Asymmetric Stripline Impedance Calculator computes characteristic impedance per IPC-2141A from trace width, thickness, dielectric heights H1/H2, and substrate dielectric constant.
Calculate Asymmetric Stripline Impedance for Off-Center PCB Traces
The Asymmetric Stripline Impedance Calculator computes characteristic impedance per IPC-2141A from trace width, thickness, dielectric heights H1/H2, and substrate dielectric constant. PCB stack-up and signal-integrity engineers rely on it whenever a signal layer cannot be centered between planes, a routine situation on layer 3 of a 4-layer board or in denser multilayer stacks.
How to Use
Enter trace width (W), trace thickness (T), the two substrate heights above and below the trace (H1, H2), and the substrate dielectric constant (εr) — each in mm, mil, inch, or µm. The tool returns characteristic impedance (Z₀) in ohms, plus derived propagation delay, capacitance, and inductance per unit length.
Formula
IPC-2141A, “Design Guide for High-Speed Controlled Impedance Circuit Boards,” Section 4.2.5, defines asymmetric stripline impedance by adapting the symmetric stripline equation from Section 4.2.4.
Step 1 — total dielectric spacing between the two reference planes:
$$B = H_1 + H_2 + T$$
Step 2 — equivalent air-dielectric impedance at that spacing, using the narrow-trace symmetric stripline formula (valid for $W/(B-T) \leq 0.35$):
$$Z_{0,SS} = 60 \times \ln\left(\frac{1.9B}{0.8W + T}\right)$$
Step 3 — IPC-2141A subtracts a correction term, $\Delta Z_{0,air}$, built from that same symmetric-stripline formula evaluated separately at H1 and H2, to account for the trace sitting off-center rather than at B/2.
Step 4 — divide by $\sqrt{\varepsilon_r}$ to bring in the real dielectric:
$$Z_{0,AS} = \frac{1}{\sqrt{\varepsilon_r}}\left(Z_{0,SS} – \Delta Z_{0,air}\right)$$
This overall structure is confirmed directly in IPC-2141A Section 4.2.5. However, the standard’s own sub-equations for $\Delta Z_{0,air}$ (numbered 4-17 through 4-19 in the published text) were later revised by an official IPC errata, so we won’t reprint a version of those sub-equations here that might not match your printing of the standard — pull the final number from the calculator above, which implements the corrected form.
Sanity check: set H1 = H2 and the asymmetric result must equal the plain symmetric-stripline value for the same total spacing B. Any asymmetric stripline tool — including this one — that doesn’t converge back to the symmetric formula under that condition is applying the correction term incorrectly.
Common input mistake: entering nominal copper weight (e.g., “1 oz”) for T instead of actual etched thickness, which typically finishes 10–20% thinner after processing. Because T sits inside the $(0.8W + T)$ term, this alone can shift Z₀ by several ohms.
FAQ
What’s the difference between symmetric and asymmetric stripline?
Symmetric stripline has the signal trace centered between two identical reference-plane spacings (H1 = H2). Asymmetric stripline has unequal spacing, which is common whenever a signal layer references two different planes in a multilayer stack.
Why does the H1/H2 split matter if H1 + H2 stays the same?
Impedance depends on proximity to each plane, not just total spacing. Moving the trace closer to one plane while holding H1+H2 constant lowers Z₀, because the nearer plane dominates the coupling.
When should I use a field solver instead of this calculator?
Once $W/(B-T)$ exceeds roughly 0.35, or your stackup mixes different dielectric materials on each side, the IPC-2141A closed-form approximation loses accuracy and a 2D field solver gives a more reliable number.
Does trace thickness (T) really change the impedance that much?
Yes. T appears directly in the $(0.8W + T)$ term, so using etched thickness instead of nominal copper weight can shift the result by several ohms on thinner dielectrics.
Can I use this for a differential pair?
No. This tool gives single-ended impedance only. Differential asymmetric stripline needs an additional coupling correction between the two traces, which this calculator doesn’t compute.