embedded-Applications.html)
Today, NAND flash memory is largely recognized as the preferred memory storage device of choice for consumer, industrial, and computing applications, with each market having its own performance, feature and density requirements.
Two of the more popular types of NAND flash memory architecture are: single-level cell (SLC) and multi-level cell (MLC). A quick snapshot of the advantages and disadvantages for both technologies reveal some of the key differences between the two:
| SLC Advantages | SLC Disadvantages |
| Fast transfer speed | High manufacturing cost per Mbit |
| Low power consumption | Low density |
| High write/erase cell endurance | |
| MLC Advantages | MLC Disadvantages |
| Low manufacturing cost per Mbit |
Slow transfer speed |
| High density | High power consumption |
| Low write/erase endurance |
SLC NAND: A great choice when performance matters
SLC NAND flash memory stores one bit of data per memory cell and so a single voltage level is required for the device to detect one of the memory’s two possible states. If a current is sensed, then the bit is stored as “0” (meaning “programmed”). If there’s no current, then the bit is “1” (meaning “erased”).
Due to this rather simple architecture, SLC NAND is able to offer quick read and write capabilities, long-term durability and rather simply error correction algorithms.
Its simplicity, though, does have one major downfall (besides its obvious limited programming density): SLC NAND flash is considerably more expensive per bit when compared to other NAND technologies. This is due to the fact that since each bit stores only one bit of data, designers must choose between cost and performance, and unless you’re ordering your SLC NAND from a flash memory manufacturer located in the woods of some back country, performance (should) always matter first and foremost.
Go with MLC NAND if density and cost are more important
MLC NAND is able to store at least two bits of data per memory cell and so multiple voltage levels are necessary to decipher between the memory’s four possible states. The different states per memory cell are recognized as “00”, “01”, “10”, and “11” and they range in value from fully programmed (“00”) to partially programmed (“01”), partially erased (“10”) and fully erased (“11”).
Two advantages of this increase in bits per cell are that the memory’s capacity is greater and production costs are reduced. As a result, MLC NAND can be sold at a much lower price-per-bit than SLC NAND and also offer twice the density.
Due to the complexity of its architecture, though, MLC NAND’s programming/reading capabilities are much slower than SLC NAND. As a matter of fact, SLC NAND is nearly three times faster than MLC NAND, and several times more durable. Also, because multiple levels of voltage are needed in order to read between the memory’s four states, there is greater power consumption and wear on the cells.
Important considerations when choosing between MLC and SLC NAND flash memory:
• Durability : SLC NAND lasts approximately 100,000 programming cycles. This is 10x’s more than MLC NAND, whose program and erase capabilities last about 10,000 cycles.
• Programming : Coupled with MLC NAND’s limited cycle count, one must also consider the fact that flash memory cannot perform partial page programming – the technology’s algorithm is specifically set up this way to increase efficiency and to limit wear on the cells. As a result, it cannot change or edit new data to a page after it’s already been programmed without first fully erasing the page’s data. This is an important consideration to programmers, who might be more inclined to go with the longer lasting SLC NAND instead.
• Energy : SLC NAND uses only two states per one bit of data stored and so is able to use less energy when managing the component’s electrical charge during operation. Since MLC NAND has four states, along with two bits of data stored, it must use more energy in managing its multiple levels of electrical charge.
• Temperature range : Higher temperatures can lead to leakage in cells which will, in turn, cause the voltage sensors for an MLC NAND flash memory (with its higher bit density) to read the wrong levels. In an industrial setting where NAND-based flash memory devices are exposed to greater temperature fluctuation, it makes sense to use SLC NAND flash memory, where its smaller bit density reduces the chance of this possible error.
While it might seem like MLC NAND is the lesser of the two flash memory technologies because it has more disadvantages, it should be noted that these shortcomings can be alleviated or in some cases even reversed by making changes to the solid-state drive’s (SSD) internal design structure (interleaving, changing the memory’s writing algorithm, creating even greater capacity).
Keep in mind, though, that this can drive a higher price tag due to the increased complexity of the software.
Bottom line
Professionals looking for a high-performance NAND flash memory system should go with the SLC NAND. Consumers searching for a flash memory device with substantial capacity will likely be pleased with their reasonably-priced, MLC NAND-based device instead.
For more information on MLC and SLC NAND flash memory, check out Scott Nelson’s piece, MLC and SLC NAND Flash Design Tradeoffs ( http://www2.electronicproducts.com/MLC_and_SLC_NAND_flash_design_tradeoffs-article-FAJH_Toshiba_Sep2008-html.aspx) . ■
Advertisement
Learn more about Electronic Products Magazine
