Page 298 - Semiconductor Manufacturing Handbook

Geng(SMH)_CH19.qxd 04/04/2005 20:00 Page 19.25

INSPECTION, MEASUREMENT, AND TEST

INSPECTION, MEASUREMENT, AND TEST 19.25

Future Trends in Flash Memory. As mentioned previously, Flash memory is an offshoot of
EPROM. It has virtually replaced EPROM and EEPROM due to its better capability and lower cost.
Inevitably, this will also happen with Flash memory. Several technologies are being investigated that
could replace Flash memory, such as ferroelectric RAM (FRAM), magnetic RAM (MRAM), ovonic
unified memory (OUM), and polymer memory and spin memory. There is also a desire to combine
the capabilities of SRAM, DRAM, and Flash memory into one unified memory type that does what
all three can do, but is cheaper to manufacture.
Future Trends in Flash Memory Testing. Since Flash memory is very time consuming to test, the
IC manufacturers are looking at how to reduce or eliminate the testing needs. One area that is being
pursued is BIST. By adding circuitry that can diagnose the device and log the memory array’s errors,
the job of the test system becomes greatly simplified. The addition to the device real estate must be
weighed against the cost reduction in test time and tester capital expense. This will be discussed fur-
ther in Sec. 19.5.

19.2.7 Testing Mixed-Signal Devices

Mixed-Signal Market Segments. Mixed-signal devices are very prevalent in our everyday lives.
Some of the most common include devices that allow such products as cellular telephones, hard disk
drives, Internet devices that access Internet service providers (ISPs), multimedia audio and video,
optical disk storage, games, and medical equipment to work properly in a convenient form factor.
Definition of a Mixed-Signal Device. A mixed-signal device consists of both analog and digital
components combined or integrated into a single device to meet a function or application need. The
analog section is one that deals with signals that are varying in time as well as amplitude and have
similar functionality to linear devices such as amplifiers and filters. The digital section can use logic,
state machines and processors to accomplish its tasks.

The Basic Building Blocks of a Mixed-Signal Device. The digital section functionality and orga-
nization are similar to that described in the digital section of this chapter. A more specialized digital
function includes DSP, which is a mathematical technique used to process a signal according to a
specific need. A typical use of DSP is to convert the time domain data into data more suited to CPU
math operations; for example, converting time domain data into the frequency domain ( fast Fourier
transform) and then performing the math to arrive at an actual test parameter number such as total
harmonic distortion. Other uses included complex signal processing such as signal modulation or
demodulation or digital filtering. Other digital blocks include CPUs, digital logic, and embedded
memory.
The analog is the section of the device that deals with signals that vary in time as well as ampli-
tude. These signals are usually considered real-world type signals, because all signals of this nature
usually are not related to a fixed timed event but can occur randomly and are nonperiodic. The most
common analog building blocks are amplifiers and filters.
The converters section of the device can convert an analog signal, or one that is varying in time
and amplitude into a digital representation of the applied signal (A to D converter) or can take a dig-
ital signal and convert it into an analog or time and amplitude-varying signal (D to A converter). The
converters provide the interface between the analog domain and the digital domain. The digital
domain data can be modified or transferred without any deterioration and are more suited to digital
logic and computation while the analog representations are more suited for human use, such as audio
and video signals and are vulnerable to noise and distortion.
The filtering section uses techniques to remove or limit unwanted signals from the signal of inter-
est. This filtering can be in the form of analog filtering, usually occurring in the time domain, using
active components such as op amps or passive components such as resistors, inductors, and capaci-
tors. Digital filtering is a math process using either a convolution operation in the time domain or
FFT-based operation in the frequency domain and will either subtract unwanted or enhance desired
components.

Downloaded from Digital Engineering Library @ McGraw-Hill (www.digitalengineeringlibrary.com)
Copyright © 2004 The McGraw-Hill Companies. All rights reserved.
Any use is subject to the Terms of Use as given at the website.