Loading…
Solve an offset stripline with a different Dk above and below the trace, or an embedded microstrip with a dielectric cover. The panel gives a harmonic-mean estimate in real time and runs the exact 2-D Laplace solve on the real permittivity map when the asymmetry matters.
Direct answer
An offset stripline is the normal inner-layer case, not an exception. When the trace sits closer to the lower plane, the lower half couples harder and the impedance falls below the symmetric value. When the two dielectrics also differ, the field crosses a permittivity boundary and no single-Dk table applies. On a 0.5 mm cavity with the trace 0.15 mm from the lower plane, a 0.2 mm trace is typically several ohms below the same trace centred, and the exact 2-D solve is the number to use.
Calculator
Pick the structure, enter the dielectric on each side and the target impedance. The closed form gives an instant width; the exact button solves Laplace's equation on the actual permittivity map, including the offset, the two Dk values and the cover. The solved width never drops below the fabrication minimum you enter.
Model and validity range. A harmonic-mean closed form of two symmetric stripline halves, then an exact 2-D Laplace solve with erLower below the trace and erUpper above it, finite copper, a cover layer and a causal wideband Dk. Both paths evaluate the Dk at the entered frequency; the closed form holds while the trace is not within 20% of a plane and the two Dk values are close, and outside that window the field solve is the number to read.
Direction
Target and dielectrics
Geometry
Material and frequency
Changes the causal Dk when the frequency differs from the reference; it is not a loss result on this panel.
Anchors the dispersion curve to the datasheet measurement.
Moves the causal Dk, delay and guided wavelength.
Asymmetric stripline and embedded microstrip
Offset traces, two Dk values and a dielectric cover, with the exact 2-D solve one click away.
50.00 Ω closed form at 0.153 mm
50.00 Ω closed form at 0.153 mm
Harmonic-mean closed form. Run the exact solve when the two Dk values differ or the trace is visibly offset.
Harmonic-mean closed form.
Sets the propagation velocity.
170.62 ps/in.
6.0 mil.
Share of the total dielectric that sits above the trace.
Half-line impedance of the lower dielectric.
Half-line impedance of the upper dielectric.
The same trace centred in the same total dielectric.
Run the exact solve for the delta.
At 1.000 GHz.
Model limits
What this does not cover
Free, no account. An engineer reviews these inputs before the fixed fee is issued.
Results are engineering estimates, not design verification. Confirm them against the fabrication drawing, a TDR coupon or a 3-D extraction before release. Read the tool terms.
A wrong number, a rejected in-range value or a missing model? Send it to the engineer.
Every result above traces to an equation below. Read the equations and their validity limits.
Equations
Each block gives the equation the calculator evaluates, every symbol with its unit, and the range over which the model holds, with what it cannot see stated alongside. The button in the corner copies the LaTeX source.
Each side of the trace is treated as half of a symmetric stripline, and the two capacitances add.
The lower half-line is a symmetric stripline of plane spacing twice the lower height, and the upper half is the same construction with the upper height. Their impedances are the two Z values in the formula.
Adding the two capacitances is equivalent to combining the two impedances in parallel, which is the harmonic mean. For equal heights and equal Dk the result collapses to the ordinary symmetric stripline value.
Model limits
The half-line construction assumes each side behaves like a symmetric stripline. A trace within 20% of one plane, or a copper thickness that is a large fraction of the gap, breaks that assumption and needs the exact solve.
The two dielectrics contribute in proportion to the capacitance each one stores, not in proportion to their thickness.
The field does not divide equally between the two sides. It concentrates where the capacitance is higher, so the higher-Dk side carries more of the stored energy and pulls the effective Dk toward its value.
Using a thickness-weighted average is a common shortcut and it is wrong in the direction that matters: it underestimates the pull of a thin, high-Dk layer close to the trace.
Model limits
The weighted mean is a quasi-static estimate. It ignores the field that crosses the boundary between the two dielectrics, which is exactly what the exact 2-D solve adds.
The weight in the effective-Dk average is the capacitance of each half-line, which follows from its impedance and permittivity.
For a TEM line the capacitance per unit length is the square root of the permittivity divided by the speed of light and the impedance. The same relation holds for each half of the asymmetric section.
The constant c0 cancels in the weighted average, which is why the tool can use the impedances alone to weight the two Dk values.
Model limits
The relation is exact only for a homogeneous TEM half-line. The half-line is a model of one side of the cavity, not a separate physical transmission line, so the weight is an approximation.
Once the effective Dk is known, the delay per millimetre is a division by the speed of light.
The asymmetric section is treated as a quasi-TEM line, so the delay follows from the effective permittivity in the same way as a microstrip or stripline.
This is why the two Dk values matter twice: they set the impedance and they set the effective Dk that determines the delay. A low-Dk side raises the impedance and speeds the line at the same time.
Model limits
The relation is a quasi-static result. Both the closed form and the exact solve evaluate a causal wideband Dk at the entered frequency, but the closed form still cannot see the field that crosses the boundary between the two dielectrics; the exact solve can.
The exact answer solves Laplace's equation on the actual permittivity map, then derives impedance and effective Dk from the two capacitances.
The solver puts the trace at one volt and the planes at zero, and relaxes the potential on a finite-volume grid until the flux is balanced. The charge on the conductor gives the capacitance.
The same geometry is solved twice: once with the real permittivity map and once with every cell set to air. The ratio of the two capacitances is the effective Dk, and the product gives the impedance. That split is why the answer does not need a field-filling approximation.
Model limits
The solve is two-dimensional and quasi-static. It does not see a via, a connector launch, a trace that changes width or a plane cavity mode, and its staircase mesh reports a residual rather than an exact answer.
Geometry
A symmetric stripline assumes the trace sits exactly between two planes. A real cavity often does not: the core below the trace is a different thickness from the prepreg above it, or the stackup has a thin plane pair next to a thick one. The moment the two heights differ, the cross-section is asymmetric.
The closed form treats each side as half of a symmetric stripline and adds the two capacitances. The impedances combine as a harmonic mean, which is at or below the arithmetic mean, so any offset lowers the impedance. The nearer plane couples harder than the far plane gives back.
Material
A cavity can have a low-Dk prepreg on one side and an FR-4 core on the other. The field then crosses a permittivity boundary inside the same cross-section, and the effective Dk is neither value. A single-Dk calculator cannot represent that geometry, no matter how carefully the height is measured.
The closed form uses a capacitance-weighted mean of the two Dk values, which is a reasonable first estimate. The exact solve writes erLower below the trace and erUpper above it and solves the boundary-value problem, so the field's preference for the higher-Dk side is in the answer.
The result is a cross-section, not a stackup. Reconcile it with the fabricator's prepreg and press-out data in the Stackup Planner and the Impedance Calculator before the stack is released.
Cover
An outer-layer trace is not always in air. A conformal coat, a laminated coverlay or a potting compound adds a dielectric above the copper, which adds capacitance, lowers the impedance and slows the line. The effect is small for a thin coat and significant for a thick coverlay.
The closed form mixes the substrate and cover Dk by thickness, which is the parallel-plate limit. The exact solve puts the cover on as its own layer with its own thickness and Dk, so the fringing field sees the real boundary. That is the number to compare with a coupon when the cover is part of the build.
A cover also changes the loss and the thermal path. Our high-speed PCB layout service, SI/PI verification service and DFM/DFA layout service work reconcile the electrical target with the fabrication and coverage requirements, and PCB layout pricing prices the layout once the construction is fixed.
FAQ
Send the stackup, the interface list and the offset-stripline geometry inputs. An engineer re-solves them against the real layout before the fixed fee is issued.