• 
• 

Monday, January 7, 2008

Flash Memory Basics

What's the anatomy of a memory card?

SLC, MLC, Bits And Bytes To Die 4
Once the wafers' production is complete, they're cut up into individual pieces called “die.” The die are then packaged to make chips (Figure 3). Most memory cards use what are called packaged flash chips. The most commonly used chip is called a “TSOP” or “Thin Outline Small Package” chip.

Raw flash die are extremely delicate and only can be handled with special equipment. Some smaller-card form factors, such as microSD, use bare die in special chip-onboard assembly processes, but for the most part, CompactFlash and SD cards are made with packaged TSOP flash chips that can be handled easily with robotic assembly machines in a standard surface-mount assembly process.

Flash is typically packaged one of three ways—one die per chip (mono), two die per chip (a dual-die package, DDP) or four die per chip (a quad-die package, QDP).

Raw-flash units are measured in term of bits, while memory-card units are measured in bytes. There are 8 bits to one byte. Bits are indicated with a lowercase “b” whereas bytes are indicated with a capital “B.”

Samsung makes two types of flash. The first, called Single Level Cell (SLC), currently can be produced in die sizes up to 8 gigabits (Gb) or 1 gigabyte (GB) per die (Figure 4a). This means that a mono TSOP has the capacity of 1 GB, a DDP 2 GB and a QDP 4 GB. If you consider a CompactFlash card that can typically hold up to 4 TSOP chips, the largest card you can produce using this packaged flash is 16 GB. Using this same flash in an SD card that has two chip placements would yield a maximum capacity of 8 GB.