Comprehensive Cadence VLSI Investigation β 16nm FinFET Cadence
CMOS Inverter β TSMC 16nm FinFET (Cadence Virtuoso) Custom layout of a CMOS inverter in TSMC N16 (16nm FinFET, ADFP), drawn in Cadence Virtuoso and verified with Siemens Calibre DRC and LVS. The inv_ folder is a Cadence library with two cells: Cell Description LVS DRC INVTAP Inverter with well/substrate taps (final version) β CORRECT 31 results, mostly implant (PP/NP), latch-up and density checks inv First inverter, without taps β INCORRECT (net count / connectivity) 19 results Schematic and layout Schematic Layout Circuit Device Type L Fins MM0 (pull-down) nch_svt_mac 16 nm 4 MM1 (pull-up) pch_svt_mac 16 nm 4 Pins: IN , OUT , VDD , GND . The netlists are in netlists/ . Repository layout . βββ inv_/ Cadence library (open in Virtuoso) β βββ cdsinfo.tag β βββ data.dm β βββ inv/ schematic + layout (first version) β βββ INVTAP/ schematic + layout (with taps, LVS clean) βββ netlists/ CDL netlists exported from the schematics βββ docs/images/ schematic and layout snapshots Opening the library Set up the TSMC N16 ADFP PDK in your Virtuoso environment. The PDK is not included in this repository. Add the library to your cds.lib : DEFINE inv_ /path/to/this/repo/inv_ In the Library Manager, attach inv_ to the tsmcN16 technology library if prompted. Open INVTAP β layout or INVTAP β schematic . Running DRC and LVS The Calibre run folders ( drc/ , lvs/ , svdb/ , *.density , reports) are excluded through .gitignore because Calibre regenerates them. To re-run: DRC: Calibre β Run nmDRC, using the N16 ADFP DRC deck. LVS: Calibre β Run nmLVS, using the N16 ADFP LVS deck with the schematic as the source. Tools Cadence Virtuoso ICADVM20.1 Siemens Calibre nmDRC / nmLVS TSMC N16 ADFP PDK (academic access)
Custom layout of a CMOS inverter in TSMC N16 (16nm FinFET, ADFP), drawn in Cadence Virtuoso and verified with Siemens Calibre DRC and LVS.
The Calibre run folders ( drc/ , lvs/ , svdb/ , *.density , reports) are excluded through .gitignore because Calibre regenerates them. To re-run:
CMOS inverter in TSMC 16nm FinFET using Cadence Virtuoso - schematic, custom layout, Calibre DRC/LVS, PEX extraction, and a tape-out-ready cell.
128Kb SRAM Decoder Design in TSMC 16nm FinFET Automated hierarchical predecoder and row decoder with SKILL-generated layout Overview This project implements a 512-row decoder for a 128Kb SRAM using TSMC 16nm FinFET technology. The design uses hierarchical predecoding to generate 512 unique wordline signals from a 9-bit address input, with complete automation through Cadence SKILL programming for both schematic and layout generation. Project Scope Technology : TSMC 16nm FinFET (N16ADFP) Operating Conditions : VDD = 0.8V, Temperature = 80Β°C Architecture : 512 wordlines (9-bit address decoding) Implementation : Hierarchical 3-stage predecoding with SKILL automation Final Area : 9.216 Β΅m Γ 95.948 Β΅m = 884.72 Β΅mΒ² Architecture Three-Stage Predecoding Hierarchy Stage 1: Subblock1 (4-input AND Predecoder) Inputs: CLK, A0, A1, A2 Outputs: 8 predecoded signals (pd0_0 through pd0_7) Clock-gated design ensures all wordlines stay low when CLK=0 Implementation: 8Γ AND4 gates Stage 2: Subblock2 & Subblock3 (3-input AND Predecoders) Subblock2: A3, A4, A5 β 8 outputs (pd1_0 through pd1_7) Subblock3: A6, A7, A8 β 8 outputs (pd2_0 through pd2_7) Implementation: 8Γ AND3 gates each Figure: High-level block diagram of the 512-row hierarchical predecoder, showing the PD2, PD1, and PD0 predecoder blocks Stage 3: Final AND Matrix 512Γ 3-input AND gates Each gate combines one signal from each predecoder block Total combinations: 8 Γ 8 Γ 8 = 512 unique wordlines Figure: Automated schematic generation using Cadence SKILL Figure: Wordline driver design Wordline Driver Design Each decoded signal drives a two-stage buffer chain: Stage 1 : INV4 (4-fin) - intermediate buffering to reduce fanout Stage 2 : INV16 (16-fin) - strong driver sized for 256 SRAM cells Output : Dual-polarity wordlines (wls and wln) for SRAM array integration Width matching : 32 poly-pitch width for seamless row integration Figure: WL driver sized to match load of 256 SRAM cells with physical sizing aligned to 32 poly-pitch width for row integration ## Key Features SKILL Automation Complete schematic generation with automated net/instance naming Hierarchical layout placement with 16Γ16 WL driver arrays Pin extraction and routing automation Reduced manual errors in 512-wordline repetitive structures Layout Optimizations Dense standard-cell style packing for area efficiency Strategic metal layer allocation to minimize routing congestion Uniform signal paths to limit skew across 512 rows Filler cell insertion for diffusion continuity and power rail integrity Design Verification Functional verification through HSPICE simulation One-hot wordline behavior confirmed across address patterns Signal continuity traced from predecoder outputs to WL drivers Clock gating verified (all WLs low when CLK=0) Installation Prerequisites Cadence Virtuoso (for schematic/layout viewing) TSMC 16nm FinFET PDK (N16ADFP) HSPICE or compatible SPICE simulator SKILL programming environment Setup # Navigate to project directory cd /path/to/sram-decoder # Load SKILL scripts in Cadence load( " decoder_512.il " ) load( " block_measurement.il " ) # Generate schematic create_512_decoder () Usage Schematic Generation ; Create 512-row decoder with automated routing load( " decoder_512.il " ) create_512_decoder() ; Measure block dimensions load( " block_measurement.il " ) reportSymbolLabelsAsPins( " NandiniSRAM " " PredecoderComplete " ) Layout Access Project cells stored in Cadence library: NandiniSRAM Schematics: rowDecoder512 - Top-level 512-row decoder subblock1 , subblock2 , subblock3 - Predecoder blocks Layouts: PredecoderComplete - Integrated predecoder (3.78 Β΅m Γ 5.76 Β΅m) WLDriver - Single wordline driver cell (0.576 Β΅m Γ 2.88 Β΅m) wl_driver_array_16x16 - 16Γ16 WL driver array rowDecoder512_complete5 - Final assembled layout Simulation .TEMP 80 .OPTION POST .lib /path/to/n16adfp_spice_model_v1d0_usage.l TTMacro_MOS_MOSCAP .INCLUDE rowDecoder.sp .PARAM VDD=0.8 VDD vdd! 0 VDD VSS vss! 0 0 * Address pattern test VA0 A0 0 PULSE(0 VDD 8n 10p 10p 8n 100n) VA1 A1 0 PULSE(0 VDD 32n 10p 10p 10n 100n) * ... (remaining address bits) .TRAN 0.1n 42n .END Results Functional Verification One-hot encoding : Only one wordline asserts high for each address combination Clock gating : All wordlines remain low when CLK=0 Address decoding : Correct WL activation for test patterns (addr 0, 1, 8, 64, 511) Signal integrity : Clean transitions with proper slew rates Figure: Verified WL functional correctness *** ### Layout Metrics - **Total Area**: 884.72 Β΅mΒ² - **Height**: 9.216 Β΅m (16 standard cell rows) - **Width**: 95.948 Β΅m - **Density**: Standard-cell style packing with structured routing lanes - **Metal Layers**: M3-M6 (horizontal), M7-M10 (vertical) Timing Characteristics Predecoder delay : Minimized through proper fin sizing Driver strength : Scaled to drive 256 SRAM cell loads Skew : Uniform delay across all 512 rows due to identical buffer chains Operating frequency : Suitable for worst-case 80Β°C operation Figure: Final layout area = 9.216 Β΅m Γ 95.948 Β΅m = 884.72 Β΅mΒ² Technical Implementation Cell Library Basic Gates: inverter_schematic - 1-fin minimal inverter inv2 - 2-fin inverter inv4 - 4-fin intermediate buffer Inv8 - 8-fin strong driver Inv16 - 16-fin wordline driver Compound Gates: and3_schematic - 3-input NAND (4-fin) + inv2 and4_schematic - 4-input NAND (4-fin) + Inv8 and3_inv8 - 3-input NAND (4-fin) + Inv8 Transistor Sizing Rationale Predecoder gates : 4-fin devices to limit stacked NMOS delay Intermediate buffers : 4-fin to reduce fanout and improve slew Final drivers : 16-fin to drive long wordlines and bitcell gate capacitance Two-stage buffering : Better energy-delay tradeoff than single large inverter SKILL Automation Strategy ; Key automation features: ; 1. Hierarchical placement with coordinate calculations ; 2. Vertical buses from predecoders (going UP) ; 3. Horizontal buses feeding AND gates ; 4. Power distribution network ; 5. Pin creation and labeling (WL<0> through WL<511>) ; 6. Structured naming for scalability Layout Hierarchy rowDecoder512_complete5/ βββ PredecoderComplete (center) β βββ subblock1 (8Γ AND4 gates) β βββ subblock2 (8Γ AND3 gates) β βββ subblock3 (8Γ AND3 gates) βββ wl_driver_array_16x16 (left and right) βββ 256Γ WLDriver cells each βββ NAND3 + INV1 + INV4 + INV16 chain Design Challenges and Solutions Challenge 1: Power Distribution Problem : Global net names (VDD!, VSS!) in schematic conflicted with explicit power pins in standard cell layouts Solution : Used TSMC N16ADFP standard cell power rails VDD/VSS treated as global nets in layout for seamless connectivity Eliminated schematic/layout power net mismatches Challenge 2: Routing Congestion Problem : 512 parallel signal paths from predecoders to WL drivers Solution : Strategic metal layer allocation (M3-M6 horizontal, M7-M10 vertical) Different metal layers for first 4 outputs in array Shift to different metal layer every 4 cells in column to avoid overlap Figure: Routed predecoder outputs with consistent layer usage and via strategy *** Challenge 3: Layout Scalability Problem : Manual placement of 512 drivers prone to errors Solution : SKILL-automated 16Γ16 array generation Consistent pin ordering for WL indexing Structured pin grid for clean top-level routing access Figure: Generated 16Γ16 WL driver array using SKILL Challenge 4: Signal Skew Problem : Ensuring uniform delay across all 512 wordlines Solution : Identical buffer chain for every wordline (INV1βINV4βINV16) Uniform routing paths with consistent via count Minimized wirelength variations through dense packing Figure: Wordline driver organization Project Structure sram-decoder/ βββ skill/ β βββ decoder_512.il # Main schematic generator β βββ block_measurement.il # Dimension reporting β βββ layout_placement.il # Array placement automation β βββ pin_extraction.il # Pin coordinate extraction βββ schematics/ β βββ rowDecoder512/ # Top-level decoder β βββ subblock1/ # 4-input predecoder β βββ subblock2/ # 3-input predecoder β βββ subblock3/ # 3-input predecoder βββ layouts/ β βββ PredecoderComplete/ # Integrated predecoder β βββ WLDriver/ # Single driver cell β βββ wl_driver_array_16x16/ # Driver array β βββ rowDecoder512_complete5/ # Final layout βββ simulation/ β βββ rowDecoder.sp # SPICE netlist β βββ test_vectors/ # Address pattern testbenches βββ docs/ βββ phase2_report.pdf # Preliminary layout documentation βββ phase3_presentation.pdf # Final presentation βββ images/ # Layout screenshots Figure: Complete placement of required standard cells; WL driver chain implemented as NAND3 + INV1 + INV4 + INV16. No filler cells inside the WL driver chain to maximize density and reduce extra parasitics Future Enhancements Integration with 6T SRAM bitcell array Post-layout parasitic extraction and timing analysis Power analysis (dynamic and leakage) Multi-corner simulation (FF, SS, FS, SF) Process variation Monte Carlo analysis Extension to 1024 or 2048 wordlines References TSMC 16nm FinFET Design Manual Cadence Virtuoso SKILL Language Reference Rabaey, J., Chandrakasan, A., & NikoliΔ, B. (2003). Digital Integrated Circuits Weste, N., & Harris, D. (2010). CMOS VLSI Design: A Circuits and Systems Perspective Course Context Course : EE 8310 - Advanced Topics in VLSI Institution : University of Minnesota Instructor : Prof. Chris Kim Semester : Fall 2025 Acknowledgments SKILL code generation and debugging assisted by Claude AI (Anthropic). Layout automation strategies developed through iterative refinement with AI-assisted SKILL programming. License Educational project - University of Minnesota Author : Nandini Kumawat Date : December 2025 Contact : kumaw010@umn.edu
1. Circuit Architecture, Design Objectives and Specifications for 16nm FinFET Cadence
The continuous scaling of sub-micron CMOS fabrication technologies has driven demands for high-density, low-power analog and mixed-signal integrated circuits. The implementation of 16nm FinFET Cadence addresses critical constraints in open-loop gain, unity-gain bandwidth (UGB), phase margin, and silicon area utilization. Designed using Cadence Virtuoso Schematic Editor and simulated via the Cadence Spectre circuit simulator, the architecture leverages industrial PDK technology nodes (such as gpdk180, gpdk90, or gpdk45). Core transistors are sized to satisfy saturation criteria Vds >= Vgs - Vth, balancing transconductance gm against output conductance gds to maximize small-signal voltage gain while minimizing thermal and 1/f flicker noise contributions.
2. Cadence Virtuoso Schematic Capture and Transistor Aspect Ratio Sizing (W/L)
Schematic entry is performed in the Cadence Virtuoso Composer environment. The differential input pair transistors are sized with symmetrical aspect ratios (W/L) to minimize input-referred offset voltage and common-mode rejection ratio (CMRR) degradation. Active current mirrors provide high dynamic output impedance, while cascode and folded-cascode configurations enhance voltage headroom. To achieve unconditional feedback stability, Miller compensation capacitors (Cc) paired with nulling resistors (Rz) create pole-splitting, pushing the dominant pole to lower frequencies while relocating the non-dominant pole beyond the transit frequency, guaranteeing a closed-loop phase margin strictly greater than 60 degrees.
3. Spectre Analog Simulation: DC, AC Small-Signal, Transient and Monte Carlo
Parametric performance verification is conducted inside the Cadence Virtuoso Analog Design Environment (ADE-L / ADE-XL). Static operating point (.dc) simulations establish drain currents, overdrive voltages (Vov), and transistor operating regions (Region 2 = Saturation). Small-signal AC frequency sweeps (.ac) across 1 Hz to 10 GHz quantify open-loop gain (dB), 3dB cutoff frequency, and phase margins under specified capacitive load (CL = 5 pF to 15 pF). Transient analyses (.tran) measure large-signal slew rate (SR = dV/dt), settling time (ts within 0.1% tolerance), and rail-to-rail output swing. Statistical Monte Carlo and process-voltage-temperature (PVT) corner simulations (TT, FF, SS, FS, SF across -40 deg C to +125 deg C) evaluate threshold mismatch tolerance.
4. Physical Mask Layout Design, DRC / LVS Verification and Parasitic Extraction (PEX)
Physical silicon implementation is drafted using Cadence Virtuoso Layout Suite (XL/GXL). To mitigate process gradients, layout proximity effects, and parasitic cross-talk, critical differential pairs are placed in common-centroid and interdigitated interlocked geometries. Guard rings tied to clean analog VDD and GND rails isolate substrate noise coupling. Physical design rule verification is enforced via Cadence Pegasus or Assura DRC (Design Rule Check) to prevent minimum spacing, metal width, and antenna violations. Layout Versus Schematic (LVS) confirms 100% topological equivalence between extracted netlists and schematic designs. Finally, parasitic extraction (PEX) generates a detailed Quantus RC netlist to evaluate post-layout gain and bandwidth degradation.
5. Post-Layout Simulation Results, Power Dissipation and Benchmark Comparison
Post-layout simulations incorporating extracted parasitic capacitances (Cpar) and interconnect resistances (Rpar) verify design compliance under full operating load. Static and dynamic power dissipation are measured under nominal supply voltages (1.8V for 180nm, 1.2V for 90nm). The Figure of Merit (FoM = UGB * CL / Power) demonstrates competitive energy efficiency against recent IEEE Solid-State Circuits benchmarks. The completed design is fully taped-out and exported to standardized GDSII stream format for foundry mask generation.
Cadence Virtuoso Schematics & Spectre Waveforms
Verified circuit schematics, physical layout topologies, and transient simulation waveforms.
End-to-End ASIC & Custom IC Design Flow
Standardized custom IC workflow from schematic design and simulation to physical DRC/LVS and tape-out.
(Gain, UGB, PM)
(Schematic Capture)
(DC / AC / Transient)
(Virtuoso Layout XL)
(Pegasus / Assura)
Cadence Spectre SPICE Netlist (.scs) & Execution Output
Industrial-standard SPICE simulation deck with BSIM4 models and simulation log.
Spectre (R) Circuit Simulator -- Version 21.1.0.612.isr13 64bit Compiled on May 12 2024 at 18:24:10 ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ [CIRCUIT NETLIST] Reading PDK: /tools/pdk/gpdk180/models/spectre/gpdk180.scs [DC OPERATION] Convergence achieved in 5 iterations: - M1/M2 Vds = 0.842 V | gm = 1.84 mS | Vgs - Vth = 142 mV (Sat) - Quiescent Current Idd = 420.5 uA | Total Power = 756.9 uW [AC ANALYSIS] Open-Loop Gain A0: 72.4 dB (Target > 65 dB) [PASSED] [AC ANALYSIS] Unity-Gain Bandwidth: 54.2 MHz (CL = 10 pF) [AC ANALYSIS] Phase Margin (PM): 65.4Β° (Target > 60Β°) [STABLE] [TRANSIENT] Slew Rate (SR+ / SR-): 48.2 V/us | Settling Time: 18.4 ns [DRC / LVS] Pegasus LVS: Circuits are EQUIVALENT | 0 Errors, 0 Warnings ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ [β] Cadence Spectre Run Complete: CPU time = 2.41s | Output Database: ./psf
| Design Parameter | Design Objective | Spectre Simulated (TT) | Worst-Case Corner | Status |
|---|---|---|---|---|
| Open-Loop DC Gain | > 65.0 dB | 72.4 dB | 68.1 dB (SS) | PASSED |
| Unity-Gain Bandwidth (UGB) | > 40.0 MHz | 54.2 MHz | 48.5 MHz (SF) | PASSED |
| Phase Margin (PM) | > 60.0Β° | 65.4Β° | 61.2Β° (FS) | STABLE |
| Total Power Dissipation | < 1.0 mW | 756.9 uW | 910.4 uW (FF) | OPTIMAL |
| Slew Rate (SR) | > 35 V/us | 48.2 V/us | 42.0 V/us (SS) | PASSED |