Reading time ( words)
Signal loss (attenuation) in transmission lines can have a variety of causes.
- Radiative loss
- Coupling to adjacent traces
- Impedance mismatches
- Conductor loss
- Dielectric loss
According to Reference 1, radiative loss is only a small component of loss contributors. Coupling to adjacent lines, on the other hand, can cause signal rise time degradation. It can be modeled quite accurately and proper circuit layout can minimize it.
Impedance discontinuities (i.e., variations or changes) along the transmission line cause signal degradation. The cross section of the conductor should not change along the transmission line. Such changes affect the impedance. They can be caused by plating height variation, or etch non-uniformity, or localized imperfections such as conductor “dish down” or “mouse bites.”
Less understood is the change of the properties of the dielectric that contributes to impedance mismatch, namely the dielectric constant.
Typical dielectric structures are glass weave, reinforced epoxy resins. Both the dielectric constant of the resin as well as that of the glass are bulk properties of the materials and are quite constant throughout the material. The dielectric constant of FR-4 resin is about 3.2 and the dielectric constant of glass is about 5.6. Since the glass weave reinforcement has a pattern, there are glass-rich and glass-lean domains in the dielectric. Therefore there are domains of higher and lower dielectric constant surrounding the signal path, which means the impedance changes along the signal path (i.e., there are impedance discontinuities). If a transmission line runs parallel, and on top of a glass fiber strand, or along a resin-rich area between glass fibers, then the dielectric environments of these two lines are noticeably different, which unfortunately is the case most of the time as laminate is sheared parallel to the weave pattern, and conductor lines tend to run in the same direction. Angling the transmission lines relative to the weave pattern will improve the situation by randomizing the glass/ resin concentration differences along the transmission path. References 2 and 3 have studied this problem area.
Read the full article here.
Editor's Note: This column originally appeared in the January 2015 issue of The PCB Magazine.