AutoGineer
Generated banner for Matlab Model Rf
Generated logo for Matlab Model Rf

Matlab Model Rf

MathWorksOfficialCreated by MathWorks

  • Skill
  • Guidelines
  • Science and research
  • BSD-2-Clause
  • No issues found
Download ZIP

From the author

RF Toolbox and RF Blockset in MATLAB -- S-parameter I/O, network conversions (S/Z/Y/ABCD/T/H/G, mixed-mode), cascade/de-embedding, rfbudget analysis, circuit composition, matching networks, amplifier stability, mixer spurs, rational fitting, SI channels, baseband processing, Circuit Envelope simulation. Trigger: sparameters, Touchstone, .s2p, .s4p, rfplot, smithplot, rfparam, rfwrite, zparameters, yparameters, abcdparameters, s2sdd, cascadesparams, deembedsparams, rfbudget, noise figure, OIP3, IIP3, amplifier, modulator, nport, rffilter, attenuator, seriesRLC, shuntRLC, lcladder, txline, circuit, setports, clone, matchingnetwork, stabilityk, stabilitymu, powergain, gammams, gammaml, mixerIMT, OpenIF, rational, rationalfit, stepresp, txlineWRLGC, rf.Amplifier, rf.Mixer, rf.Filter, rf.Sparameter, rfsystem, RF Blockset.
— the description by MathWorks

AutoGineer summary

Written by AutoGineer from the skill’s own files, not by the author

This skill provides a structured approach to utilizing MATLAB's RF Toolbox and RF Blockset for end-to-end RF system development. It covers workflows ranging from measured S-parameter data validation and network conversions to complex circuit-level design, amplifier stability analysis, and time-domain Circuit Envelope simulation. The documentation emphasizes best practices for script-based workflows, proper element handling, and avoiding common pitfalls in RF parameter manipulation and system modeling.

What it does

  • Validates and processes measured S-parameter data from Touchstone files.
  • Performs network parameter conversions and mixed-mode analysis.
  • Enables RF signal chain design and cascaded budget analysis.
  • Supports circuit-level composition and impedance matching network design.
  • Facilitates amplifier stability analysis and mixer intermodulation planning.
  • Provides tools for rational fitting and time-domain system simulation.

When to use it

  • When working with measured S-parameter data for validation or de-embedding.
  • When designing RF signal chains and performing cascaded budget analysis.
  • When composing arbitrary RF circuits using node-based netlists.
  • When analyzing amplifier stability or mixer spur performance.
  • When performing time-domain signal integrity or Circuit Envelope simulations.

Advantages

  • Unified workflow for both passive and active RF device characterization.
  • Strict adherence to MATLAB identifier and object handling rules prevents silent errors.
  • Comprehensive support for modern visualization tools like tiledlayout and smithplot.
  • Clear separation between script-based design workflows and interactive analysis.
  • Detailed guidance on avoiding common errors in port ordering and parameter units.

Usage examples

“How do I load an S-parameter file and plot it on a Smith chart?”
The agent will guide you to use sparameters() to load the file, followed by smithplot() for visualization, ensuring you handle reference impedance correctly via newref() if needed.
“Calculate the cascaded gain and noise figure for my RF chain.”
The agent will help you define elements using rfbudget, ensuring you set the SignalBandwidth and use clone() for reused elements, then run the budget analysis.
“How do I analyze the stability of an amplifier?”
The agent will instruct you to use stabilityk and stabilitymu, and calculate power gain using powergain with the correct linear-to-dB conversion.
“I need to export my S-parameters to a new file.”
The agent will provide the rfwrite command, specifically reminding you to include 'ForceOverwrite', true to avoid interactive dialogs.

Install

  1. Download the skill as a ZIP:

    https://autogineer.com/agent-skills/matlab/matlab-model-rf/download

  2. Unzip it into the cross-client skills folder of your project:

    mkdir -p .agents/skills && unzip matlab-model-rf.zip -d .agents/skills

Documentation by MathWorks

RF Toolbox and RF Blockset

Design, analyze, and simulate RF systems in MATLAB -- from measured S-parameter data through full time-domain Circuit Envelope simulation. One unified workflow covering passive networks, active device characterization, system cascade budgets, and behavioral time-domain modeling.

When to Use

Measured data workflows

  • Loading Touchstone files (.s1p through .snp), validating quality (passivity, causality, IEEE P370)

  • Plotting on Smith charts and rectangular plots, interpolating, re-referencing impedance

  • Converting between S, Z, Y, ABCD, T, H, G parameters; computing mixed-mode differential

  • Cascading and de-embedding S-parameter networks; fixture removal

System design and analysis

  • Building RF signal chains from behavioral elements (amplifier, modulator, rffilter, nport, etc.)

  • Computing cascaded gain, noise figure, IP3, SNR, output power with rfbudget

  • Comparing Friis vs Harmonic Balance solver accuracy

  • Exporting to RF Blockset Simulink models, Communications Toolbox, or MATLAB scripts

Circuit-level design

  • Composing arbitrary RF circuits with node wiring (SPICE-like netlist)

  • Designing impedance matching networks (L, Pi, Tee topologies, Richards stubs)

  • Analyzing amplifier stability (K-factor, mu), power gain, and matching conditions

  • Mixer intermodulation analysis and spur-free IF frequency planning

Time-domain and signal integrity

  • Fitting S-parameters to rational models; computing TDR, impulse, step responses

  • Exporting broadband circuit models to SPICE or Verilog-A

  • Building SI channel models from nport blocks, lossy transmission lines, and parasitics

  • Processing complex baseband signals through rf.Amplifier, rf.Mixer, rf.Filter System Objects

  • Full Circuit Envelope time-domain simulation via rfsystem (no Simulink knowledge required)

When NOT to Use

  • Antenna design and radiation patterns -- use Antenna Toolbox

  • RF propagation modeling (path loss, fading, ray tracing) -- use Communications Toolbox or Phased Array System Toolbox

  • SerDes/high-speed serial link equalization (FFE, CTLE, DFE, IBIS-AMI) -- use SerDes Toolbox

  • Radar waveform design or phased arrays -- use Phased Array System Toolbox

Must-Follow Rules

These apply across ALL RF Toolbox workflows. Violating them produces silent errors or blocks execution.

Universal

  1. rfwrite blocks on existing files -- Always pass 'ForceOverwrite', true. Without it, MATLAB opens an interactive overwrite dialog that halts unattended execution.

  2. Elements cannot be reused -- The same element object cannot appear twice in an rfbudget or circuit. Use clone(element) for independent copies.

  3. Element Name must be a valid MATLAB identifier -- No spaces, no special characters. Use 'IFAmp' not 'IF Amp'.

  4. Use tiledlayout/nexttile -- Not subplot. For all multi-panel figures.

  5. rfparam returns complex, not dB -- Always use 20*log10(abs(rfparam(...))) for dB magnitude.

S-Parameters and Network Parameters

  1. sparameters(filename, Z0) is invalid -- Cannot pass reference impedance at load time. Load first, then newref(s, Z0).

  2. rfinterp1 interpolates real/imag independently -- Not magnitude and angle. Can produce artifacts near sharp resonances.

  3. Frequency alignment required for cascade/de-embed -- cascadesparams and deembedsparams require identical frequency vectors. Always rfinterp1 to a common grid first.

  4. deembedsparams always requires 3 arguments -- (sMeasured, sFixture1, sFixture2). For one-sided, pass an ideal thru for the unused side.

  5. Mixed-mode port ordering is the #1 error source -- s2sdd option 1 (default) treats ports 1,2 as positive and 3,4 as negative. Verify which option matches your VNA port numbering.

  6. snp2smp argument order: ports before impedance -- snp2smp(s, portList, Z0), not snp2smp(s, Z0, portList).

rfbudget and Elements

  1. SignalBandwidth is required -- The 4th positional arg to rfbudget. Without it, NF/IP3/SNR remain empty.

  2. Passive NF equals loss -- For rfelement, NF should equal abs(Gain). Prefer attenuator which handles this automatically.

  3. nport Name is positional -- nport('file.s2p', 'MyName'), not name-value syntax.

  4. rffilter: ResponseType is shape, FilterType is algorithm -- rffilter('FilterType', 'Bandpass') errors. Use rffilter('ResponseType', 'Bandpass').

  5. show(b) opens a GUI -- Produces no command-window output. Use properties (b.TransducerGain, b.NF) for programmatic access.

Circuit Composition

  1. setports/setterminals can only be called once -- Plan topology before defining terminals.

  2. 2-port RF elements use 4-node mapping -- add(ckt, [in out inRef outRef], elem) for amplifier, modulator, nport, all txline types. Primitive 2-terminal elements (resistor, capacitor, inductor) use 2-node mapping.

  3. Primitive constructors are positional -- resistor(100, 'R1'), not resistor('Resistance', 100).

  4. setterminals takes a node vector -- setterminals(ckt, [1 2]), not setterminals(ckt, 1, 2).

  5. All txline constructors use name-value pairs only -- No positional arguments.

Amplifier Analysis

  1. powergain gain type string must be last -- powergain(s, 50, 50, 'Gt'), not powergain(s, 'Gt', 50, 50).

  2. powergain returns linear, not dB -- Convert with 10*log10(g) (power gain uses 10x, not 20x).

  3. Gmag returns NaN when conditionally stable -- Use Gmsg at frequencies where K < 1.

Rational Fitting

  1. rational vs rationalfit -- Use rational for new code (AAA). Use rationalfit only for delay extraction.

  2. Single-param fit returns plain vectors; multi-port returns cell arrays -- Always check with iscell(result).

  3. freqresp takes Hz -- Not rad/s. Pass s.Frequencies directly.

System Objects (rf.*)

  1. rf.Amplifier/rf.Mixer use 1-ohm power -- Vin = sqrt(Pin_W). This differs from rfsystem which uses 50-ohm.

  2. rf.Filter frequencies are ABSOLUTE RF -- PassFreq_bp=[2.395e9 2.405e9] means absolute GHz. Not baseband offsets.

  3. Properties locked during use -- Call release(obj) before modifying Nontunable properties.

rfsystem

  1. Requires rfbudget -- rfsystem takes an rfbudget, not raw elements.

  2. close_system needs save flag -- Use close_system(rfs, 0) to discard changes.

  3. IdealizedBaseband only supports amplifier/modulator/rffilter/nport -- Other elements require CircuitEnvelope.

Workflow Quick Reference

TaskEntry PointDetails
Load/plot S-parameterssparameters, rfplot, smithplotreference/sparameters-io.md
Convert S/Z/Y/ABCD, mixed-modezparameters, s2sdd, snp2smpreference/network-conversions.md
Cascade or de-embed networkscascadesparams, deembedsparamsreference/cascade-deembed.md
Create rfbudget elementsamplifier, modulator, rffilter, etc.reference/rfbudget-elements.md
Run cascade budget analysisrfbudget, rfplot(b, 'Pout')reference/rfbudget-analysis.md
Compose RF circuitscircuit, add, setportsreference/circuit-composition.md
Design matching networksmatchingnetwork, richardsreference/matching-networks.md
Analyze amplifier stability/gainstabilityk, powergain, gammamsreference/amplifier-analysis.md
Mixer spurs and IF planningmixerIMT, OpenIFreference/mixer-analysis.md
Fit rational models, TDRrational, stepresp, generateSPICEreference/rational-fitting.md
Build SI channel modelsnport + txlineWRLGC + circuitreference/si-channel-modeling.md
Process complex basebandrf.Amplifier, rf.Mixer, rf.Filterreference/baseband-processing.md
Time-domain system simulationrfsystem, Circuit Envelopereference/system-simulation.md

Conventions

Code Style

  • Use tiledlayout/nexttile for multi-panel figures (not subplot)

  • Always label axes with units (GHz, dB, ns, dBm) and include figure titles

  • Name elements descriptively for readable topology ('LNA', 'DCBlock', 'PCBTrace')

  • Use clone() liberally for element reuse

Script-First Workflow

For design, analysis, or sweep tasks -- write code to .m files on disk, not inline snippets. For quick one-off checks, inline evaluate_matlab_code is fine.

Modern APIs

  • tiledlayout/nexttile instead of subplot

  • datetime instead of datenum

  • rational instead of rationalfit (unless delay extraction needed)

  • smithplot instead of smith

  • rfplot for S-parameter visualization (handles dB conversion automatically)

References

Load the relevant reference file before writing code for a specific workflow. Each reference contains correct calling conventions, constructor arguments, property names, gotchas, and executable code patterns.

Copyright 2026 The MathWorks, Inc.

Files

reference/cascade-deembed.md

Cascade and De-embed S-Parameter Networks

Chain RF components in series (cascade) and remove fixture effects from measurements (de-embed) using cascadesparams and deembedsparams.

Workflow

  1. Load networks -- Load all S-parameter blocks

  2. Align frequencies -- Interpolate all networks to a common frequency grid with rfinterp1

  3. Cascade or de-embed -- Chain networks or remove fixtures

  4. Verify -- Compare results against expectations; round-trip cascade+de-embed to check accuracy

Prerequisite: Frequency Alignment

All networks must share the same frequency vector and reference impedance. This is the most common source of errors.

% Find overlapping frequency range
fCommon = linspace(max(s1.Frequencies(1), s2.Frequencies(1)), ...
                   min(s1.Frequencies(end), s2.Frequencies(end)), 500);

% Interpolate all networks to the common grid
s1i = rfinterp1(s1, fCommon);
s2i = rfinterp1(s2, fCommon);

Gotcha: cascadesparams and deembedsparams require identical frequency vectors. If you skip rfinterp1, you get a dimension mismatch error or silently wrong results. Always interpolate first.

Gotcha: Networks must share the same reference impedance. If impedances differ, re-normalize with newref before cascading: s2i = newref(s2i, s1i.Impedance).

Gotcha: rfinterp1 interpolates the real and imaginary parts independently (like interp1), not magnitude and angle. This can produce artifacts near sharp resonances or rapid phase transitions.

Cascading Networks

Connect networks in series -- output ports of one feed into input ports of the next.

% Two networks
sTotal = cascadesparams(s1i, s2i);

% Three or more networks
sChain = cascadesparams(s1i, s2i, s3i);

Gotcha: cascadesparams assumes standard port ordering: ports 1:N are inputs, N+1:2N are outputs. If your ports follow a different convention, use snp2smp to reorder before cascading.

Identifying Input vs Output Ports

To determine which ports are inputs and which are outputs, inspect the S-parameter matrix at the lowest frequency: off-diagonal elements with the largest magnitude are the through connections.

s4 = sparameters('device.s4p');
S = s4.Parameters(:,:,1);          % S-matrix at lowest frequency
disp(abs(S));                       % Through paths have the largest off-diagonal magnitudes
% If S(2,1) and S(4,3) are the largest, then ports 1,3 are inputs and 2,4 are outputs

For typical 4-port differential pairs: through paths are S21 and S43, meaning ports 1,3 are near-end (input) and ports 2,4 are far-end (output). To use with cascadesparams, reorder so inputs are first:

% Reorder from [1,2,3,4] to [1,3,2,4] so ports 1,2=input and 3,4=output
sReordered = snp2smp(s4, [1 3 2 4]);
sCascade = cascadesparams(sReordered, sReordered);
% Reorder result back to original convention
sCascade = snp2smp(sCascade, [1 3 2 4]);
Multi-Port Cascading

cascadesparams supports networks of different sizes when you specify how many ports connect via the third argument:

% Default (no 3rd arg): both must be 2N-port, connects last N to first N
sTotal = cascadesparams(s2a, s2b);      % Two 2-port -> 2-port result

% With connection count (3rd arg): works with different-sized networks
sTotal = cascadesparams(s4, s2, 1);     % 4-port + 2-port, connect 1 port
% Result: (4 + 2 - 2*1) = 4-port

The third argument specifies how many ports connect between the networks. The last M output ports of the first network connect to the first M input ports of the second. The result has (N1 + N2 - 2*M) ports.

For complex topologies where cascadesparams port ordering doesn't match your physical setup:

  • Use snp2smp to reorder ports before cascading

  • Or use the circuit object for full topology control

De-embedding Fixtures

Remove fixture (test fixture, adapter, cable) effects from a measured DUT to recover the true device response.

Gotcha: deembedsparams always requires 3 arguments: deembedsparams(sMeasured, s1, s3) where s1 is the port-1 fixture and s3 is the port-2 fixture. Calling with only 2 arguments errors with "Not enough input arguments." For one-sided de-embedding, pass an ideal thru for the unused side.

Two-Sided De-embedding
% Measured: fixture1 + DUT + fixture2
% Goal: extract DUT S-parameters
sDUT = deembedsparams(sMeasured, sFixture1, sFixture2);
One-Sided De-embedding

To de-embed only one side, construct an ideal thru for the other side:

nf = numel(sMeasured.Frequencies);
thruData = zeros(2, 2, nf);
thruData(1,2,:) = 1;
thruData(2,1,:) = 1;
sThru = sparameters(thruData, sMeasured.Frequencies, 50);

% Remove only the left fixture
sDUT = deembedsparams(sMeasured, sFixture1, sThru);

% Remove only the right fixture
sDUT = deembedsparams(sMeasured, sThru, sFixture2);
Half-Fixture Extraction from 2x-Through

When you only have a through measurement of two identical fixtures back-to-back (no separate fixture characterization), extract one fixture using T-parameter matrix square root:

% Load 2x-through measurement (two identical fixtures cascaded)
s2xThru = sparameters('2x_through.s2p');

% Convert to T-parameters and take matrix square root at each frequency
T2x = s2t(s2xThru.Parameters);
nf = numel(s2xThru.Frequencies);
Tfixture = zeros(2, 2, nf);
for k = 1:nf
    Tfixture(:,:,k) = sqrtm(T2x(:,:,k));
end

% Convert back to S-parameters
sFixture = sparameters(t2s(Tfixture), s2xThru.Frequencies, s2xThru.Impedance);

% Now de-embed DUT using the extracted fixture
sDUT = deembedsparams(sMeasured, sFixture, sFixture);

Note: This assumes symmetric fixtures (identical left and right). For asymmetric fixtures, more advanced techniques are needed.

Synthetic Fixture Data (for testing)
freq = linspace(1e9, 20e9, 201);
loss = 10^(-0.5/20);  % 0.5 dB per fixture
phase = reshape(exp(-1j*2*pi*freq*15e-12), 1, 1, []);  % 15 ps delay
sData = [0.03 0; 0 0.03] + [0 loss; loss 0].*phase;  % 2x2xN via implicit expansion
sOneFixture = sparameters(sData, freq, 50);
s2xThru = cascadesparams(sOneFixture, sOneFixture);
4-Port De-embedding

deembedsparams works directly with 4-port (and N-port) networks -- same 3-argument syntax:

% 4-port fixture and 4-port measurement
sFixture4p = sparameters('fixture_4port.s4p');
sMeasured4p = sparameters('measured_4port.s4p');

% De-embed both sides (symmetric fixtures)
sDUT4p = deembedsparams(sMeasured4p, sFixture4p, sFixture4p);

% Or asymmetric fixtures
sDUT4p = deembedsparams(sMeasured4p, sFixtureLeft4p, sFixtureRight4p);

All networks must have the same number of ports. Port ordering must be consistent across fixture and measurement.

Verification Pattern

Cascade then de-embed to verify round-trip accuracy:

% Build chain
sChain = cascadesparams(sFixture, sDUT, sFixture);

% Recover DUT
sRecovered = deembedsparams(sChain, sFixture, sFixture);

% Check error
origS21 = rfparam(sDUT, 2, 1);
recS21 = rfparam(sRecovered, 2, 1);
maxErr = max(abs(origS21 - recS21));
fprintf('Round-trip S21 max error: %.2e\n', maxErr);
VNA Calibration Check (De-embed Through Standard)

Verify calibration by de-embedding fixtures from a known through measurement -- the result should be an ideal through (0 dB insertion loss, 0 phase):

% Measure through standard with fixtures in place
sThruMeasured = sparameters('through_measured.s2p');

% De-embed fixtures
sThruDeembedded = deembedsparams(sThruMeasured, sFixtureL, sFixtureR);

% Check deviation from ideal (S21 = 1, S11 = 0)
s21 = rfparam(sThruDeembedded, 2, 1);
s11 = rfparam(sThruDeembedded, 1, 1);
fprintf('S21 mag error: %.4f dB\n', max(abs(20*log10(abs(s21)))));
fprintf('S21 phase error: %.4f deg\n', max(abs(angle(s21)*180/pi)));
fprintf('S11 max: %.1f dB\n', max(20*log10(abs(s11))));

Gotchas

  1. Frequency alignment is required -- cascadesparams and deembedsparams require identical frequency vectors. Always rfinterp1 to a common grid first, or you get dimension mismatch errors or silently wrong results.

  2. deembedsparams always requires 3 arguments -- deembedsparams(sMeasured, s1, s3) where s1 is the port-1 fixture and s3 is the port-2 fixture. Calling with only 2 arguments errors. For one-sided de-embedding, pass an ideal thru for the unused side.

  3. rfinterp1 interpolates real/imag independently -- (like interp1). Not magnitude and angle. Can produce artifacts near sharp resonances or rapid phase transitions.

  4. cascadesparams assumes standard port ordering -- Ports 1:N are inputs, N+1:2N are outputs. If your ports follow a different convention, use snp2smp to reorder before cascading.

Conventions

  • Always interpolate to a common frequency grid before cascading or de-embedding

  • Use tiledlayout/nexttile when comparing original vs. cascaded/de-embedded results

  • Label plots clearly to distinguish standalone vs. cascaded responses

  • Verify de-embedding accuracy by round-tripping when fixture data is available

Copyright 2026 The MathWorks, Inc.

Requirements and compatibility

Works with: Claude Code, Claude apps, OpenAI Codex, GitHub Copilot, Cursor, Gemini CLI, Kiro, OpenCode, Goose, Amp, Roo Code, JetBrains Junie, Factory Droid, Devin (ex-Windsurf), Mistral Vibe, OpenHands, Letta

“Works with” means the tool reads the open Agent Skills format; a skill written for one tool may use features others lack.

Needs: Needs Node.js

Security

No issues found by our static scan· No issues found

A static scan is not a guarantee. Skills are third-party content: review a skill before you let an agent use it.

Attribution and licence

Created by
MathWorks
Publisher
MathWorks (Official)
Source
matlab/matlab-agentic-toolkit · skills-catalog/rf-and-mixed-signal/matlab-model-rf · commit 2a7f215
Licence
BSD 2-Clause License (redistribution allowed)
  • LICENSE (added by AutoGineer from LICENSE.md)
  • ATTRIBUTION.md (written by AutoGineer)
Synced

AutoGineer is not affiliated with or endorsed by MathWorks or MathWorks.

Download ZIP

The ZIP holds the skill’s files, its licence files and an ATTRIBUTION.md. Download matlab-model-rf.zip