PROBLEMS SOLVED BY FLASH BASED MICROCONTROLLERS
Charles Melear and Kevin Kilbane
6501 William Cannon Dr. West
Austin, Texas 78735-8598
Using microcontrollers with integrated FLASH memory modules provides greater flexibility with regard to field
programmability, remote data storage, and prototyping. In fact, using FLASH based microcontrollers for volume
production now makes economic sense. FLASH, being a programmable, non-volatile memory technology, allows a
FLASH device or a microcontroller with integrated FLASH to be programmed just before final assembly of a printed
circuit board and in many cases, after the microcontroller is actually installed on the board. FLASH EEPROMs are
fundamentally different from One Time Programmable devices in that FLASH modules are electrically erasable.
One Time Programmable devices are essentially Ultraviolet Erasable devices that are put in a plastic packages
with no window. Being in a plastic package, there is not a mechanism for applying an ultraviolet light to the chip.
Thus, once it is programmed, the chip cannot be erased. FLASH, on the other hand, is electrically erasable and
does not require ultraviolet light. FLASH devices can be repeatedly erased and reprogrammed while installed in
the user application. In fact, a FLASH based microcontroller can execute a program that will erase its own FLASH
memory and reprogram the memory while installed in the end application.
WHAT DOES FLASH BRING TO THE PARTY
The economic justification for using FLASH based microcontrollers has generally been along the line of inventory
control and program modification once the microcontroller is installed in its final application. Because FLASH
based devices can be programmed at any time in standard programming fixtures, inventory control is facilitated.
Only the exact number of devices needs to be programmed as there are modules to build. Obviously, any
"unprogrammed" devices can be used for something else. Previously, high volume applications generally used
ROM-based devices. Logistical problems with ROM-based microcontrollers have to be carefully managed. Hardly
anyone, particularly accountants, like to hold inventory. Because of supply liabilities, ROM-based devices must be
ordered well in advance of when they are needed. Ten weeks and greater of ordering lead time is quite common.
Then, someone must have accurately determined how many devices will actually be needed. Otherwise, if
manufacturer using the microcontroller got an unexpected upside on the demand for their units, the waiting time to
get additional microcontrollers might be unacceptable. For instance, Christmas always occurs on December 25.
Toys must be on the shelves by that date and there is just no forgiveness for being late. On the other hand,
ordering in excess of current needs will leave one with excess inventory which may or may not be usable at a later
FLASH MAKES ECONOMIC SENSE
As has already been stated, there are economic justifications for using the more expensive, reprogrammable
FLASH-based units. As always, the technology steamroller continues to charge forward. Advanced technology
has greatly affected the cost of FLASH when integrated onto a microcontroller in both absolute and relative terms.
The cost of a silicon chip is determined by the cost of an integrated circuit wafer. Count the number of good chips
and divide that into the cost of the wafer and die cost is obtained. FLASH memory has, and still is, physically larger
than its fixed ROM counterpart. However, the differential in size between ROM and FLASH has dramatically
narrowed. For instance, when FLASH was first introduced, a FLASH device was about four times bigger than its