Estimate near-end and far-end crosstalk from simple coupled-line geometry, or calculate from odd/even-mode impedance data when a solver-backed model is available.
Inputs and tolerances
Estimate NEXT and FEXT with either a simple geometry model or direct odd/even-mode impedance data.
Use this for an early geometry check when you have dimensions, Er/Dk, rise time, and aggressor voltage but do not yet have odd/even-mode impedance data.
Calculate NEXT and FEXT for the entered coupled length.
Outer-layer edge-coupled routing over one reference plane. Uses dielectric height H and the 3H spacing guideline.
Microstrip simple mode uses dielectric height H, estimates k from edge-to-edge spacing, and checks the common 3H spacing guideline.
NEXT and FEXT estimate
Near-end and far-end results are reported as percentage, dB, and coupled voltage.
NEXT0.521%Range: 0.3058 to 0.8721 %
NEXT-45.66dBRange: -50.29 to -41.19 dB
NEXT voltage17.19mVRange: 9.586m to 30.22m V
FEXT0.01274%Range: 0.006723 to 0.02408 %
FEXT-77.9dBRange: -83.45 to -72.37 dB
FEXT voltage420.4µVRange: 210.8µ to 834.3µ V
Kb0.007906ratioRange: 0.005293 to 0.01181 ratio
Kf2.548ps/mRange: 1.643p to 3.94p s/m
Coupling coefficient k0.03163ratioRange: 0.02117 to 0.04724 ratio
Odd-mode Z₀68.99ΩRange: 61.31 to 77.06 Ω
Even-mode Z₀73.5ΩRange: 64.63 to 83.39 Ω
Saturation length92.96mmRange: 82.03m to 104.4m m
S/H ratio3ratioRange: 2.455 to 3.667 ratio
Design rulePass
Crosstalk limitPass
Model limit: This is a first-pass analytical estimate for early PCB layout screening. It is not a field solver and does not replace stackup-aware SI simulation or measurement on critical nets.
This calculator separates near-end crosstalk and far-end crosstalk so a quick spacing check does not hide the difference between coupling strength, edge rate, and coupled length.
Simple geometry mode
Use trace width, edge-to-edge spacing, dielectric height, copper thickness, Er/Dk, rise time, and aggressor voltage for an early PCB layout estimate.
Engineering odd/even mode
Use Zeven, Zodd, Er,eff, and Kf when a stackup tool, coupled-line calculator, field solver, or measured model already provides transmission-line data.
Use as a screening tool
Treat the result as a first-pass noise estimate. Critical nets still need stackup-aware simulation, return-path review, and measurement.
The two workflows answer the same NEXT/FEXT question with different levels of input confidence.
Fast layout screen
Simple geometry
Choose simple geometry when you are still moving traces around and need to compare spacing, height, length, and edge-rate sensitivity. It estimates k from dimensions and reports design-rule pass/fail guidance.
Good for early routing trade-offs.
Supports microstrip and stripline assumptions.
Not a substitute for odd/even-mode impedance extraction.
Solver-backed analysis
Engineering odd/even
Choose engineering odd/even mode when the stackup is known and you can enter Zeven and Zodd directly. The calculator then computes k and Kb from those values instead of estimating coupling from spacing.
Use field-solver or stackup-tool odd/even data.
Enter Kf explicitly for FEXT.
Best for documenting a known coupled-line model.
The calculator uses the same NEXT/FEXT structure in both workflows, but changes how k and Kf are obtained.
Core relationships
NEXT, FEXT, and coupling
NEXT
Near-end crosstalk rises toward the backward coupling coefficient as the coupled length approaches saturation.
Odd/even coupling
Engineering mode calculates k and Kb directly from odd-mode and even-mode impedances.
FEXT
Far-end crosstalk scales with the FEXT coefficient, coupled length, and inverse rise time.
Saturation length
The simple NEXT estimate uses velocity from effective dielectric constant.
Variables and outputs
NEXT: Near-end crosstalk
Unit: % / dB / V
Victim noise observed near the aggressor source end. It rises with coupling strength and approaches a saturation value with long coupled length.
FEXT: Far-end crosstalk
Unit: % / dB / V
Victim noise observed at the far end. It depends on edge rate, coupled length, and even/odd-mode propagation differences.
k: Coupling coefficient
Unit: ratio
Dimensionless odd/even-mode coupling ratio. Engineering mode calculates it directly from Zeven and Zodd.
Kb: Backward crosstalk coefficient
Unit: ratio
NEXT saturation coefficient. In the odd/even model, Kb = k / 4.
Kf: Forward crosstalk coefficient
Unit: s/m
FEXT coefficient used by FEXT = Kf × L / Tr. Use field-solver or stackup-tool data where possible.
Lsat: Saturation length
Unit: m
Approximate coupled length where NEXT stops increasing strongly: Lsat = Tr × v / 2.
Model boundary
The simple geometry workflow estimates coupling from limited dimensions. Real crosstalk also depends on solder mask, copper roughness, neighbouring copper, plane cavities, return-path continuity, stackup tolerances, and driver/receiver impedance.
Use engineering mode when odd/even-mode data is available.
Use simulation or measurement for critical clocks, high-speed links, and sensitive analogue nets.
Do not use this calculator as compliance evidence.
These examples show the difference between a geometry estimate and a solver-backed odd/even calculation.
Worked example
Simple microstrip geometry
A 3.3V aggressor with a 0.1ns edge couples for 50mm beside a nearby victim trace.
Inputs
Trace geometry
W = 150µm, S = 150µm, H = 100µm, T = 35µm
Signal
L = 50mm, Er = 4.3, Tr = 0.1ns, Va = 3.3V
Equation and substitution
k
≈0.1265
NEXT
≈3.16%
NEXT voltage
≈104mV
Next check
Move to odd/even data if this route is timing-critical, high impedance, or noise-sensitive.
Worked example
Engineering odd/even inputs
A coupled-line model reports Zodd = 45Ω and Zeven = 55Ω.
Inputs
Odd/even data
Zodd = 45Ω, Zeven = 55Ω
Forward coefficient
Kf = 10.4ps/m
Equation and substitution
k
0.1
Kb
0.025
FEXT form
Kf × L / Tr
Use case
Document solver-derived coupling without relying on geometry heuristics.
Use the calculator as one step in a broader layout review rather than as the final answer.
Start with simple geometry
During placement and early routing, compare spacing, coupled length, and edge rate quickly enough to change the layout.
Escalate when the result is close
If the crosstalk limit fails or sits near the target, extract odd/even data from a stackup tool or field solver and re-run engineering mode.
Close the loop with verification
For critical nets, confirm the result with SI simulation, oscilloscope measurement, or compliance test data using the real driver and receiver environment.
Design follow-up
Next engineering checks
Use these follow-up checks before turning the calculated value into a component choice, layout decision, or production limit.
Confirm edge rate at the driver
Crosstalk follows edge rate more than clock frequency. Use the real 10–90% or 20–80% transition time where possible.
Check coupled length, not total route length
Only the parallel section with meaningful field overlap should be entered as coupled length.
Keep the return path continuous
Plane splits, reference changes, stitching gaps, and loose return current can dominate coupling even when spacing looks acceptable.
Use odd/even data for critical nets
For clocks, SerDes, high-impedance analogue, sensitive reset lines, or compliance-sensitive nets, move from simple geometry to field-solver or SI data.
Related calculators and next checks
Move from crosstalk estimation to edge-rate, propagation, conversion, and power-integrity context when the layout risk is higher.
