Low Loss Laminate Selection For High Speed Boards

The laminate in a high speed design is no longer just a mechanical carrier with a dielectric constant. Its loss characteristics set how far a signal can travel before it needs to be retimed, and the choice between a standard FR-4 and a low loss material can decide whether a link closes at all.

Material selection is therefore a system decision rather than a purchasing one. The dissipation factor, the dielectric constant, the glass weave and the price all interact, and improving one of them usually costs something elsewhere.

What Makes A Material Low Loss

Loss in a laminate comes from the polar groups in the resin, which absorb energy as the electric field alternates. A resin with fewer polar groups, such as a modified epoxy, a cyanate ester or a hydrocarbon, absorbs less and therefore has a lower dissipation factor.

The filler system and the glass also contribute. A material with a high filler content can have a lower expansion and better dimensional stability, at the cost of a slightly different dielectric behaviour.

Dissipation Factor And Insertion Loss

The dissipation factor is quoted at a specific frequency, usually one or ten gigahertz, and it is the parameter that most directly predicts dielectric loss. Standard FR-4 sits around 0.015 to 0.020, a mid loss material around 0.008, and a low loss material below 0.005.

Dielectric loss scales with frequency and with the length of the trace, so the benefit of a better material grows as the link gets longer or faster. On a very short interconnect a low loss material may make no measurable difference at all.

Dielectric Constant And Impedance

The dielectric constant sets the trace width needed for a given impedance and the propagation delay along the line. A lower dielectric constant allows a wider trace for the same impedance, which reduces conductor loss and makes etching easier.

The value also varies with frequency and with the resin content of the particular prepreg. A material with a nominal constant of 3.7 may measure 3.9 at one gigahertz, and the design should use the value at the frequency of interest rather than the data sheet headline.

Glass Weave And Skew

The glass fabric in a laminate is woven, and the dielectric constant over a glass bundle is higher than over the resin rich area between bundles. A differential pair whose two traces sit over different parts of the weave then travels at slightly different speeds.

The result is skew, which eats into the timing budget of a differential link. Spread glass, where the weave is flattened and randomised, reduces the effect, and it is often specified together with a low loss resin for a high speed design.

Temperature And Frequency Behaviour

Loss and dielectric constant both change with temperature. A material that performs well at 25 C may be noticeably worse at 85 C, and the specification should be checked across the operating range rather than at room temperature only.

Frequency behaviour matters as well. Some materials have a dissipation factor that rises sharply above a certain frequency, which makes a data sheet figure quoted at one gigahertz misleading for a design operating at ten times that value.

Cost And Availability

Cost rises steeply as the loss falls. A low loss material can cost several times as much as standard FR-4 for the same area, and the difference is magnified by the minimum order quantity and by the lead time.

Availability is a practical constraint too. A material that is only made in one place and in a limited range of thicknesses creates a supply risk, and a design that depends on it should have a qualified alternative identified in advance.

Choosing The Right Grade

The starting point is a loss budget for the longest and fastest link, calculated with the dielectric constant and the dissipation factor of the candidate materials. If standard FR-4 closes the link with margin, there is no reason to move.

Where the budget is tight, the options in order of cost are usually a mid loss material, a low loss material and finally a very low loss material with spread glass. Each step should be justified by the calculation rather than by a preference for better material.

Process Compatibility

A low loss material often has a higher glass transition temperature and a different resin flow, which changes the lamination cycle and the drilling parameters. Some materials are more brittle and need a different drilling recipe to avoid cracked barrels.

Those differences have to be discussed with the fabricator before the stackup is released. A material that is technically ideal but cannot be processed on the available equipment is not a usable choice. The stackup documentation is described under layer stackup from one to eight layers.

Qualification And Verification

Qualification is by measurement on a coupon built from the actual material, using a known trace geometry and a network analyser. The measured insertion loss is compared with the simulation to confirm that the material data used in the model was accurate.

The coupon should also be used to check the impedance, since the dielectric constant of the finished laminate can differ from the data sheet. The measurement practice is described under impedance coupon design.

Process Control and Verification

Running a first article through the same checks as the production panel confirms that the two agree, and that comparison is the cheapest form of process control available at prototype stage. Keeping a sample from the panel turns a dispute into a measurement, because the same coupon can be re-examined by both parties without rebuilding the batch.

Related reading: our fabrication notes, board quality and design release notes cover the same ground.

A first article check confirms that the process and the drawing agree on the points listed above, and that the coupon data supports the values used in the design.

Where a measurement falls outside the expected window, the sample is retained so that the cause can be established before the balance of the batch is released.

FAQ

Is low loss material needed for every high speed design? No. It is needed when the loss budget cannot be met with a cheaper material. A short link at a moderate data rate is often served perfectly well by a mid loss laminate.

Does a lower dissipation factor always reduce loss? It reduces the dielectric part of the loss. If the conductor loss dominates, because the traces are narrow or the foil is rough, the improvement will be smaller than expected.

Can a low loss material be mixed into a stack? Yes, and it is a common way to control cost. The high speed layers use the better material while the power and low speed layers use standard FR-4, provided the lamination is compatible.