VNA fixture
Convert length, one-way delay, and electrical angle, or estimate a line length from measured phase.
Blank length, delay, angle, phase values, or extra turns mean zero. Frequency and dielectric values are required. Use cable velocity factor or effective permittivity. Bulk board permittivity is not generally the effective value of a microstrip. These calculations assume a uniform, nondispersive line.
Port extension is entered as physical one-way delay per port; reflection compensation accounts for the round trip. Trace electrical delay follows the measured trace delay. The sign and entry convention depend on the instrument control; consult its help.
Enter two measured phase values at increasing frequencies. For a uniform line, decreasing unwrapped phase gives positive delay.
Unwrapped Δφ = end − start + 360° × extra turns. Example: +170° → −170° across a wrap uses +1 turn for +20°. Two samples cannot resolve the number of turns; choose it from the measured trace. Negative delay is shown as a signed estimate, not a physical cable length.
A length mismatch between the P and N halves delays one of them. The differential signal is scaled by cos(πfΔt) and the remainder converts to common mode, sin(πfΔt). No power is lost; it changes mode. Evaluated at the frequency and dielectric entered above. Blank mismatch means zero.
Mode conversion usually sets the limit: a pair can look flat on the differential response and still convert enough to radiate. The limit tightens in direct proportion to frequency, so a budget met at 5 GHz is missed by 6 dB at 10 GHz.
Order-of-magnitude only
| Item | Thumb | Why it matters |
|---|---|---|
| Free space | 1 ns ≈ 300 mm ≈ 11.8 in | About a foot per nanosecond, close enough for planning |
| FR4 stripline | ≈ 176 ps/in, 6.9 ps/mm | Fully buried, so it sees the full εr ≈ 4.3 |
| FR4 microstrip | ≈ 147 ps/in, 5.8 ps/mm | Part of the field is in air, so εeff ≈ 3 and it is faster |
| PTFE cable | ≈ 123 ps/in | Velocity factor near 0.7; use the datasheet, not bulk εr |
| Lumped or distributed | lumped below λ/10 | Past that, a trace is a transmission line whether you planned one or not |
| Reflection | trace delay = 2× one way | Port extension and S11 phase both count the round trip |
| Intra-pair skew | −30 dB conversion at 1% of a period | About 6 mil of FR4 at 10 GHz, and half that at 20 GHz |
| Skew null | total at half a period | P and N arrive in phase, so the differential signal disappears entirely |
| Sweep step for unwrapping | Δf < 1/(2τ) | A 10 ns cable needs steps under 50 MHz or the phase wraps between points |
Reflection travels out and back through a fixture. Its trace phase slope represents twice the physical one-way delay; transmission through that same line represents one traversal. Phase-slope estimates can include DUT dispersion and resonance.
Calculations stay in your browser. Shared links and exports include entered values; use nonsensitive examples.