When should I use TRL instead of SOLT calibration?

Use TRL when measuring devices in non-coaxial environments: on-wafer probing, waveguide, fixture-mounted devices, or PCB test structures. TRL uses only a thru connection, a reflect (open or short), and one or more transmission lines of known impedance, which can be fabricated in any medium. SOLT requires precision coaxial standards with factory-characterized reflection coefficients and is limited to coaxial connector interfaces. Above 50 GHz, even coaxial SOLT becomes problematic because the standards themselves have significant uncertainty, making TRL the better choice even in coax.

How do I verify that my calibration is good?

Measure a known verification standard that was not used during calibration. Common choices: (1) a precision airline, which should show return loss better than 40 dB across the band, (2) a precision offset short, which should show the expected phase rotation, or (3) a 20 dB attenuator, which should read within plus or minus 0.1 dB of its calibrated value. Also check the residual source match and directivity by examining the calibration coefficients on the VNA. Residual directivity should be better than 46 dB for a 3.5 mm calibration kit.

Does cable movement after calibration affect the measurement?

Yes, significantly. Bending a test cable changes its phase and magnitude response. The calibration captured the cable state at the time of cal; any subsequent movement introduces error. Phase stability of better than 5 degrees over a 90-degree bend is typical for high-quality flexible test cables (Gore, Sucoflex). For critical measurements, torque the cable connections to spec, route the cables, calibrate, and then do not move anything until the DUT measurement is complete.

Test & Measurement

Calibration

Without calibration, a vector network analyzer does not measure the device under test. It measures the device, plus the cables, plus the connectors, plus the adapter, plus the VNA's own internal imperfections, all tangled together. Calibration untangles them. By measuring a set of known standards (short, open, load, and thru for coaxial; thru, reflect, and line for non-coaxial), the VNA solves for the 12 systematic error terms in its signal flow graph and mathematically removes them from every subsequent measurement. The result is S-parameters that represent the DUT alone.

Category: Test & Measurement

Error Model: 12-term (full 2-port)

Key Standards: Short, Open, Load, Thru, Line

Twelve Errors, Four Standards, One Solution

A full two-port VNA calibration must solve for 12 error terms: three each in the forward and reverse reflection paths (directivity, source match, reflection tracking) and three each in the forward and reverse transmission paths (load match, transmission tracking, isolation). Each calibration method uses a different set of known standards to construct enough independent equations to solve for all 12 unknowns.

Calibration Method Selection

Method	Standards Needed	Best For	Frequency Range	Accuracy
SOLT	Short, Open, Load, Thru	Coaxial connectors	DC to 50 GHz	Good (limited by standard models)
TRL	Thru, Reflect, Line	On-wafer, waveguide, fixtures	Band-limited by line length	Excellent (self-calibrating)
Multiline TRL	Thru, Reflect, 3+ Lines	Broadband non-coaxial	Extended by multiple lines	Best available
eCal	Electronic cal module	Production, automated	DC to 67 GHz	Good (convenience vs. accuracy)
Unknown Thru (SOLR)	SOL on each port + any reciprocal thru	Mixed connectors	DC to 50 GHz	Good

The Verification Step Everyone Skips

Airline check: Measure a precision airline (beadless coaxial air line). Return loss should exceed 40 dB across the band. If you see ripple, something moved or a connector is damaged.
Known attenuator: Measure a 20 dB calibrated attenuator. If the reading deviates by more than ±0.15 dB from its calibrated value, recalibrate.
Offset short: Measure a precision short with a known offset length. The phase should track the expected electrical delay. A phase error indicates a calibration plane or port extension problem.
Residual error check: Examine the calibration coefficients on the VNA display. Residual directivity should be >46 dB for 3.5 mm, >40 dB for 2.4 mm, and >36 dB for 1.0 mm connectors.

Common Questions

Frequently Asked Questions

When should I use TRL instead of SOLT?

Use TRL in non-coaxial environments: on-wafer probing, waveguide, PCB fixtures. TRL requires only a thru, a reflect, and transmission lines of known impedance, all fabricable in any medium. Above 50 GHz, even coaxial SOLT becomes problematic due to standard uncertainty, making TRL preferred even in coax.

How do I verify my calibration is good?

Measure a known standard not used during calibration. A precision airline should show >40 dB return loss. A 20 dB attenuator should read within ±0.1 dB of its calibrated value. Also check residual directivity in the cal coefficients: better than 46 dB for 3.5 mm connectors.

Does cable movement after calibration affect measurements?

Yes. Bending a test cable changes phase and magnitude, introducing error that the calibration cannot correct. Use high-quality phase-stable cables (Gore, Sucoflex) and do not move cables or connections between calibration and measurement.

Related Terms

← Butterfly Stub Capture Range →

Calibration Services

Rent or Purchase Calibration Kits

We carry SOLT and TRL calibration kits for 3.5 mm, 2.92 mm, 2.4 mm, and 1.85 mm connector interfaces, plus on-wafer ISS substrates.

View Cal Kits