© 2011 Rochester Electronics, LLC

Page 3

July 2011

All Rights Reserved

 

Gate array – A prefabricated silicon chip integrated circuit made up of digital logic gates (such as

standard NAND or NOR logic gates and transistors) in predefined positions. The digital gates

remain unconnected, and the chip is not intended to have a particular function. To customize the

chip to perform a specific function, the manufacturer adds a final surface layer (or layers) of metal

interconnects that join the gates in such a way as to produce the required functionality.

Standard cell – An integrated circuit made of a combination of circuit blocks from previous

designs that are combined to create a new design. Each layer is a unique design, which

differentiates this from a gate array. This is one of the most expensive substitute components, but

may be required depending upon design content.

Field programmable gate array – A programmable part that is customized by the user through a

program resident in a companion memory device or from a processor. FPGAs can be a less expensive

choice up front; however, they are particularly vulnerable to near-future obsolescence issue themselves.

 

Why aren’t these substitutions true form-fit-function replacements?

All of these substitute solutions are created using state-of-the-art silicon foundry technology. While this sounds

like a good thing, it can actually introduce some negative aspects. This new technology may be suitable to

replace the functionality of the original device, but it is almost certainly does not replicate all characteristics of the

original. This is because semiconductor manufacturers are always trying to get more functional elements onto

their silicon, and, to do so, they are continually shrinking the geometry of the elements that go into the silicon.

This can bring improved performance to systems and also lower costs, but these benefits come with

disadvantages that are likely to be significant in high-reliability and safety-critical applications.

 

What are the differences between the substitutes and the originals?

 

The major difference is the smaller functional cell size used in the new technology. This leads to:

•

Higher device switching speeds (more susceptibility to noise)

•

Different capacitance (board-level loading changes)

•

Different radiation tolerance

•

Different EMC performance

These parameters are usually not recognized by the manufacturers who make the substitutes because they are

not usually fully specified and tested for during device production.

 

What problems do these differences cause?

 

Higher switching speeds and different capacitance can cause devices to see spikes and signals or to produce a

race condition in an application that the original parts did not. Any of these conditions could lead to equipment

malfunction.

Semiconductors based on smaller cell size are usually less radiation tolerant than semiconductors produced

using older technologies. This can lead to latch-up or loss of data. If such a component were used in an avionics

systems operating at high altitude, for example, aircraft safety would be compromised.

Semiconductors based on smaller cell size are also usually less robust electrically. This can make them more

prone to interference from radio and TV transmitter signals and more vulnerable to damage from electrostatic

discharges. The latest consumer-driven semiconductor technologies are not designed for 10-year life at the full

industrial/military performance limits when it comes to metal migration. These technologies compromise on

environmental temperature limits and voltage ranges to achieve similar lifetime guarantees as older technologies.