Atmel CPLD Reference Designs Prove Logic Doubling™ Works White Paper March 1, 2001 R is the registered trademark of Atmel Corporation, 2325 Orchard Parkway, San Jose, CA 95131 Rev. 2310A–03/01 Summary The first section of this paper describes the development of limitations in I/O connectivity and logic reusability in Complex Programmable Logic Devices (CPLDs) over their 20year history. The second section describes the architectural enhancements (Logic Doubling) in Atmel’s ATF15xx Families of CPLDs and how they address these limitations. The third section describes several reference designs that illustrate both the limitations and how Logic Doubling overcomes them and in each case compares Atmel’s ATF15xx Families performance to that of typical industry-standard CPLDs. The programmable logic designer is encouraged to download these reference designs and Atmel’s design and fitter software, repeat these experiments and replicate results. Using these examples and tools, the PLD designer can then apply Logic Doubling techniques to new product designs, obtaining the benefits of more features in a smaller, and possibly less expensive chip, or spare logic resources for future revisions and reduced risk of PCB re-spin. Logic Doubling Background The first PAL devices in the late 1970’s offered a single layer of simple logic: Inputs were routed to an AND/OR block and then to the outputs. PALs had a single, risingedge CLK pin and a single register OE pin. If more layers of logic were needed, more pins were required. Although extremely limited by current standards, PALs had two advantages: all I/O signals were available to all logic cells, and relatively little logic was wasted. Each new generation of programmable logic improved on the many other PAL shortcomings but not these two. Over the first decade, PALs evolved into “CMOS” SPLDs, and the 16V8, 20V8 and 22V10 became standard parts. These devices remained 100 percent connected between the I/O pins and their logic cells. In the following decade, as the logic cells grew in complexity, adding Product Term clocks, multiple OE terms, etc., more of the logic in each cell was potentially left unused in the finished designs. The metric “usable gates” (generally a fraction of about half of the total number of physical gates on the chip) came into use as better way to describe the amount of logic typically accessible for use in a finished design. However, as the number of macrocells in a single CPLD is increased, the required signal routing area and loading on these nodes also increases according to the square law. The resulting increased die size and speed penalty is simply too great, and so in larger devices all nodes cannot be fed into all macrocells. As 44-pin CPLDs emerged, new layers of hierarchy were added to their structures. Macrocells were grouped into blocks usually of 16. Output enable functions were added, but often at the logic block level of hierarchy, thus having limited flexibility. In defining their CPLD architecture, most manufacturers decided to sacrifice connectivity for minimized die size and maximized speed. Fan-in to the blocks became limited, logic utilization took another drop and routing flexibility, both within the macrocells and to the I/O pads, was compromised. The term “pin-locking” was introduced to describe the ability to preassign pins. Lack of pin-locking became a CPLD issue and was debated hotly by leading competitors. Over the last decade, the CPLD version of Moore’s law drove logic density higher. However, routing density, while improving, did not keep pace. The extra fuses and interconnect required to make the increasing amount of unusable logic more accessible would simply take up too much area on the chip, and rather than drive the already high cost of CPLDs even higher, the trend toward inaccessible logic in each macrocell continued. 2 White Paper 2310A–03/01 White Paper Likewise, CPLDs still carried the limitation that to have more than one logic layer, another whole macrocell was consumed. At Atmel, these trends were identified early, and our PLD architects took care to retain as much I/O connectivity and logic reusability as possible, introducing the 100 percent connected ATF1500 with 32 enhanced macrocells in 1996, followed by the ISP family members, ATF1502 with 32 macrocells, ATF1504 with 64 macrocells, and ATF1508 with 128 macrocells. Atmel’s commitment to efficient, flexible architecture has continued and this paper will describe the current state of our art. We have coined the term “Logic Doubling” to refer to our efficient, flexible CPLD architecture, now available with secondgeneration EDA, second-generation fitters and current products as well as the enhanced second generation silicon. 3 2310A–03/01 Logic Doubling Theory Readers already familiar with Logic Doubling concepts may wish to skip this section. Atmel has incorporated a number of features to the ATF15xx Family to address connectivity and reusability problems. • More cross-point MUXs available for input node fan-in. • Wider MUX channels into the Logic Blocks. • Dual, independent feedback paths for each macrocell. The buried and pin-driver feedback paths are split, so a register output may be buried while independently driving a combinatorial pin, or vice versa. Thus, the unused macrocell logic can be accessed and used. • EVERY macrocell may have separate Output Enable, a feature necessary for software control of data direction and often overlooked by other manufacturers who provide only a few OE terms. • Selectable Global Clock Polarity, either rising or falling edge. • Global RESET can combine (OR) with a local Product Term. Global Routing The 44-pin, 32-macrocell Atmel ATF1500 CPLD is the patriarch of the ATF15xx Family and provides 100 percent connectivity. The Global bus in the ATF1500 has all input and feedback signals available to all logic blocks, for a fan-in of 68, and the ATF1500 is thus ideal for designs needing maximum fan-in. Because the ATF1500 is 100 percent connected and all the logic cells look the same, most logic changes can be made without touching the pinouts. The 1500 is not immune to pin-locking problems, but it does avoid MUX induced pin shifts. (Atmel also offers the proprietary, 100 percent connected ATF2500, which has 48 registers with 17 product terms per macrocell in a 44-pin package.) Of course this 100 percent connected approach has scaling issues, and so the rest of the ATF15xx Family devices use the global/regional bus hierarchy for cross-point allocation. Even so, Atmel’s CPLDs maintain the highest connectivity of any CPLD Family. Logic Block Routing and Fan-In Number of Macrocells 32 32 64 128 Atmel Part Number ATF1500 ATF1502 ATF1504 ATF1508 Atmel Mux Structure 68 (No MUX) 40 x 5 Matrix x 2 40 x 9 Matrix x 4 40 x 27 x 8 Typical Mux Structure N/A 36 x 4 Matrix x 2 36 x 8 Matrix x 4 36 x 14 x 8 Because Atmel provides both wider fan-ins than typical, and more cross-points, the ATF15xx Family users have a higher Logic Ceiling (more headroom) for their designs. If you have ever moved up to a larger, costlier device just to get a bit more logic resource, you might have delayed, or avoided that move altogether with the ATF15xx Family. 4 White Paper 2310A–03/01 White Paper Macrocell Routing The following summarizes the Atmel Macrocell signal resources. Each Macrocell has two main input paths into the five AND type product terms: 16 Unipolar Foldback Terms 40 Bipolar Signals selected from global pool, by UIM Each Macrocell has additional inputs: 3 Global Clock choices 6 Output enable choices 1 Global Clear option 1 Cascade in, from adjacent MC Each Macrocell has four output paths: 1 Regional Foldback (16 within each 16-macrocell logic block) 1 Buried Feedback (global) 1 Pin Drive (global) 1 Cascade output, for sharing unused product terms to and adjacent, higher macrocell Compared to typical CPLDs, the ATF15xx Family provides extra Multiplexers within the macrocell for the register and combinatorial inputs and outputs. This extra routing allows: 1. Toggle flip-flop synthesis. 2. A transparent latch mode for the flip-flop, for ALE style BUS. 3. A buried D-type register, while the remaining combinatorial term drives the PIN. 4. A buried Latch-type register, where LE can be held at logic High, obtaining another combinatorial feedback, while the remaining combinatorial terms drive the PIN. 5. Fast input from the I/O pin. 5 2310A–03/01 Atmel Macrocell with Logic Doubling Clock Resources The Atmel ATF15xx Family provides three global clocks, with a choice of rising or falling edge, as well as a local product term clock. One of the global clocks is derived from an I/O pin, allowing a complex, shared clock. Typical CPLDs, from most other suppliers, provide two, with polarity fixed. I/O Control Resources Atmel’s Logic Doubling architecture also provides an extra Multiplexer, so every macrocell can have it’s own output enable. This allows fully soft control of Data Direction on every pin – surprisingly many CPLD families do not offer this. Fuse Resources Since fuses control the number of logic choices and paths, more fuses is an indication of more flexibility and usability. For CPLDs that have similar structures, the fuse count is proportionate to the number of resource usage alternatives the device affords, and can be considered a simple “IQ” indicator. 6 White Paper 2310A–03/01 White Paper Fitter Capabilities The second generation Atmel device fitters are designed to take optimal advantage of these Logic Doubling features. The fitters place and route the design onto a specific Atmel ATF15xx Family device and report results to the designer for review. Besides pinout, the fitters produce a detailed report file that details logic resource usage as well as spare resources remaining for each macrocell. These details allow the designer to see what is going on and make seamless informed design choices as the design progresses. Note: At the time of this writing, detailed fitter reports were still being enhanced. Atmel’s second generation Atmel ATF15xx Family device fitters support EDIF input (VHDL and Verilog®) as well as legacy PLA formats and produce SDF timing output files. They are fully integrated in Atmel’s ProChip Designer™ EDA tool suite with fitter controls to make it easy to pack more features into the chip or to leave additional logic resources for future revisions, reducing the chance of a PCB re-spin. Verification Capabilities The SDF output files, when combined with the ATF15xx Family VITAL models, allow the designer to run timing simulations of the final design, including the effects of placement and routing. If a timing problem is detected, the designer can correct it before hardware testing begins. The Accolade PeakVHDL™ simulator is fully integrated in Atmel’s ProChip Designer EDA tool suite, so this enhanced verification is easy to include in the design flow. 7 2310A–03/01 Logic Doubling Practice: Atmel Reference Designs So the reader can see how all this stacks up, Atmel has placed these reference designs on it’s web site (most CPLD vendors keep them secret) where the designer can easily access them and personally verify our claims. These reference designs are free, compiler ready, “full chip” designs based on common design problems. Using these reference designs, you can: • Quickly evaluate how much logic fits into a CPLD. • Change the template details to suit your system. • Use them as rapid language training examples. • Use them as IP, merging their elements with other Logic into a larger Atmel CPLD. • Anything else you can think of. Parallel I/O Expander (ATF8255N) Description The Parallel I/O expander function is a widely used tool for getting high-bandwidth I/O from microcontrollers and microprocessors. Legacy silicon is still widely used for this, partly because alternatives have come up short. Often these “Legacy” devices are not available in QFP, have limited BUS speeds, and have low-drive capabilities, but there have been few upgrades until now. Design Summary To implement a parallel I/O expander, with software control of Port Direction on every pin, a device needs at least 25 OE terms. This design will fit in all ATF15xx Family devices, creating a scalable 82C55 and fits in none of the EPM7XXX devices! Relative to the 146823, the CPLD has much higher pin drive and is MUCH faster. Enhancements This design merges the best features of 82C55 and 68230 in that it provides 24 I/O pins, and also individual DDR control on each pin. To keep the pin count down, a Multiplex memory interface is used. The DDR design allows software control of pin types, and suits fully SOFT system designs, where the system software configures the in/out mix, and reprogram of the I/O devices is not needed. Some Applications Areas KEYPAD Scan: A single ATF1502 could scan 144 keys, needing just 12 pull-up resistors, or 132 keys + 12 status LED’s multiplexed with the KeyScan, etc. CABLE LOOM TESTERS/PCB bed of nails testers: Looking for OPEN, and Illegal Shorts, and Valid connections. A high speed parallel interface is needed for this, as the required number of nodes goes as the square of the wires. For each Node Write, typically ALL others are READ. Fast Parallel Memory Interface: ALE.WRN.RDN.CEN0 – Standard C51 Multiplex BUS Memory interface DB0..7– Has Address when ALE is high, controlled by RDN, WRN otherwise 8 White Paper 2310A–03/01 White Paper Results 9 2310A–03/01 Comparative Analysis FIT1502 ATF8255N Fitter Report The parallel I/O expander simply does not fit in the typical CPLD, due to lack of OE terms. pin_num pin_name output_type feedback foldback cascade_out output_slew MC1 4 DB0 reg bAD0 -- -- slow MC2 5 Pa2 reg oePa2 NA -- slow MC3 6 Pa1 reg oePa1 NA -- slow MC4 7 Pa0 reg oePa0 NA -- slow MC5 8 Pb3 reg oePb3 NA -- slow MC6 9 Pb2 reg oePb2 NA -- slow MC7 11 Pb1 reg oePb1 NA -- slow MC8 12 Pb0 reg oePb0 NA -- slow MC9 13 Pc3 reg oePc3 NA -- slow MC10 14 Pc2 reg oePc2 NA -- slow MC11 16 Pc1 reg oePc1 NA -- slow MC12 17 Pc0 reg oePc0 NA -- slow MC13 18 Pa3 reg oePa3 NA -- slow MC14 19 DB1 reg GlOE FbFWrPbL -- slow MC15 20 DB2 reg bAD1 -- -- slow MC16 21 DB3 reg FbFWrPaL FbFWrPcL -- slow MC17 41 Pa4 reg oePa4 NA -- slow MC18 40 Pa5 reg oePa5 NA -- slow MC19 39 Pa6 reg oePa6 NA -- slow MC20 38 Pa7 reg oePa7 NA -- slow MC21 37 Pb4 reg oePb4 NA -- slow MC22 36 DB4 reg HnWrPc FbFWrPcU -- slow MC23 34 Pb6 reg oePb6 NA -- slow MC24 33 Pb7 reg oePb7 NA -- slow MC25 32 Pc4 reg oePc4 NA -- slow MC26 31 Pc5 reg oePc5 NA -- slow MC27 29 Pc6 reg oePc6 NA -- slow MC28 28 DB7 reg bAD2 -- -- slow MC29 27 Pc7 reg oePc7 NA -- slow MC30 26 DB6 reg HnWrPa FbFWrPaU -- slow MC31 25 DB5 reg HnWrPb FbFWrPbU -- slow MC32 24 Pb5 reg oePb5 NA -- slow MC0 2 ALE -- -- -- -- slow MC0 1 CEN0 -- -- -- -- slow MC0 44 RDN -- -- -- -- slow MC0 43 WRN -- -- -- -- slow Array Block Nodes/MCells I/O Pins Foldbacks Cascades - LC16 32/16(200%) 16/16(100%) 2/16(12%) 0 B: LC17 - LC32 32/16(200%) 16/16(100%) 3/16(18%) 0 A: LC1 Total dedicated input used:4/4 (100%) Total I/O pins used32/32 (100%) Logic Nodes+FB/MCells used69/32 (215%) Total Foldback logic used 5/32 (15%) Total cascade used 0 Total input pins 4 Total output pins 32 ---------------- 10 End fitter, Design FITS in ATF1502AS White Paper 2310A–03/01 White Paper Altera ATF8255N (All Devices) Fitter Summary ** ERROR SUMMARY ** Error: Project requires too many (25/10) Output Enable signals Serial I/O Expander and LED Driver (LED32ser) Description Serial I/O expanders like 4094/HC595 are widely used for low bandwidth I/O from microcontrollers and microprocessors. This example uses 100 percent of the CPLD I/O as output drive and so compares directly with these logic devices. More typically a CPLD design would also have some input support. The CPLD has pin swap and a higher drive than 4094/HC595 devices and the low static IDD of the ATF15xx Family “L” versions means all the package power ability can be used for LED drive. Many other CPLDs have static powers of hundreds of mW. Results Comparative Analysis Currently, this cannot fit into any other 32-macrocell device, one must double the macrocell count to obtain a fit. 11 2310A–03/01 FIT1502 LED32Ser Fitter Summary pin_num pin_name output_type feedback foldback cascade_out output_slew MC1 4 IO0 reg Shift0 -- -- slow MC2 5 IO1 reg Shift1 -- -- slow MC3 6 IO2 reg Shift2 -- -- slow MC4 7 IO3 reg Shift3 -- -- slow MC5 8 IO4 reg Shift4 -- -- slow MC6 9 IO5 reg Shift5 -- -- slow MC7 11 IO6 reg Shift6 -- -- slow MC8 12 IO7 reg Shift7 -- -- slow MC9 13 IO8 reg Shift8 -- -- slow MC10 14 IO9 reg Shift9 -- -- slow MC11 16 IO10 reg Shift10 -- -- slow MC12 17 IO11 reg Shift11 -- -- slow MC13 18 IO12 reg Shift12 -- -- slow MC14 19 IO13 reg Shift13 -- -- slow MC15 20 IO14 reg Shift14 -- -- slow MC16 21 IO15 reg Shift15 FbFollowL -- slow MC17 41 IO16 reg Shift16 FbFollowU -- slow MC18 40 IO17 reg Shift17 -- -- slow MC19 39 IO18 reg Shift18 -- -- slow MC20 38 IO19 reg Shift19 -- -- slow MC21 37 IO20 reg Shift20 -- -- slow MC22 36 IO21 reg Shift21 -- -- slow MC23 34 IO22 reg Shift22 -- -- slow MC24 33 IO23 reg Shift23 -- -- slow MC25 32 IO24 reg Shift24 -- -- slow MC26 31 IO25 reg Shift25 -- -- slow MC27 29 IO26 reg Shift26 -- -- slow MC28 28 IO27 reg Shift27 -- -- slow MC29 27 IO28 reg Shift28 -- -- slow MC30 26 IO29 reg Shift29 -- -- slow MC31 25 IO30 reg Shift30 -- -- slow MC32 24 IO31 reg Shift31 -- -- slow MC0 2 DATI -- -- -- -- slow MC0 1 CSEL -- -- -- -- slow MC0 44 BPlane -- -- -- -- slow MC0 43 CLK -- -- -- -- slow Array Block Nodes/MCells I/O Pins Foldbacks A: LC1 - LC16 32/16(200%) 16/16(100%) 1/16( 6%) 0 Cascades B: LC17 - LC32 32/16(200%) 16/16(100%) 1/16( 6%) 0 Total dedicated input used:4/4 (100%) Total I/O pins used32/32 (100%) Logic Nodes+FB/MCells used66/32 (206%) Total Foldback logic used 2/32 (6%) Total cascade used 0 Total input pins 4 Total output pins 32 ---------------- 12 End fitter, Design FITS in ATF1502AS White Paper 2310A–03/01 White Paper Altera (7032) Fitter Summary Led32ser ** ERROR SUMMARY ** Error: Project requires too many (64/32) logic cells Altera (7064) Fitter Summary Total dedicated input pins used: 4/4 (100%) Total I/O pins used: 32/32 (100%) Total logic cells used: 64/64 (100%) Total shareable expanders used: 0/64 ( 0%) Total Turbo logic cells used: 64/64 (100%) Total shareable expanders not available (n/a): 32/64 ( 50%) Average fan-in: Total fan-in: 3.00 192 Pulsewidth Modulator (PWM8x4) Description This design packs the maximum number of 8-bit resolution DACS, and PWM modulators, into a 44-pin PLCC 32-macrocell device. Pulse width modulation is often used for digital-to-analog conversion (with an output filter) and to drive Motors and Solenoids efficiently using switched mode. Using the Logic Doubling, an ATF1502 can swallow four 8-bit PWM generators and their Value Latches. The design has an 8-bit microcontroller BUS interface for write of the PWM values. The circuit can run at full Clock Speed and has special trap for SetPoint of 0, so the output remains DC low when stopped. If always having a waveform is important to the application, removing the trap can force 1/256 to both 0 and 1. Typical add-on features could be a FAST protect reset, if Motor Driving, or system RESET input, to define the PWM output at reset. Results 13 2310A–03/01 Comparative Analysis FIT1502 PWM8x4 Fitter Summary PWM8X4 cannot fit into a typical 32-macrocell device, roughly double the macrocell count is required. pin_num pin_name output_type feedback foldback cascade_out output_slew MC1 4 CEN reg PwCtr0 -- -- slow MC2 5 WRN reg PwCtr1 -- -- slow MC3 6 DB4 reg PwmA6 -- -- slow MC4 7 Adr1 reg PwCtr3 -- -- slow MC5 8 PwmD6 reg PwCtr4 -- -- slow MC6 9 PwmD7 reg PwCtr5 -- -- slow MC7 11 PwmC6 reg PwCtr6 XXL_257 -- slow MC8 12 PwmC7 reg PwCtr7 XXL_256 -- slow MC9 13 PwmAOUT reg -- XXL_255 -- slow MC10 14 PwmBOUT reg -- XXL_254 -- slow MC11 16 DB0 com NEqB0_4 NA -- slow MC12 17 DB1 com NEqA0_4 NA -- slow MC13 18 -- reg PwmB6 fbWrPwmC -- slow MC14 19 -- reg PwmB7 fbWrPwmD -- slow MC15 20 DB6 reg PwCtr2 fbZeroAN -- slow MC16 21 DB7 reg PwmA7 fbZeroBN -- slow MC17 41 PwmC0 reg PwmA0 WrPwmCN -- slow MC18 40 PwmC1 reg PwmA1 WrPwmDN -- slow MC19 39 PwmC2 reg PwmA2 -- -- slow MC20 38 PwmC3 reg PwmA3 -- -- slow MC21 37 DB2 com NEqC0_4 NA -- slow MC22 36 PwmC5 reg PwmA5 -- -- slow MC23 34 PwmC4 reg PwmA4 -- -- slow MC24 33 DB3 com NEqD0_4 NA -- slow MC25 32 PwmD0 reg PwmB0 XXL_253 -- slow MC26 31 PwmD1 reg PwmB1 XXL_252 -- slow MC27 29 PwmD2 reg PwmB2 XXL_251 -- slow MC28 28 PwmD3 reg PwmB3 XXL_250 -- slow MC29 27 PwmD4 reg PwmB4 fbWrPwmC -- slow MC30 26 PwmD5 reg PwmB5 fbWrPwmD -- slow MC31 25 PwmCOUT reg -- fbZeroCN -- slow MC32 24 PwmDOUT reg -- fbZeroDN -- slow MC0 2 DB5 -- -- -- -- slow MC0 1 Reset -- -- -- -- slow MC0 44 Adr0 -- -- -- -- slow MC0 43 CLK -- -- -- -- slow Array Block A: LC1- LC16 Nodes+FB/MCells I/O Pins 20/16(125%)14/16(87%) B: LC17- LC32 28/16(175%)16/16(100%) Foldbacks Cascades 8/16(50%)0 10/16(62%)0 Total dedicated input used:4/4 (100%) Total I/O pins used30/32 (93%) Logic Nodes+FB/MCells used66/32 (206%) Total Foldback logic used 18/32 (56%) Total cascade used 0 Total input pins 14 Total output pins 20 ---------------- 14 End fitter, Design FITS in ATF1502AS White Paper 2310A–03/01 White Paper Altera PWM8x4 (7032) Fitter Summary pwm8x4 Altera PWM8x4 (7064) Fitter Summary Total dedicated input pins used: ** ERROR SUMMARY ** Error: Project requires too many (48/32) logic cells 2/4 ( 50%) Total I/O pins used: 32/32 (100%) Total logic cells used: 48/64 ( 75%) 9/64 ( 14%) 48/64 ( 75%) Total shareable expanders not available (n/a): 8/64 ( 12%) Average fan-in: 7.39 Total shareable expanders used: Total Turbo logic cells used: Total fan-in: 355 Latch for Multiplex BUS (LatchSyn) Description This modular example demonstrates ways to synthesize a Transparent Latch, commonly used in Multiplex BUS system, so you can place a HC573/373 into a corner of your CPLD, and save PCB real estate. The macrocell of the ATF15xx family offers options on the method used to create a transparent latch: • Use the macrocell .L, LE support directly. • Use a make-before-break 2:1 MUX, using the PIN Feedback. • Use a .D register, like a -ve clock HC374/574. This file implements both and packs both into one macrocell as a demonstration. In a real system, one can choose to bury some address terms, or drive all 8 to Pins, and bury 8-D registers in the same macrocells. Besides the pins used, the HC573/373 can have minimal impact on the LOGIC used, and give a significant PCB area saving, and better security. FIT1502 LatchSyn Fitter Summary pin_num pin_name output_type feedback foldback cascade_out output_slew MC1 4 uDB0 reg -- -- -- slow MC2 5 uDB1 reg -- -- -- slow MC3 6 uDB2 reg -- -- -- slow MC4 7 uDB3 reg -- -- -- slow MC5 8 uDB4 reg -- -- -- slow MC6 9 uDB5 reg -- -- -- slow MC7 11 uDB6 reg -- -- -- slow MC8 12 uDB7 reg -- -- -- slow MC9 13 PinL0 reg StdL0 -- -- slow MC10 14 PinL1 reg StdL1 -- -- slow MC11 16 PinL2 reg StdL2 -- -- slow MC12 17 PinL3 reg StdL3 -- -- slow MC13 18 PinL4 reg StdL4 -- -- slow MC14 19 PinL5 reg StdL5 -- -- slow MC15 20 PinL6 reg StdL6 -- -- slow MC16 21 PinL7 reg StdL7 -- -- slow MC17 41 -- com PinLFollowN -- -- slow MC18 40 -- -- -- -- -- slow MC19 39 -- -- -- -- -- slow MC20 38 -- -- -- -- -- slow MC21 37 -- -- -- -- -- slow MC22 36 -- -- -- -- -- slow 15 2310A–03/01 MC23 34 -- -- -- -- -- slow MC24 33 -- -- -- -- -- slow MC25 32 DLat0 reg -- -- -- slow MC26 31 DLat1 reg -- -- -- slow MC27 29 DLat2 reg -- -- -- slow MC28 28 DLat3 reg -- -- -- slow MC29 27 DLat4 reg -- -- -- slow MC30 26 DLat5 reg -- -- -- slow MC31 25 DLat6 reg -- -- -- slow MC32 24 DLat7 reg -- -- -- slow MC0 2 -- -- -- -- -- slow MC0 1 -- -- -- -- -- slow MC0 44 RDN -- -- -- -- slow MC0 43 ALE -- -- -- -- slow ----------- End fitter, Design FITS in ATF1502AS Conclusion Atmel’s ATF15xx Family of CPLDs provides enhanced I/O connectivity and logic utilizability. Atmel’s Logic Doubling architecture combined with Atmel’s second-generation device fitters, extends CPLD place and route efficiency. For example with double independent buried feedback designers can pack more logic (particularly shifters and latches) into smaller CPLDs. This stretching of CPLD resources in some cases increases the nominal register/latch count to 200 percent or more. The more flexible routing and denser packing of logic enables the designer to use a smaller, often less costly device or leave spare room for later design revisions at no additional design effort. 16 White Paper 2310A–03/01 Atmel Headquarters Atmel Operations Corporate Headquarters Atmel Colorado Springs 2325 Orchard Parkway San Jose, CA 95131 TEL (408) 441-0311 FAX (408) 487-2600 Europe Atmel SarL Route des Arsenaux 41 Casa Postale 80 CH-1705 Fribourg Switzerland TEL (41) 26-426-5555 FAX (41) 26-426-5500 Asia Atmel Asia, Ltd. Room 1219 Chinachem Golden Plaza 77 Mody Road Tsimhatsui East Kowloon Hong Kong TEL (852) 2721-9778 FAX (852) 2722-1369 Japan 1150 E. Cheyenne Mtn. Blvd. Colorado Springs, CO 80906 TEL (719) 576-3300 FAX (719) 540-1759 Atmel Rousset Zone Industrielle 13106 Rousset Cedex France TEL (33) 4-4253-6000 FAX (33) 4-4253-6001 Atmel Smart Card ICs Scottish Enterprise Technology Park East Kilbride, Scotland G75 0QR TEL (44) 1355-357-000 FAX (44) 1355-242-743 Atmel Grenoble Avenue de Rochepleine BP 123 38521 Saint-Egreve Cedex France TEL (33) 4-7658-3000 FAX (33) 4-7658-3480 Atmel Japan K.K. 9F, Tonetsu Shinkawa Bldg. 1-24-8 Shinkawa Chuo-ku, Tokyo 104-0033 Japan TEL (81) 3-3523-3551 FAX (81) 3-3523-7581 Fax-on-Demand e-mail North America: 1-(800) 292-8635 [email protected] International: 1-(408) 441-0732 http://www.atmel.com Web Site BBS 1-(408) 436-4309 © Atmel Corporation 2001. Atmel Corporation makes no warranty for the use of its products, other than those expressly contained in the Company’s standard warranty which is detailed in Atmel’s Terms and Conditions located on the Company’s web site. The Company assumes no responsibility for any errors which may appear in this document, reserves the right to change devices or specifications detailed herein at any time without notice, and does not make any commitment to update the information contained herein. No licenses to patents or other intellectual property of Atmel are granted by the Company in connection with the sale of Atmel products, expressly or by implication. Atmel’s products are not authorized for use as critical components in life support devices or systems. Verilog is a trademark of Gateaway Design Automation Corporation. Accolade PeakVHDL is a trademark of Protel International Logic Doubling and ProChip Designer are trademarks of Atmel Corporation. Terms and product names in this document may be trademarks of others. Printed on recycled paper. 2310A–03/01/xM