6
Specifications
ispGDX240VA
Control
Inputs
(from
μ
P)
The ispGDXVA Family architecture has been developed
to deliver an in-system programmable signal routing
solution with high speed and high flexibility. The devices
are targeted for three similar but distinct classes of end-
system applications:
Programmable, Random Signal
Interconnect (PRSI)
This class includes PCB-level programmable signal rout-
ing and may be used to provide arbitrary signal swapping
between chips. It opens up the possibilities of program-
mable system hardware. It is characterized by the need
to provide a large number of 1:1 pin connections which
are statically configured, i.e., the pin-to-pin paths do not
need to change dynamically in response to control in-
puts.
Programmable Data Path (PDP)
This application area includes system data path trans-
ceiver, MUX and latch functions. With today
’
s 32- and
64-bit microprocessor buses, but standard data path glue
components still relegated primarily to eight bits, PCBs
are frequently crammed with a dozen or more data path
glue chips that use valuable real estate. Many of these
applications consist of
“
on-board
”
bus and memory inter-
faces that do not require the very high drive of standard
glue functions but can benefit from higher integration.
Therefore, there is a need for a flexible means to inte-
grate these on-board data path functions in an analogous
way to programmable logic
’
s solution to control logic
integration. Lattice
’
s CPLDs make an ideal control logic
complement to the ispGDXVA in-system programmable
data path devices as shown below.
Data Path
Bus #1
Address
Inputs
(from
μ
P)
Control
Outputs
System
Clock(s)
Data Path
Bus #2
Configuration
(Switch)
Outputs
ISP/JTAG
Interface
ispLSI/
ispMACH
Device
ispGDXVA
Device
Buffers / Registers
Decoders
Buffers / Registers
State Machines
Figure 4. ispGDXVA Complements Lattice CPLDs
Applications
Programmable Switch Replacement (PSR)
Includes solid-state replacement and integration of me-
chanical DIP Switch and jumper functions. Through
in-system programming, pins of the ispGDXVA devices
can be driven to HIGH or LOW logic levels to emulate the
traditional device outputs. PSR functions do not require
any input pin connections.
These applications actually require somewhat different
silicon features. PRSI functions require that the device
support arbitrary signal routing on-chip between any two
pins with no routing restrictions. The routing connections
are static (determined at programming time) and each
input-to-output path operates independently. As a result,
there is little need for dynamic signal controls (OE,
clocks, etc.). Because the ispGDXVA device will inter-
face with control logic outputs from other components
(such as ispLSI or ispMACH) on the board (which fre-
quently change late in the design process as control logic
is finalized), there must be no restrictions on pin-to-pin
signal routing for this type of application.
PDP functions, on the other hand, require the ability to
dynamically switch signal routing (MUXing) as well as
latch and tri-state output signals. As a result, the pro-
grammable interconnect is used to define possible signal
routes that are then selected dynamically by control
signals from an external MPU or control logic. These
functions are usually formulated early in the conceptual
design of a product. The data path requirements are
driven by the microprocessor, bus and memory architec-
ture defined for the system. This part of the design is the
earliest portion of the system design frozen, and will not
usually change late in the design because the result
would be total system and PCB redesign. As a result, the
ability to accommodate arbitrary any pin-to-any pin re-
routing is not a strong requirement as long as the designer
has the ability to define his functions with a reasonable
degree of freedom initially.
As a result, the ispGDXVA architecture has been defined
to support PSR and PRSI applications (including bidirec-
tional paths) with no restrictions, while PDP applications
(using dynamic MUXing) are supported with a minimal
number of restrictions as described below. In this way,
speed and cost can be optimized and the devices can still
support the system designer
’
s needs.
The following diagrams illustrate several ispGDXVA ap-
plications.