FPGA programming technology and

Interconnect architecture

Dr. D. V. Kamath
Professor, Department of E&C Engg.,
Manipal Institute of Technology, Manipal

1
 Programming technology for FPGAs

The logic cells within an FPGA are configured using a programming

technology. There are two classes of programming technology :
[i] OTP(One time programmable) [ii] Reprogrammable
Different FPGAs use different programmable (switching) elements :
[i] Antifuse in ACTEL FPGA
[ii] Static RAM cell in Xilinx FPGA
[iii] EPROM/ EEPROM in Altera CPLD
Programmable interconnect (PI) is used to connect any two logic
cells.
Different FPGAs use different interconnect architectures
[i] Segmented Channel routing in ACTEL FPGA
[ii] LCA interconnect architecture in Xilinx FPGA
2
 Antifuse technology for ACTEL FPGAs

 An antifuse is the opposite of a regular fuse

 An antifuse is a normally open(N.O) type of switch. A programming
current of 5 mA through it closes the switch.
e.g. In a poly-diffusion antifuse, the high current density causes a
large power dissipation in a small area, which melts a thin insulating
dielectric between polysilicon and diffusion electrodes and forms a
thin (about 20 nm in diameter), permanent, and resistive silicon.

 Antifuses separate interconnect wires on FPGA and programmer

blows an antifuse to make a permanent connection. This
programming process cannot be reversed.

3
 Programming technology for ACTEL FPGAs

Actel’s Antifuse
 Two types of antifuses are used :
[i] Poly-diffusion antifuse (Actel) [ii] Metal-metal antifuse (Quick
Logic)

 Actel calls its antifuse as Programmable Low Impedance Circuit

Element (PLICE)

 Advantage: small area overhead (size of the antifuse switching

element is very small in comparison with size of sRAM cell)

 Disadvantage : OTP
4
 ACTEL antifuse

2λ
A cross section of Actel antifuse
 The ONO (oxide-nitride-oxide) dielectric layer having thickness less
than 10 nm is deposited between conducting polysilicon and
diffusion layers.
 ONO dielectric is a combination of silicon dioxide and silicon nitride 5
 ACTEL antifuse

Actel antifuse (a) A cross section of a poly-diffusion antifuse (b) A

simplified drawing. The ONO (oxide–nitride–oxide) dielectric is less
than 10 nm thick, so this diagram is not to scale. (c) From above, an
antifuse is approximately the same size as a contact.

6
 ACTEL ‘s PLICE

• The size of an antifuse is limited by the resolution of the

lithography equipment used to makes ICs.

 The fabrication process and the programming current

control the average resistance of a blown antifuse.

 For programming current of 5 mA, antifuse resistance is

of the order of about 500 Ω. For 15 mA current, antifuse
resistance is about 100 Ω.

 The number of antifuse switches used in an ACTEL FPGA

is large. Eg. An Actel 1010 contains 1,12,000 antifuses. 7
 Metal-metal antifuse (Quick Logic)

Cross section of a Metal-metal antifuse

 Uses a two-level metal process
 The conductive link is an alloy of tungsten, titanium and silicon
with a bulk resistance of 500 μΩcm 8
 Metal-metal antifuse

Advantages of metal-metal antifuse over poly-diffusion antifuse

(i) Connections form antifuses to wiring layers (metal) are direct.

Hence, associated parasitic capacitance is less.
(ii) Because of direct connection, area requirement is also reduced.

(i) Also due to direct connection, it is easier to use larger

programming current which results in reduced antifuse
resistance.
 For Quick logic antifuse resistance is given by
R ≈ 0.8/ I Ω, where I in mA
 Resistance of metal-metal antifuse is around 80 Ω for a
programming current of 15 mA (for quick logic).
9
 Programming technology for Xilinx FPGAs

Xilinx sRAM configuration cell

 Consists of two cross-coupled inverters

 Uses a standard CMOS process
 The outputs of configuration cell are used to drive the gates of other
transistors (i.e., pass transistors or transmission gates) to make or
break the connection 10
 Xilinx sRAM configuration cell

Advantages:
 Designers can reuse chips
 Easy to upgrade the circuit design

Disadvantages:
 Volatile and power supply should be kept applied (Alternatively,
the configuration data can be downloaded from a PROM or Flash
memory) every time the system is turned ON)
 Larger in size
11
 sRAM cell

Word line

5V

 SRAM cells are found in Xilinx FPGA

 Allow for efficient implementation
of memory
BIT BIT
Line Line

Six transistor SRAM cell

12
 EPROM cell

Double-Poly EPROM Cell

13
 EPROM

 Altera MAX CPLDs and most of the PLDs use UV-erasable EPROM cells
as their programming technology.
 Size of the EPROM cell is almost same as that of an antifuse.
 EPROM technology is sometimes called floating-gate avalanche MOS
(FGMOS or FAMOS ).
 An unprogrammed n-channel EPROM transistor (having normal
threshold voltage) acts as ON switch for normal operating voltage.
 A programmed n-channel EPROM transistor (with a high threshold
voltage) acts as OFF. 14
 Double poly EPROM memory cell

 Applying programming voltage Vpp > 12 V to the drain of n-

channel EPROM transistor programs the EPROM cell.
 A High electric field causes electrons flowing toward the drain to
move so fast they jump across the insulating gate oxide, where
they are trapped on the bottom of the floating gate (gate1).
 These energetic electrons are hot and this effect is known as hot-
electron injection or avalanche injection . 15
 Double poly EPROM memory cell

 Electrons trapped on floating gate raise the threshold voltage of n-

channel EPROM transistor.

 Once programmed, an n-channel EPROM device remains OFF for

normal operating voltages (i.e., even with VDD applied to the top
gate).
16
 Double poly EPROM memory cell

 UV light is used to erase the EPROM cell. UV light provides

enough energy for the electrons stuck on gate1 to “jump” back to the
bulk, allowing the transistor to operate normally.

17
 EEPROM

 Programming an EEPROM transistor is similar to programming

an UV-erasable EPROM transistor, but the erase mechanism is
different.

 In an EEPROM transistor an electric field is also used to remove

electrons from the floating gate of a programmed transistor.

 This is faster than using a UV lamp and the chip does not have to
be removed from the system.

18
 Actel ACT family interconnect

 The channel routing uses dedicated rectangular areas of fixed size

within the chip called wiring channels or just channels.
 The horizontal channels run across the chip in the horizontal direction.
 The vertical channels run over the top of the basic logic cells in the
vertical direction .
 Within the horizontal or vertical channels wires run horizontally or
vertically, respectively, within tracks.
 The channel capacity refers to the number of tracks it contains. Each
track holds one wire.
 In a channeled gate array the designer decides the location and length
of the interconnect within a channel. But, in an FPGA the interconnect
is fixed at the time of manufacture.
 Each Logic Module is provided with the input stubs and output stubs.
19
Interconnect architecture used in an Actel ACT family FPGA
20
 Segmented channel routing

 To allow programming of the interconnect, Actel divides the fixed

interconnect wires within each channel into wires of various
lengths called wire segments.

 This is referred as segmented channel routing, a type of channel

routing.

 Antifuses join the wire segments. The designer then programs

the interconnections by blowing antifuses and making
connections between wire segments; unwanted connections are
left unprogrammed.

 A statistical analysis of many different layouts determines the

optimum number and the lengths of the wire segments.
21
 22
Detailed view of the channel and the connections to each Logic Module
 ACT1 interconnection architecture

Horizontal tracks

 The ACT1 interconnection architecture uses a total of 25

tracks per channel ( 22 horizontal tracks per channel for
signal routing + 3 tracks dedicated to VDD, GND, and the
global clock (GCLK))

 Horizontal segments vary in length from 4 columns of

Logic Modules to the entire row of modules (Actel calls
these long segments as long lines).

23
 ACT1 interconnection architecture
Vertical tracks
 Each ACT1 logic module has 8 inputs ( 4 input stubs on top and 4 on
bottom)

 8 vertical tracks per LM are available for inputs (4 from the LM above
the channel and 4 from the LM below).These connections are the input
stubs.

 The single LM output connects to a vertical track that extends across

the 2 channels above the module and across the 2 channels below the
module (output stub). Since this is a dedicated connection, no antifuse
is needed.

 Thus module outputs use 4 vertical tracks per module (counting 2 tracks
from the modules below, and 2 tracks from the modules above each
channel).
24
 ACT1 interconnection architecture

Vertical tracks

 One vertical track per column is a long vertical track (

LVT ) that spans the entire height of the chip (Actel
1020 contains some segmented LVTs).

 There are thus a total of 13 vertical tracks per column

in the ACT 1 architecture (8 for inputs, 4 for outputs,
and 1 for an LVT).

25
 ACT1 interconnection architecture

 The ACT 1 devices are very nearly fully populated (an antifuse at
every horizontal and vertical interconnect intersection).

 If the Logic Module at the end of a net is less than two rows away
from the driver module, a connection requires 2 antifuses, 1
vertical track, and 2 horizontal segments.

 If the modules are more than two rows apart, a connection

between them will require a long vertical track (LVT) together with
another vertical track (the output stub) and two horizontal tracks.
To connect these tracks will require a total of 4 antifuses in series
and this will add delay due to the resistance of the antifuses.

26
 Xilinx LCA
 The vertical lines and horizontal lines run between CLBs.
 The general-purpose interconnect joins switch boxes (also known
as magic boxes or switching matrices).
 The long lines run across the entire chip. It is possible to form
internal buses using long lines and the three-state buffers that are
next to each CLB.
 The direct connections (not used on the XC4000) bypass the
switch matrices and directly connect adjacent CLBs.
 The Programmable Interconnection Points (PIP’s) are
programmable pass transistors that connect the CLB inputs and
outputs to the routing network.
 The bidirectional (BIDI) interconnect buffers restore the logic level
and logic strength on long interconnect paths.
27
 Xilinx LCA

Xilinx LCA interconnect. (a) The LCA architecture (notice the matrix element size is
larger than a CLB). (b) A simplified representation of the interconnect resources.
Each of the lines is a bus. 28
Components of interconnect delay in a Xilinx LCA array. (a) A portion of the interconnect
around the CLBs. (b) A switching matrix. (c) A detailed view inside the switching matrix
showing the passtransistor arrangement. (d) The equivalent circuit for the connection
between nets 6 and 20 using the matrix. (e) A view of the interconnect at a Programmable
Interconnection Point (PIP). (f) and (g) The equivalent schematic of a PIP connection. 29
 Altera MAX 5000 and 7000 Interconnect scheme

A simplified block diagram of the Altera MAX interconnect scheme. (a) The PIA
(Programmable Interconnect Array) is deterministic— delay is independent of the
path length. (b) Each LAB (Logic Array Block) contains a programmable AND array. 30
 Altera MAX Interconnect scheme

• Altera MAX 5000/7000 devices use a Programmable

Interconnect Array ( PIA ).
• The PIA is a cross-point switch for logic signals traveling
between LABs.
• The advantages of this architecture is it uses a fixed
number of connections so the routing delay is also
fixed.
• Simpler and regular structure in nature that improved
speed of the placement and routing software.
• The delay between LAB1 and LAB2 is the same as the
delay between LAB1 and LAB6
31
 Contact

• [email protected]
• [email protected]

32