How the Teflon Inside an Aging Coax Slowly Drifts the Impedance It Was Trusted to Hold

A good coaxial cable is supposed to be a constant. You buy it rated at fifty ohms, you install it, and you trust that number for years, building matching networks and calibrations on the assumption that the cable will present the same impedance tomorrow that it did the day it was made. For most of a cable's life that trust is well placed. But the white insulator at the heart of a high-performance coax, polytetrafluoroethylene, is not the inert eternal solid its reputation suggests. It creeps under pressure, it lurches through a crystal transition right at room temperature, and over years it settles and flows in ways that quietly shift its dimensions and its dielectric constant. Because the characteristic impedance of a coax depends directly on both, the cable's defining number drifts, slowly and invisibly, and the calibration that was perfect at installation develops an error no one can see by looking.

The surprise for most builders is that the chemically inert material, the one prized for shrugging off acids and solvents and temperature extremes, is mechanically restless. Its molecules are slippery, its crystalline structure rearranges at the wrong temperature, and it lacks the stiffness to resist sustained pressure. The very properties that make it an electrical jewel, low loss and low permittivity, come bundled with a mechanical softness that makes it dimensionally unreliable over the long haul. Understanding how that softness translates into impedance drift turns an unsettling mystery into a quantifiable, manageable effect.

What sets the impedance and why two soft variables control it

The characteristic impedance of a coaxial line is fixed by its geometry and its dielectric through a clean formula:

Z0 = (138 / sqrt(epsilon_r)) * log10(D / d)

where D is the inner diameter of the outer conductor, d is the diameter of the inner conductor, and epsilon_r is the relative permittivity of the dielectric filling the space between them. Three quantities set the impedance, and two of them, the dielectric constant and the spacing ratio, are governed by the polytetrafluoroethylene. The dielectric constant appears directly under a square root. The diameter ratio depends on the dielectric holding the inner conductor concentric and at the right radial position. If the dielectric changes its permittivity, or lets the conductor spacing shift, the impedance moves.

This is the vulnerability. The dielectric is not a passive spacer but an active determinant of the impedance, and it is made of a material that changes. A one percent change in the dielectric constant shifts the impedance by about half a percent through the square root, and a change in the geometry from a flowing or settling dielectric shifts it through the logarithm. Neither change announces itself. The cable looks identical; only an impedance measurement reveals the drift.

The room-temperature lurch that no other plastic has

Polytetrafluoroethylene hides a peculiar trap precisely in the range where equipment lives. At approximately 19 degrees Celsius the material undergoes a crystalline phase transition, a rearrangement of its molecular packing that produces an abrupt change in dimensions. This is not a gentle thermal expansion but a near step-function change. The published figure is striking: the phase change equates to about a 1.5 percent volumetric change, with a corresponding change in dielectric constant, producing an abrupt change in the cable's electrical length right around room temperature.

A 1.5 percent volume change is enormous for a dimension-critical component, and its placement at 19 degrees is the cruelest possible location, because that is squarely within the temperature range a cable experiences indoors and out. A cable warming through 19 degrees on a sunny morning crosses the transition and its dielectric lurches; cooling back through it that evening, the dielectric lurches back, but not along the same path. The transition shows hysteresis, beginning at different temperatures for rising versus falling temperature, so the cable does not retrace its steps. After a temperature excursion through the transition, the dielectric can settle at a slightly different state than it started, and repeated cycling can ratchet the material toward a new configuration. This is one route by which aging happens: a cable that has been thermally cycled through 19 degrees many times is not dimensionally identical to the cable that left the factory.

Cold flow, the slow creep that never stops

The second aging mechanism is mechanical and even slower. Polytetrafluoroethylene has poor creep resistance, meaning it deforms permanently under sustained pressure rather than springing back. Pure material has a hardness around Shore 55 and cannot withstand high pressure without deforming, and gaskets made from it are notorious for plastic deformation or cold flow when tightened. Inside a coax, the dielectric is under constant mechanical stress from the conductors pressing on it, from bends that squeeze it, from connectors that clamp it, and from its own weight in a long vertical run.

Under that sustained stress the material slowly flows. The technical literature on these cables describes the mechanism plainly: the dielectric flows under pressure over time, and if it cannot change diameter because the outer conductor confines it, the material tries to expand lengthways instead. So a cable under radial pressure gradually redistributes its dielectric, thinning where it is squeezed and bulging where it can, subtly changing the conductor spacing and the uniformity of the fill. Every such change moves the local impedance, and because the flow is uneven along the cable, it creates small impedance variations that show up as structural return loss, reflections born not of a fault but of the dielectric slowly rearranging itself. Cables with the lowest-density dielectric, the expanded polytetrafluoroethylene prized for the lowest loss and lowest permittivity, are the most vulnerable, because the airy low-density structure offers the least mechanical support and is described as extremely prone to mechanical damage and degradation.

A numerical estimate of the impedance drift from a permittivity shift

Put numbers to how much these mechanisms move the impedance. Standard full-density polytetrafluoroethylene has a relative permittivity of about 2.01, while low-density expanded versions run around 1.73 and ultra-low-density around 1.42. Start with a cable using full-density dielectric at epsilon_r of 2.01, giving a nominal impedance. Now let the phase transition or aging shift the effective permittivity by the 1.5 percent associated with the volumetric change, so epsilon_r moves from 2.01 to about 2.04.

The impedance depends on permittivity as one over its square root, so the fractional change in impedance is half the negative fractional change in permittivity:

dZ0 / Z0 = -0.5 (d_epsilon / epsilon) = -0.5 (0.03 / 2.01) = -0.5 * 0.0149 = -0.0075

a 0.75 percent shift. On a 50 ohm cable that is

dZ0 = -0.0075 * 50 = -0.37 ohms

so the cable drifts from 50 ohms to about 49.6 ohms purely from the permittivity change. That seems small, but consider its effect on a precision measurement. A 0.37 ohm mismatch against a 50 ohm reference produces a reflection coefficient of magnitude

gamma = dZ0 / (2 * Z0) = 0.37 / 100 = 0.0037

corresponding to a return loss of

RL = -20 * log10(0.0037) = 48.6 dB

which for many purposes is acceptable, but for a metrology-grade calibration expecting 50 dB or better return loss, the aging has just pushed the cable out of specification. And this is from the permittivity alone; the geometric drift from cold flow adds further reflections distributed along the length, and the phase instability through the transition adds an electrical-length error on top. The individual effects are fractions of a percent, but they stack, and they stack differently at different temperatures and after different histories.

Why the phase stability problem is the same problem wearing a different hat

The drift shows up most painfully as phase instability, which is the same dimensional and permittivity drift expressed in the cable's electrical length rather than its impedance. The electrical length of a cable is its physical length times the square root of the dielectric constant, so anything that changes either changes the phase a signal accumulates traveling through. As temperature rises, the polytetrafluoroethylene expands and becomes less dense, lowering its permittivity and thus the electrical length, while the metal conductors expand and lengthen the cable, raising the electrical length. The two effects partly oppose, which is why cable designers carefully balance the conductor and dielectric contributions, but the room-temperature phase transition overwhelms the balance with its abrupt step.

The consequence is that a cable used as a phase reference, the way it is in a phased array or an interferometric measurement, cannot be trusted to hold its phase through a temperature swing that crosses 19 degrees, and cannot be trusted to return to the same phase after the swing because of the hysteresis. The physics is unforgiving here: the stability of a polytetrafluoroethylene-based cable can only be reduced to finite limits, never eliminated, because the limits are set by the material itself. This is why the most demanding applications abandon the material entirely, moving to dielectrics such as silicon dioxide or specially engineered low-flow formulations that hold their dimensions where the original cannot.

How the creep rate decays and what it means for a cable's lifetime

Cold flow is not a steady march but a decelerating one, and knowing its time course tells a builder how a cable ages over years. Creep in a polymer under constant stress follows an approximately logarithmic law in time, meaning the deformation grows with the logarithm of elapsed time rather than linearly:

strain(t) = strain_0 + C * log10(t / t_0)

where C is a creep coefficient that rises with stress and temperature and t_0 is a reference time. The logarithmic form has a reassuring implication and a discouraging one. The reassuring part is that the creep rate falls steeply with time: most of the deformation that will ever happen happens early. The discouraging part is that it never truly stops, since the logarithm keeps growing without bound, so a cable continues to drift, slower and slower, for its entire service life.

Quantify it. Suppose a cable's dielectric creeps to a strain of 0.1 percent in its first year under installed stress. The logarithmic law predicts the strain at later times by adding creep proportional to the log of the time ratio. Going from one year to ten years multiplies the time by ten, adding one decade to the logarithm, so if the first year produced 0.1 percent, the interval from year one to year ten produces roughly another 0.1 percent, reaching about 0.2 percent total. From year ten to year one hundred adds yet another increment of similar size. So the dimensional drift roughly doubles each decade of service rather than each year, which is why a cable can seem stable for a long time and yet measurably differ from new after many years. Translate a 0.2 percent geometric strain into impedance through the logarithmic dependence on the diameter ratio, and for a typical fifty ohm geometry it produces an impedance shift of a few tenths of an ohm, comparable to the permittivity-driven drift calculated earlier and adding to it. The two aging channels, the lurching permittivity and the creeping geometry, thus operate on different timescales, the permittivity responding fast to each temperature cycle and the geometry drifting slowly across years, and a cable's total deviation from its original number is the accumulated sum of both.

Living with a restless dielectric

The practical wisdom is to treat a high-performance coax as a precision instrument that drifts, not a fixed constant. For routine work the drift is negligible and the cable's rated impedance is fine. For precision work, the defenses are concrete. Recalibrate at the temperature of use rather than relying on a calibration done at a different temperature, because the room-temperature transition means a calibration at 22 degrees may not hold at 16 degrees. Avoid cycling the cable through 19 degrees during a measurement run, keeping it in a stable thermal environment so the transition never engages. Respect the cable mechanically, avoiding tight bends, overtightened connectors, and sustained pressure points that accelerate cold flow. And for the most demanding phase work, choose a cable whose dielectric was designed to escape the material's restlessness rather than fighting the physics of the standard material.

The deeper lesson is that no component is truly inert, and the ones we trust most are the ones whose slow changes we have stopped watching for. Polytetrafluoroethylene earned its reputation honestly with its chemical inertness and low loss, and that reputation lulls builders into treating its dimensions as eternal. They are not. The material creeps under the pressure it lives under, lurches through a transition at the temperature it lives at, and settles over the years it spends in service, and each of those motions nudges the impedance and the phase the cable was trusted to hold. The builder who knows this stops being surprised when an old cable measures a little off, and starts accounting for the drift before it quietly corrupts a measurement that depended on a number assumed to be fixed.