FR4 vs Rogers PCB High-Frequency Sizer
Compare microstrip width, impedance, propagation delay, and dielectric loss before choosing a board stackup. Start with a target impedance, or see what changes when the same trace is placed on a different material.
Size the trace. Compare the trade-offs.
Use material data for your laminate construction and operating conditions. Default values are starting assumptions, not a production stackup approval.
Microstrip sizing results
Waiting for the calculator.
Material A
- Quasi-static Z₀
- —
- Effective Dk
- —
- Propagation delay
- —
- Guided wavelength
- —
- Dielectric loss / 100 mm
- —
- Length at allocation¹
- —
Material B
- Quasi-static Z₀
- —
- Effective Dk
- —
- Propagation delay
- —
- Guided wavelength
- —
- Dielectric loss / 100 mm
- —
- Length at allocation¹
- —
Dielectric loss over the entered length
Both bars share one relative scale. These values exclude conductor, roughness, mismatch, radiation, and transition losses.
¹ Theoretical length for the dielectric-only allocation, with fixed Dk and Df. It is not the maximum usable PCB route length.
The report includes your inputs, assumptions, and source references.
What changes when the material changes?
Compare the geometry and the loss together. A material change can reduce one contribution to attenuation while requiring a different trace width.
Check the routing space
Use Size for target Ω to estimate a separate width for each material. Use Compare same width to reveal the impedance change if you keep the existing layout. The tool does not include mismatch loss in either mode.
Allocate only part of the budget
The loss result covers the dielectric only. Reserve separate allowances for copper, roughness, connectors, vias, and discontinuities before judging the end-to-end channel.
Review electrical length
Effective Dk determines the estimated propagation delay and guided wavelength. These are nondispersive estimates; phase-sensitive RF paths need a frequency-dependent stackup model.
FR4 is a baseline, not one fixed Dk
Confirm the laminate grade, glass style, resin content, test method, and frequency. A value copied from a different construction may not describe your board.
| Material | Dk used | Df used | How to interpret it |
|---|---|---|---|
| FR4 example | 4.20 | 0.0200 | Illustrative inputs only. No specific grade or frequency characterization is implied. |
| Rogers RO4003C | 3.55 design Dk | 0.0027 | Design Dk: 8–40 GHz. Typical Df: 10 GHz, 23°C. |
| Rogers RO4350B | 3.66 design Dk | 0.0037 | Design Dk: 8–40 GHz. Typical Df: 10 GHz, 23°C. |
| Custom material | Your input | Your input | Use supplier or fabricator data appropriate to your construction and analysis frequency. |
Rogers presets use the manufacturer’s design Dk, not its process Dk. The RO4000 laminate datasheet, typical-properties table distinguishes the measurement methods. Values remain constant when you change frequency; the calculator does not interpolate a material dispersion curve.
For an example of construction-specific FR4 data, compare the frequency and resin-content rows in Isola’s 370HR Dk / Df tables. Those tables are not the source of the illustrative FR4 preset above.
Choose against the actual channel
- Short interconnect or generous loss allocation? Evaluate the proposed FR4 grade using the actual route length and the complete loss budget.
- Longer RF path or tight phase requirements? Compare a characterized high-frequency laminate, then include conductor effects and transitions in the channel model.
- Changing material on an existing board? Recheck impedance, trace width, stackup tolerances, and component launches before reusing the layout.
- High-speed digital rather than a single RF tone? A clock frequency alone does not describe the signal spectrum. Review edge rate and the full channel in a signal-integrity model.
Bring these details to your fabricator
- Stackup: exact laminate and construction, finished dielectric heights, copper weight, plating, and solder-mask coverage.
- Controlled impedance: target, tolerance, reference layer, finished trace dimensions, and coupon measurement plan.
- RF path: operating band, route lengths, phase or delay constraints, connectors, vias, and permitted insertion loss.
- Release evidence: agreed material documentation, stackup confirmation, and verification results needed for your design.
A transparent first-pass model
Understand what is calculated before using the numbers in a design review.
Geometry and propagation
Quasi-static impedance and effective permittivity use the zero-thickness Hammerstad–Jensen microstrip equations. Width sizing inverts that model numerically.
u = w / h
Delay = L × √εeff / c
Guided wavelength = c / (f × √εeff)Here, εeff is the calculated effective permittivity, L is length in metres, and c is the speed of light in vacuum.
Model reference: Qucs microstrip technical documentation, equations 11.4–11.6 and 11.15–11.18.
Dielectric attenuation only
αd = [πf / c] × [Dk / √εeff]
× [(εeff − 1) / (Dk − 1)] × Df
Loss (dB) = (20 / ln 10) × αd × Lαd is attenuation in nepers per metre. With a dielectric-only allocation B, the theoretical length is B divided by the dielectric attenuation in dB per metre.
Dielectric-loss reference: Qucs microstrip transmission-loss model, equation 11.79. For the wider simulation context, see Rogers: Using Simulation to Assist with PCB Design.
Tool input limits: 0.1–40 GHz; 1–2,000 mm route length; 0.05–3.2 mm dielectric height; Dk 1.2–15; Df 0–0.1; 0.1 ≤ w/h ≤ 10. Target-impedance mode accepts 20–120 Ω only where a solution exists within that width-to-height range. These input limits are not a guarantee of physical model accuracy.
Excluded: finite copper thickness, copper conductivity and roughness, solder mask, dielectric dispersion, weave anisotropy, coupling, vias, connectors, radiation, fabrication tolerance, and thermal variation. At h/λ₀ ≥ 0.05, an additional caution is shown; this is a screening flag, not a universal validity boundary. Final dimensions require a fabricator-approved stackup and an appropriate field solver.
FR4 vs Rogers: practical answers
At what frequency must I switch from FR4 to Rogers?
There is no single cutoff that this calculator can establish. The decision depends on the specific FR4 grade, length, channel loss and phase targets, stackup, copper, and fabrication capability. Use the comparison to identify whether dielectric loss is a concern, then evaluate the complete channel.
Why is the trace width different at the same impedance?
Width-to-height ratio and effective permittivity both affect microstrip impedance. When Dk changes, the width generally needs to change to retain the same target impedance. Keeping the width unchanged is a different comparison, which you can select in the tool.
Does the displayed loss include copper or connectors?
No. It is the dielectric contribution only. Real insertion loss also depends on conductors, roughness, reflections, discontinuities, launches, and other channel details. The displayed allocation length is not an approved maximum route length.
Should I enter the full PCB thickness?
Enter the dielectric distance from the surface trace to its reference ground plane. On a multilayer board, this may be much smaller than the full board thickness. Both materials use the same height in this comparison.
Can I use the Rogers presets at 2.4 or 5 GHz?
You can run a constant-property scenario, but the built-in Df values are specified at 10 GHz and the design Dk values cover 8–40 GHz in the cited datasheet. The results are not characterized material performance at 2.4 or 5 GHz. Use appropriate supplier or fabricator data for your operating band.
Does this size stripline, differential pairs, or coplanar waveguides?
No. The model is for an isolated, single-ended surface microstrip with air above the trace and a continuous reference plane below. Other geometries need their own models and inputs.
Related tools and sourcing support
Have the circuit. Need the components?
Send your RF, analog, interface, or power-component part numbers and quantities. YURUNOX can review sourcing options, available evidence, and delivery requirements for your BOM.