ComputeEM Studios/Microwave FDTD Studio
MICROWAVE FDTD STUDIO / MULTIBAND GUIDE

Inside the guide.
Across the irises.

A two-dimensional section through the original waveguide simulation.

2D electric-field section · original solver
SCROLL TO MOVE THE WAVESOriginal FDTD solver · recorded field section
t+
PREDICTIVE SIGNAL CONTINUATION

Predict the tail.
Shorten the simulation.

Resonant filters can keep ringing long after excitation. Microwave FDTD Studio uses the recorded port-signal decay to estimate its continuation, helping evaluate the frequency response without always waiting for the full tail to decay in the FDTD run.

Recorded transientPredicted decayFrequency-response comparison

Why the signal tail matters

Stopping a transient while appreciable energy remains can leave truncation ripples in a frequency-domain response. Signal continuation estimates the remaining decay from the acquired record, so its contribution can be included in the response calculation.

01 · Record

Run FDTD and retain the computed port signals. These samples remain the direct simulation record.

02 · Estimate

Fit the observed late-time behavior and extend the port response beyond the recorded interval. The continuation is predicted, not additional FDTD time steps.

03 · Compare

Inspect the time-domain join and compare the frequency responses before and after estimation. Where provided below, the original exports show both views.

The practical advantage: a suitable, validated continuation can reduce the acquisition needed for slowly decaying resonant responses. The benefit depends on the structure, recording length and quality of the fit; these examples do not establish a universal speedup.

How to assess a predicted response

A smoother curve alone is not an accuracy test. Check prediction against withheld or longer recorded data, and check sensitivity to the fitting window and acquisition length. Grid convergence, port normalization, reference planes and losses still matter. Port-signal prediction does not extend spatial field volumes or generate new heat maps. Recorded and predicted results should be distinguished when reporting device performance.

MICROWAVE FDTD STUDIO · APPLICATION EXAMPLES

Explore the devices.

Ten waveguide, coaxial and microstrip examples. Open each model to inspect its original results.

MICROWAVE FDTD STUDIO

From guided fields to device response.

01

Passive-device workflows

Explore filters, couplers and guided-wave structures in a workspace focused on microwave engineering.

02

Predictive port-signal continuation

Estimate the long-lived tail of port voltage and current responses from a shorter FDTD record. Check the continuation before using it in the response analysis.

03

Field inspection

Visualize field distributions to understand propagation, resonances and coupling within the structure.

04

Power-loss investigation

Explore electromagnetic loss and heat-estimation workflows, with attention to the assumptions and convergence of each model.

A FOCUSED WORKFLOW

Move from setup to understanding.

01

Choose a structure or template

02

Set materials, ports and boundaries

03

Run and check convergence

04

Inspect response, fields and losses

LET’S TALK ELECTROMAGNETICS

Have a problem you’d like to simulate?

Tell us about your geometry, frequency range and the results you need. Start a conversation about Microwave FDTD Studio and a suitable demonstration.

Connect with MohammadMohammad Marvasti · Founder of ComputeEM Studios and its software family
ALSO IN THE COMPUTEEM FAMILY
About these recorded fields

These fields come from complete runs of the original waveguide and coaxial solver equations on their original grids. The four earlier port records were compared with their bundled original baselines. The multiband and dual-mode waveguides use clean 2D sections; the four-pole, zero-free and six-pole, two-zero coaxial filters use CAD cutaways. Each earlier example has 160 recorded time samples. The silver-plated chapter uses the original lossless PEC model of that CAD geometry; it does not model finite-conductivity silver loss. The ninth-order, silver-plated and third-order chapters are additional captures from their original solver setups; they are not extra frames inferred from the exported demonstration videos. The final microstrip chapter shows a horizontal field section from a fresh run of the original three-port T-junction template. Contours are drawn from the recorded scalar fields, with a separate high-resolution CAD layer. DualMode-WG uses the original planar solver; the other examples use the original 3D grids. Coaxial magnitude views combine the three electric components only after Yee-to-cell-center collocation; waveguide close-ups show actual central cavities without stretching their geometry. These are recorded fields, not invented propagation; display selection and compression do not alter the retained raw results.