IEEE Electrification Magazine - December 2017 - 51

cores among microprocessor manufacturers, and they are
suitable options for the CCPS digital core.

FPGA in the Aerospace Industry
The FPGA concept was first introduced by Ross Freeman
in 1984. FPGAs are digital integrated circuits that consist of
logic blocks and interconnections that can be programed
to conduct a wide array of functions. The combination of
different components within the FPGA, the flexible nature
of the FPGA architecture, and the consistently increasing
library of IP modules ensure that FPGAs can handle a wide
variety of tasks. Due to their reprogrammable nature,
FPGAs have advantages over their fixed counterpart, the
application-specific integrated circuit. These benefits
include significantly less nonrecurring expense, a simpler
design cycle, faster time-to-market, and the ability to
modify the device at any stage of its life-cycle, including
when it is being used by an end user. The ability to modify
the device in the field is particularly attractive as it offers a
way of at least partially avoiding obsolescence, because
the FPGA can be updated to include new functionality and
remove obsolete modules.
FPGA memory is based on flash memory and static RAM
(SRAM). SRAM-based FPGAs are volatile, which means that,
every time they lose power, they need to be reconfigured using information from an external, nonvolatile memory source. Flash-based FPGAs, on the other hand, are

nonvolatile and, thus, keep their configuration when powered down. In logic systems, events called single-event effects
(SEE) can occur as a result of cosmic rays or high-energy particles. These SEEs can take many different forms, such as
single-event upsets (SEUs) and single-event latch-ups (SELs).
SEEs can flip bits in either memory or registers as well as
create transient pulses that propagate through the system,
while SELs cause higher-than-safe operating currents. SEUs
are a major concern for safety-critical systems such as avionics and can be especially dangerous when they cause
changes in the configuration memory of the FPGA. The flash
memory is more resistant to SEUs with the limited number
of writes that can be used to reprogram it. There are different FPGA suppliers, including Xilinx Inc., Altera Corporation
(now part of Intel), Lattice Semiconductor, Microsemi, Atmel,
QuickLogic, and Achronix Semiconductor Corporation. FPGA
manufacturers offer a wide selection of FPGA families that
address varying needs, ranging from high-capacity, performance-intensive networking to small-footprint FPGAs. The
composition of the embedded math blocks that each device
has varies; e.g., Altera and Lattice devices available on the
market only have a single multiplier for each of their math
blocks, while Xilinx and Microsemi devices have multiple
math components per one math block. Table 3 compares
devices from some of these manufacturers.
The maximum frequency of FPGA clock trees and the
number of logic cell (LC)/logic element (LE)/adaptive logic

TABLE 3. Comparison of various FPGAs.
Devices

LCs (k)

Maximum Internal
Clock (MHz)

Math
Blocks

Operating
Temperatures (°C)

Maximum Approximation
Price (US$)

6SLX45T (Xilinx Spartan-6)

43.661

400

0-85

88.83

0-100

102.9

−40-125

115.22

0-85

60.97

0-100

80.57

−40-100

70.07

0-85

94.68

−40-100

137.68

0-85

90.72

−40-100

108.87

−40-125

110.32

0-85

117.23

−40-100

208.587

−55-125

204.467

0-85

67.6

−40-100

67.987

7A35T (Xilinx Artix-7)

EP4CGX30 (Altera Cyclone IV)

10M40 (Altera MAX10)

M2GL050 (Microsemi IGLOO2)

LFE5UM-45 (Lattice ECP5)

33.28

56.34

628

500

450

400

370

125

Source: Data sheets and official price quotes from Xilinx Corp., Altera Corp., Microsemi Corp., and Lattice Corp.

IEEE Electrific ation Magazine / D EC EM BE R 2 0 1 7

Table of Contents for the Digital Edition of IEEE Electrification Magazine - December 2017