勛圖tv 繚 RF ENGINEERING
Copper Roughness & Signal Integrity: Loss and Foil Choices
Copper roughness can influence signal integrity through conductor loss and propagation behavior, especially when the surface scale becomes relevant to the current distribution. The practical design task is to identify the actual foil and interface, use compatible roughness data and models, and compare representative circuits. A label such as low profile is not a complete electrical specification.
This guide focuses on the conductor contribution to a PCB interconnect. It separates material descriptors, ideal skin-depth calculations and measured channel performance so an engineer can ask for useful evidence without turning a roughness number into a fictional loss estimate.
Identify which copper surface the data describe
Use the conductor comparison below to identify the foil surface, treatment, roughness metric and manufacturing stage before entering model data. The worksheet then provides an ideal skin-depth reference; it does not extract a foil-loss model.
| Record | What it resolves | Why a short label is insufficient |
|---|---|---|
| Foil and treatment | Identifies the supplied conductor system | Similar marketing categories may include different products |
| Measured side | Locates the surface being characterized | The two interfaces need not be equivalent |
| Metric and method | Explains how the number was obtained | Different metrics cannot be substituted without justification |
| Construction | Places the foil in the actual transmission line | Geometry changes conductor participation |
| Model | States how roughness enters the calculation | Model parameters are not universal roughness conversions |
Use skin depth as a scale reference
The worksheet runs locally in your browser with JavaScript. You can contact 勛圖tv directly without completing it.
Ideal copper skin-depth reference
The expression is delta equals the square root of one divided by pi times frequency, permeability and conductivity. The tool uses permeability equal to the free-space value and a conductivity of 58 million siemens per metre as an idealized room-temperature copper assumption. Real processed conductors and temperature conditions can differ.
Specify the conductor before comparing two loss curves
A dielectric-only comparison can mislead when the two boards use different copper foils or surface treatments. The field interacts with the conductor surfaces and the dielectric boundary, so conductor condition belongs in the transmission-line model. Start by identifying the foil option used on the relevant layer, not merely the copper weight or the laminate brand.
Ask which surface is described by a roughness value and at which manufacturing stage. A foil can have different characteristics on its two sides, and subsequent treatment can matter. Preserve the measurement metric and method. Ra, Rq and Rz are different descriptors; a number without its metric is not a complete model input.
Do not assume that a marketing label such as very low profile uniquely determines a solver's roughness parameters. Different models require different inputs, and an arbitrary conversion can produce misleading confidence. Request the supplier's appropriate model information or disclose the assumption. Unknown conductor data should remain visible in the loss budget.
This is not an instruction to choose the smoothest available foil for every board. Adhesion, handling, processing compatibility, availability and cost also belong in the construction review. The useful comparison holds the circuit requirement constant and evaluates a complete approved construction, including the conductor option that the manufacturer can actually supply.
Use skin-depth scale without turning it into a universal cutoff
For an ideal smooth conductor, skin depth decreases as frequency increases. Using copper conductivity of 5.8 times ten to the seventh siemens per metre and free-space permeability, the familiar good-conductor expression gives approximately 2.09 micrometres at 1 GHz and 0.66 micrometres at 10 GHz. These are calculated reference scales under stated assumptions, not measured foil properties.
The calculation explains why micrometre-scale surface features can matter at microwave frequencies. It does not establish a universal frequency above which a particular foil becomes unacceptable. The relevant geometry, roughness model, route length and allowed loss all matter. A short interconnect may tolerate an effect that a long channel cannot.
Do not compare a quoted peak-to-valley roughness directly with skin depth and infer an exact attenuation multiplier. The physical surface and the selected model contain more information than one height ratio. A model should be used within its documented assumptions and checked against representative evidence where the design margin depends on it.
The provided illustration plots the ideal skin-depth relationship only. It deliberately does not plot invented measured loss for commercial foils. To compare actual candidates, place their documented model inputs and a consistent geometry into the same analysis, then identify the remaining uncertainty rather than labeling the smoother candidate automatically verified.
Separate dielectric, conductor and fixture contributions
A measured insertion-loss curve includes the complete measurement path unless the setup removes specified fixtures. Connector launches, line sections, transitions and calibration choices can all contribute. If two test boards differ in more than foil, the observed difference cannot automatically be assigned to roughness.
Define a comparison with the same route length, cross-section, dielectric system, finish and reference planes. Record any unavoidable differences. If an extraction method is used, retain its assumptions and the raw measurements. A fitted model can be useful, but it should not hide a change in geometry or fixture behavior inside an adjusted dielectric value.
For a channel budget, allocate uncertainty as well as nominal loss. Suppose a proposed foil change is expected to improve a modeled route by a small amount, while the uncharacterized connector launches have a larger uncertainty. Improving the launch evidence may be a better immediate investment than optimizing the foil to an unsupported decimal place. This is a decision principle, not a numerical performance claim.
Keep phase behavior in view when relevant. A conductor model can affect more than amplitude, and a material parameter derived from circuit behavior may already incorporate some physical effects. Check parameter compatibility before combining vendor design Dk, fitted loss and a separate roughness correction in one model.
Make foil changes visible to purchasing and fabrication
The release package should name the foil option or the accepted requirement, identify the critical layers and define how alternatives are reviewed. A generic note saying low-loss material does not preserve the conductor assumptions used in simulation. Include the relevant model source and construction revision with the engineering record.
The comparison worksheet has fields for original and proposed foil, roughness metric, treatment, material pairing, model parameters and affected channels. It also asks what evidence supports the claimed benefit. A blank measurement field does not block preliminary planning, but it prevents the proposal from being misrepresented as a verified improvement.
For fabrication review, ask whether the requested conductor option is compatible with the chosen laminate, bonding and process. For quotation, compare the entire scope including coupons and any requested RF characterization. A material upgrade without matching verification may leave the original uncertainty unresolved.
Use the signal-integrity worksheet to describe the route and the review sheet to make conductor assumptions explicit. It is not a replacement for a field solver or a guarantee that a smoother foil produces a specified loss on the finished board. The useful conversion is an enquiry containing the actual channel and construction question.
Keep roughness and dielectric interpretation connected
A propagation measurement can reflect both dielectric and conductor effects. If a material value is extracted from a circuit with particular copper, it may not transfer unchanged into a solver that separately applies another conductor correction. Review the supplier's model guidance and the software's definitions before combining them.
Rogers demonstrates the influence of copper surface condition on circuit-derived dielectric interpretation in its . The engineering lesson is to keep the extraction and conductor model consistent, not to assign a fixed Dk offset to every rough foil.
For the dielectric terminology, use design Dk versus process Dk. This page owns the copper data and channel-comparison task rather than duplicating a complete laminate-selection catalogue.
Choose a model that matches the available evidence
Some roughness models use a correction based on a surface descriptor; others use a more detailed representation. Identify the model and the meaning of each parameter. A software field with units of length does not mean any published roughness length is a valid input.
Where supplier-supported parameters are available for the exact foil and construction, preserve their provenance. Where parameters are fitted from measurements, document the extraction and valid range. An excellent fit over a narrow band should not be presented as proof of accuracy for a different geometry or much higher frequency.
Keep a smooth-conductor baseline when it helps separate contributions, but label it as an idealized comparison. Do not use it as a manufacturing option unless an actual specified construction supports the assumption. The purpose of the baseline is understanding the model, not promising a physically perfect copper surface.
Compare channels with matching boundaries
Specify the transmission-line geometry, dielectric, copper, finish and length. Identify the ports and reference planes. A measured insertion-loss curve can include connectors, launches and fixtures, so its boundary must match the simulated or supplier data used for comparison.
If using multiple line lengths to investigate distributed behavior, keep their cross-sections and processing representative and apply a justified extraction method. Simply subtracting unrelated measurements can include mismatched launches or different calibration errors. Save the raw data and processing details with the coupon drawing.
For a high-speed digital channel, evaluate the frequency content relevant to the interface and its rise times, not only the clock label. For a narrowband RF path, use the operating band and appropriate out-of-band requirements. The test span should follow the circuit task rather than a convenient instrument default.
A useful foil comparison experiment
Suppose two otherwise comparable builds use different documented copper options. A controlled comparison preserves the stackup, routed geometry, finish, connectors and measurement procedure as closely as practical. Record the actual construction evidence and distinguish intended differences from process variation.
If the measured channel loss changes, do not attribute the entire difference to roughness without checking thickness, geometry, dielectric and interfaces. The comparison may still support a practical material choice, but the strength of the causal conclusion depends on how well those other variables were controlled.
Use the result to refine the design and sourcing record. A purchaser needs the exact approved copper option and an alternative-approval process, not just a statement that smooth copper is better. This is particularly important when a laminate substitution changes the foil silently.
Balance electrical goals with manufacturing needs
The conductor system must also support the selected fabrication and reliability requirements. Adhesion, handling, processing and availability should be reviewed with the material supplier and fabricator. An electrical model alone does not approve the laminate process or the resulting board reliability.
Surface finish deserves separate attention. Its electrical relevance depends on where it is applied, the geometry and frequency. Do not assume that every exposed RF feature should use the same finish decision as an ordinary solder pad. Identify the actual construction in the model and request a project-specific review where needed.
The RF PCB design guide connects these choices to the line and transition geometry. RF PCB manufacturing is the commercial route for agreeing a construction. Broader build requirements remain with PCB Manufacturing.
Write a copper requirement that can be quoted
Include the laminate grade, copper option, finished conductor requirements, relevant surface data and any model-specific parameters supplied by the vendor. Identify which RF or channel features are sensitive and what verification is requested. Separate a typical material descriptor from a contractual acceptance limit.
Ask for proposed alternatives to be identified before release. If the original foil is unavailable, compare the alternative's construction and model data instead of accepting a nominally similar label. Include any prototype or coupon work needed to evaluate the change in the quotation scope.
You can request a conductor and stackup review with the model assumptions and measurements available. The skin-depth worksheet is a supporting reference, not a substitute for a channel-loss specification or a verified manufacturing capability.
Copper and loss questions
Does low-profile copper mean zero conductor loss?
No. Finite conductivity and the actual current distribution still contribute to loss, and the complete channel has other mechanisms. A low-profile designation needs a specific foil and data source. It does not establish a universal insertion-loss figure.
Can Ra and Rq be entered into the same model field?
Only if the model and data conversion explicitly support that use. They are different metrics. Keep the measurement definition and the model parameter definition together rather than assuming equal units make the quantities interchangeable.
Can a coupon prove every trace on the board?
A coupon provides evidence for its defined structure and conditions. Its relevance depends on how well it represents the production path. Different transitions, geometry or surface processing may require additional assessment. State the scope of the coupon result in the release record.
What should be retained for a repeat build?
Retain the approved foil and laminate construction, the drawing and stackup revisions, the model assumptions and the measurement method. Include any accepted alternatives explicitly. This allows a repeat-order review to identify a real construction change instead of relying on a generic material description.
Engineering guide 繚 Content prepared 1 October 2026. Diagram and example labels distinguish explanations from product or test evidence.