Skip to main content
Insertion Loss and Trace Width: A Practical Guide for High-Frequency PCB Design, featured image

Insertion Loss and Trace Width: A Practical Guide for High-Frequency PCB Design

GSAS Engineering · · 7 min read

# Insertion Loss and Trace Width: A Practical Guide for High-Frequency PCB Design

At low frequencies, a PCB trace is a wire. At high frequencies, it is a transmission line, and transmission lines lose energy. This energy loss, quantified as insertion loss in decibels per unit length, determines whether a signal arrives with enough amplitude to be correctly received. For Indian teams designing 5G infrastructure, military radar, satellite communications, and high-speed serial links, managing insertion loss is the difference between a board that works and one that does not.

As Indian industry pushes into mmWave 5G (Jio, Airtel), X-band radar (Indian defence agencies), and interfaces like PCIe Gen5 and 112G SerDes, the design choices that control insertion loss, trace width, copper roughness, dielectric material, and stackup geometry, must be made early, informed by simulation rather than rules of thumb. This article provides a practical guide to the physics, design parameters, and Siemens EDA tools, HyperLynx SI and Z Planner, that enable Indian teams to optimise their high-frequency designs.

Why Insertion Loss Matters

A signal travelling through a PCB trace loses amplitude as it propagates. If insertion loss is too high, the eye diagram closes, voltage and timing margins shrink until the receiver cannot distinguish logic levels. At 1 Gbps on a 15 cm trace, FR-4 is manageable. At 10 Gbps, insertion loss becomes a significant constraint. At 28 Gbps, standard FR-4 may make the channel non-functional.

High-frequency PCB design requires explicit management of insertion loss through material selection, trace geometry, and stackup design, decisions that must be made during board planning and validated through simulation before routing begins.

The Physics of Insertion Loss

Insertion loss in a PCB trace has three primary components, each with distinct frequency dependence:

Conductor Loss (Skin Effect)

At DC, current flows uniformly through the trace cross-section. As frequency increases, current crowds toward the surface, the skin effect. At 1 GHz, skin depth is approximately 2.1 micrometres; at 10 GHz, approximately 0.66 micrometres. Current flows through an increasingly thin shell, effectively increasing resistance.

Wider traces have lower conductor loss because the current-carrying perimeter is longer. A 200-micrometre trace has roughly twice the conducting perimeter of a 100-micrometre trace, and therefore roughly half the conductor loss. However, wider traces require thicker dielectric to maintain impedance, increasing board thickness, a tradeoff requiring simulation to optimise.

Dielectric Loss

The dielectric material absorbs electromagnetic energy, characterised by its dissipation factor (Df). FR-4 has Df of approximately 0.02 at 1 GHz. Low-loss materials like Megtron 6 achieve 0.002 or less, an order of magnitude improvement.

Dielectric loss increases approximately linearly with frequency and overtakes conductor loss between 3-8 GHz for most configurations. Above a few gigahertz, material selection has more impact on insertion loss than copper geometry. Indian teams designing 5G mmWave, Ku-band satellite links, or 56G/112G SerDes channels cannot use standard FR-4.

Copper Roughness Loss

This loss mechanism catches many engineers by surprise. Copper foil has a deliberate roughness profile for dielectric adhesion, invisible at low frequencies but significant when skin depth approaches roughness height. Roughness is characterised by the Rz value (average peak-to-valley height):

  • Standard (STD) copper: Rz approximately 6-10 micrometres
  • Low-profile (LP) copper: Rz approximately 3-5 micrometres
  • Very low-profile (VLP) copper: Rz approximately 1.5-3 micrometres
  • Hyper very low-profile (HVLP) copper: Rz approximately 1-2 micrometres

At 10 GHz, when skin depth (0.66 micrometres) is comparable to roughness features, current follows the rough contour, increasing path length and resistance. The difference between standard and VLP copper can be 0.5-1.0 dB per inch at 10 GHz, significant over a 10-20 inch channel. For Indian teams designing above 10 GHz, copper roughness is a primary design variable.

Trace Width: The Direct Lever

Trace width is the most directly controllable parameter affecting conductor loss. Halving trace width roughly doubles conductor loss. But width cannot be considered in isolation, trace width and dielectric thickness together determine characteristic impedance (typically 50 ohms single-ended, 85-100 ohms differential).

This creates a design optimisation problem:

  • Narrower trace, thinner dielectric: Higher conductor loss, but thinner stackup, more routing layers possible
  • Wider trace, thicker dielectric: Lower conductor loss, but thicker stackup, fewer routing layers

The optimal trace width depends on target frequency, channel length, board thickness constraints, available materials, and loss budget, precisely the multi-variable optimisation that simulation tools solve.

Dielectric Material Selection

For Indian teams working at frequencies above 3-5 GHz, the choice of dielectric material is the single most impactful decision for insertion loss management. The spectrum of available materials ranges from standard FR-4 to ultra-low-loss materials, with corresponding differences in performance and cost:

Standard FR-4 (Dk ~4.3, Df ~0.02 at 1 GHz). Adequate for designs operating below 2-3 GHz. Widely available from Indian PCB fabricators. Cost-effective. Not suitable for high-frequency or long-channel high-speed serial links. Mid-loss materials (Dk ~3.6-3.8, Df ~0.008-0.012). Materials like Megtron 4 and similar mid-loss laminates provide a significant improvement over FR-4 at a moderate cost premium. Suitable for PCIe Gen4, 25G Ethernet, and similar mid-frequency applications. Low-loss materials (Dk ~3.3-3.6, Df ~0.002-0.005). Materials like Megtron 6 and Rogers RO4000 series are used for demanding high-frequency applications. Required for PCIe Gen5, 56G SerDes, and 5G sub-6 GHz systems. Available from specialised fabricators. Ultra-low-loss materials (Dk ~3.0-3.5, Df <0.002). Materials like Rogers RO3000 series and similar PTFE-based laminates are used for mmWave, radar, and satellite applications. Required for 28 GHz 5G, X-band radar, and Ku-band satellite systems. Limited fabricator availability in India, most Indian teams source these boards from specialised international fabricators or the few domestic facilities with mmWave capability.

Material choice must be validated through simulation. HyperLynx SI models frequency-dependent dielectric properties, allowing engineers to compare materials and select the most cost-effective option meeting the loss budget.

Z Planner: Stackup Optimisation

Z Planner allows engineers to define the PCB stackup, copper weights, dielectric materials, thicknesses, and roughness profiles, and calculate resulting impedance and insertion loss. It addresses the fundamental problem of balancing impedance targets with loss budgets:

  • Characteristic impedance for single-ended and differential traces at specified widths and spacings
  • Insertion loss per unit length as a function of frequency, accounting for conductor loss, dielectric loss, and copper roughness
  • Stackup thickness and its impact on board manufacturability and mechanical requirements

An Indian team designing a 5G base station backplane, for example, might use Z Planner to evaluate several stackup options:

  • Standard FR-4 with 1 oz copper and standard roughness, lowest cost, highest loss
  • Megtron 6 with 1 oz LP copper, moderate cost, significantly lower loss
  • Megtron 6 with 0.5 oz HVLP copper and wider traces, higher cost, lowest loss

Z Planner quantifies the insertion loss for each option across the frequency range of interest, enabling an informed engineering decision rather than a guess.

HyperLynx SI: Channel Simulation

HyperLynx SI provides the simulation engine for validating high-frequency signal integrity on the actual PCB layout. Key capabilities for insertion loss management include:

S-Parameter Analysis

S-parameters characterise frequency-dependent behaviour including insertion loss (S21), return loss (S11), and crosstalk. HyperLynx SI extracts S-parameters from PCB layout geometry, accounting for trace width, dielectric properties, copper roughness, via transitions, and connectors.

For Indian radar and satellite teams, S-parameter analysis is the primary validation tool. An X-band radar front-end must maintain insertion loss across 8-12 GHz, HyperLynx SI models these transitions using actual geometry and material properties.

Channel Simulation

For high-speed serial links, HyperLynx SI cascades S-parameters of each channel segment, traces, vias, connectors, cables, and evaluates overall performance against the interface specification, accounting for all loss mechanisms and producing a realistic loss budget assessment.

Eye Diagrams with Realistic Loss Models

HyperLynx SI generates eye diagrams incorporating frequency-dependent insertion loss and transmitter/receiver equalisation (pre-emphasis, CTLE, DFE). For Indian teams designing PCIe Gen5 or 100G Ethernet, this simulation confirms the design will work, or identifies problems while changes are still possible.

Indian Applications

The need for insertion loss management spans several critical Indian industry segments:

5G Infrastructure. Jio and Airtel are deploying mmWave bands at 26-28 GHz. Base station antenna feed networks, beamforming arrays, and backhaul interfaces all require insertion loss management at frequencies where FR-4 is inadequate. Military Radar. Indian defence agencies and OEMs develop radar systems from L-band through Ku-band. At X-band and above, every decibel of feed network loss reduces detection range. Material selection, trace geometry, and stackup must minimise loss at the operating frequency. Satellite Communications. ISRO’s programmes, including NavIC and broadband satellite constellations, require ground terminal and payload electronics achieving specified insertion loss targets across L-band through Ku-band. High-Speed Digital. Indian companies working with advanced processors encounter PCIe Gen4/Gen5, DDR5, USB4, and 100G/400G Ethernet, where insertion loss management is essential for first-pass success.

A Methodology, Not Just a Tool

Managing insertion loss in high-frequency PCB design is not a single analysis step, it is a methodology that spans the entire design process:

1. Requirements definition: Establish the loss budget for each critical channel based on the interface specification and the transmitter/receiver capabilities
2. Material selection: Use Z Planner to evaluate dielectric materials and copper roughness profiles against the loss budget
3. Stackup optimisation: Use Z Planner to determine trace widths, dielectric thicknesses, and layer assignments that meet both impedance and loss targets
4. Layout implementation: Route critical channels in Xpedition with the constraints defined during stackup optimisation
5. Channel validation: Use HyperLynx SI to extract S-parameters, simulate the complete channel, and generate eye diagrams
6. Iteration: If the channel does not meet its loss budget, adjust the material, stackup, or routing, informed by simulation rather than guesswork

This methodology ensures that insertion loss is managed proactively throughout the design, rather than discovered as a problem during board bring-up.

Getting Started with GSAS

GSAS Micro Systems provides high-frequency design methodology support for Indian teams, from stackup planning and material selection through channel simulation and validation.

Contact GSAS Micro Systems today for high-frequency design methodology support. Whether you are designing 5G infrastructure, military radar, satellite communications, or high-speed digital systems, GSAS can help you establish the insertion loss management methodology that ensures first-pass success. Our field application engineers are based in Bengaluru, Hyderabad, Chennai, Coimbatore, Pune, and Delhi NCR. Reach us through gsasindia.com or visit our offices in Bengaluru, Chennai, or Vadodara.

Interested in Siemens EDA tools?

Talk to our application engineers for personalized tool recommendations.

Stay in the Loop

Get monthly compliance updates, product insights, and engineering best practices delivered to your inbox.

Related Articles

Embedded engineering bench in India, illustrating the hardware a developer can act on today, supported by GSAS
Technical Guides Arm

Arm's Agentic AI Push: What Indian Embedded Teams Can Actually Use Today

Arm has set out its platform strategy for the agentic era, putting inference everywhere it has compute. Indian embedded teams can start on that direction this quarter with the Cortex-M toolchain they already own: Helium, CMSIS-NN, Ethos-U and Vela, on boards you can order today.

13 Sept 2026 · 7 min read
Side by side comparison of a 10BASE-T1S multidrop mixing segment, one balanced pair with four nodes on short stubs and a termination at each end, against a point to point star of four separate links into switch ports, from GSAS Micro Systems India
Automotive Ethernet Automotive & Mobility

10BASE-T1S and PLCA: Multidrop Ethernet Explained

10BASE-T1S is the one member of the T1 single-pair Ethernet family that keeps a shared medium, and PLCA is the reconciliation sublayer that stops the nodes on it from colliding. This article covers what IEEE 802.3cg standardises, how the beacon and transmit opportunities schedule a cycle, the node count and segment length figures the OPEN Alliance interoperability test suite works to, and the failure modes that put a segment quietly back into contention while every link still looks up. Written by the GSAS Micro Systems engineering team in India for teams bringing up multidrop segments on the bench.

29 Aug 2026 · 12 min read
The bench to harness adapter path drawn left to right: the ECU connector, a test lead, a media converter or pluggable T1 module, RJ45, and the host, showing where each connector family sits between the device under test and the laptop, from GSAS Micro Systems India
Automotive Ethernet Automotive & Mobility

Automotive Ethernet Connectors: H-MTD, MATEnet, MQS

An ECU arrives on the bench with a connector nobody has a mate for, and the day is gone. Search the family names and you get connector product pages that describe their own part and stop there. This article puts the families side by side in one table using only what their public pages state, then makes the point those pages leave out: the IEEE link segment definition, not the connector, is what sets reach and loss limits, and shielding is a channel decision rather than a preference. Written by the GSAS Micro Systems engineering team in India.

29 Aug 2026 · 12 min read