TF-1/2 PTFE laminate: identity and RF comparison

Read the property, the test method and the condition as one record. A dielectric constant without frequency, temperature or method can be ambiguous. A typical value is also different from a purchasing tolerance. The table on this page separates supplier process Dk from design Dk so that an engineer does not accidentally transfer a quality-control measurement into a circuit model with a different definition.

Effective permittivity belongs to the transmission-line structure. For an air-covered microstrip, some of the field occupies air and some occupies the substrate. For a homogeneous centered stripline, the field arrangement is different. Neither quantity should be casually relabeled as the material's process Dk. The same laminate can therefore appear in two correct calculations with different effective permittivities because the geometries differ.

Preserve document revisions when recording values. A supplier may publish a design reference that depends on thickness, copper configuration, frequency range or a particular extraction method. The presence of a newer document does not mean that an old approved circuit should be changed without review. Compare the relevant construction and determine whether the design model, procurement specification or acceptance record needs revision.

A useful material comparison holds the circuit task constant. Begin with the same electrical length or board geometry, then state which variable is allowed to change. Comparing one material at a different thickness, foil type and fixture can obscure the reason for a performance difference. Keep the comparison small enough to answer a defined question, such as trace-width change at equal impedance or dielectric attenuation under an explicitly constant-Df assumption.

When the material is selected, carry the chosen construction into the fabrication drawing. Specify the supplier grade, core thickness, copper type, finish, bond layers and critical RF dimensions. Leave unresolved items visibly open for engineering review. This page is the data interpretation step; the linked manufacturing page carries the selected material into a quote-ready board description.

Establish the local cross-section before choosing a trace width. Record the dielectric height, material model, copper thickness and reference-plane arrangement. A total board thickness is not a valid substitute for the distance to the reference plane unless the structure is truly a two-layer core with that definition. For stripline, check whether the trace is centered and whether the dielectric is homogeneous on both sides.

Use a preliminary calculation to estimate scale, then evaluate the actual construction in a suitable solver. The ideal line model provides a reproducible starting point, but it does not include solder mask, trapezoidal etch shape, conductor roughness, dispersion or a three-dimensional connector launch. A calculator output with many decimal places is still only as useful as the assumptions that generated it.

Translate manufacturing variation into circuit questions. A width tolerance, dielectric-height tolerance and Dk variation do not necessarily combine as a single worst-case impedance number for every circuit. Establish which variables can move together and which are independently controlled. For a filter, sensitivity of resonator length and coupling gap may be more useful than one line-impedance estimate. For an array, phase tracking across paths can matter more than a nominal delay value.

Keep a continuous and intentional return structure. Slots, voids, antipads and layer changes can alter the fields even when the trace width remains constant. Where a signal transitions to another layer, review the return connection and launch geometry. Where a metal enclosure or shield is close to the line, include it in the final electromagnetic model. The schematic net name does not describe those physical boundaries.

Release a drawing that links the electrical model to measurable construction. Identify critical dimensions, stackup revision, allowed substitutions and the test method. If a prototype leads to a geometry adjustment, record the reason and the revised model rather than allowing the change to remain an undocumented fabrication compensation. That record makes later procurement and repeat production substantially easier to compare.

Where the cost changes, and which savings are real

For a small RF order, setup, procurement and test preparation may matter as much as the area of laminate. A panel that holds more circuits spreads some shared effort across more boards, but only if the orientation, tooling and coupon requirements permit that layout. The budget tool uses an explicit rectangular nesting model so that its area assumption can be inspected. It does not assume every visible square millimeter of a panel becomes a delivered board.

Copper foil, core thickness and bonding system are part of the purchased construction. A quotation based on a different thickness or foil is not a clean price comparison, even when the headline material name matches. Ask whether the offered material is standard for the supplier's available construction and whether minimum purchases, special handling or qualification work affect the lot. Availability and lead time must be confirmed for the actual order rather than inferred from a web page.

Complexity can move cost in several directions. A hybrid stack may reduce the amount of premium RF laminate while adding press planning and interface review. Metal backing can improve a thermal path while adding bonding and machining requirements. Extra coupons or RF fixtures can reduce uncertainty but add material, setup and measurement work. Compare the complete project scope instead of ranking designs by laminate price alone.

The optional numerical budget uses cost allowances entered by the visitor. It contains no calibrated RAYPCB tariff and does not claim current market pricing. The model combines the costed panels, one-time setup, batch testing and per-board processing; assembly is included only when selected and entered. Shipping, taxes, process yield loss and unentered qualification are excluded. A zero value means already included or deliberately zero, not unknown.

When geometry, quantity or supply scope changes, revisit the allowances. A per-board processing cost from one batch may not apply to another. The tool clears old allowances on those changes to avoid presenting a stale total. Hybrid, metal-backed, unresolved and more complex multilayer constructions go to engineering quotation. The downloaded brief remains useful because it carries the requirements and assumptions even when a numerical price would be misleading.