Y21 Choke Measurement: Strengths, Limits and Alternatives
Y21 Choke Measurement: Strengths, Limits and Alternatives
Y21 is a powerful VNA technique for extracting high common-mode impedance when fixture shunt capacitance matters. Its useful range still depends on the circuit model, calibration, dynamic range and control of unwanted transfer paths.
In a suitable two-port π model, Y21 separates the desired series branch from local shunt admittances at the two fixture ports. Reliable results still require a sound fixture, calibration at defined reference planes, enough transmission dynamic range and validation against known impedances or an overlapping method.
Measure the right quantity first: a choke is best described by its complex common-mode impedance, ZCM(f) = RCM + jXCM. Installed current suppression is a separate system result that also depends on the external common-mode source and return path.
What Y21 Actually Is
A calibrated VNA measures complex scattering parameters. For a general two-port that means S11, S21, S12 and S22. Those values can be converted mathematically into the two-port admittance matrix:
I1 = Y11V1 + Y12V2
I2 = Y21V1 + Y22V2
Now represent the DUT and fixture near the calibration planes as a π network: a series branch between the ports plus one shunt branch from each port to the fixture reference. If the series-branch admittance is Y3, the matrix contains:
Y21 = Y12 = −Y3
Zseries = 1/Y3 = −1/Y21
That is the attraction. In the ideal π model, capacitance from port 1 to the reference plane contributes to Y11, and capacitance from port 2 contributes to Y22. Those local shunts do not appear in Y21, so the extracted series impedance is much less sensitive to them than a simplistic one-port high-impedance reading.
What Remains in Y21
S-to-Y conversion separates branches that fit the assumed model; it is not a general fixture de-embedding operation. Y21 still includes any admittance that transfers current directly between the two ports, regardless of which physical object created it.
- Direct port-to-port capacitance bypassing the choke appears in the transfer term.
- Magnetic or electric coupling between fixture halves can contaminate Y21.
- Radiation and cable common mode can create another transfer path through the room.
- Calibration residuals become serious when the wanted transmission is near the receiver noise floor.
- Distributed behaviour can make a single lumped π model inadequate at the upper end of the sweep.
Y21 separates model branches; it does not automatically subtract an empty fixture. Open/short fixture characterisation, port extension or a more complete de-embedding network may still be needed, depending on construction and frequency.
Choose the Method for the Measurand and Range
Impedance analysers and VNAs offer several methods because each has a useful impedance, frequency and fixture range. Select the method for the measurand, expected impedance and available calibration, then confirm it in an overlap region whenever practical.
| Method | Where it is useful | Main limitation |
|---|---|---|
| One-port reflection | Simple, calibrated measurement for low-to-middle impedances; can provide complex Z directly | At high Z, Γ approaches 1 and small phase/calibration errors or shunt capacitance create large impedance errors |
| Series-through S21 conversion | Middle-to-high impedance with a well-calibrated series fixture; valid absolute conversion is possible | A raw S21 dB trace is fixture-dependent, and shunt/direct coupling can bypass the DUT |
| Y21 π extraction | High series impedance where local shunts at the two ports matter | Requires a valid model, adequate complex data and control of mutual leakage |
| Impedance analyser or RF bridge | Direct component characterisation within the instrument and fixture’s specified range | Fixture compensation, geometry and upper-frequency parasitics still matter |
| Clamp-current measurement | Shows whether common-mode current fell in the actual station | Measures a system result at a position on the line, not the choke’s standalone impedance |
| Defined EMC injection setup | Repeatable immunity or emissions work under the applicable standard and cable configuration | Answers a standards-specific system question, not automatically the choke’s series ZCM |
One-Port S11: Accurate Within a Verified Range
For a one-port reflection measurement referenced to Z0:
Z = Z0(1 + S11) / (1 − S11)
This is exact for the calibrated one-port model. The difficulty is numerical sensitivity at high impedance. When S11 lies extremely close to +1, a tiny error in magnitude or phase becomes a large error after division by (1 − S11). A few picofarads across the fixture can also dominate a multi-kilohm DUT at HF.
One-port S11 is useful within a verified range and becomes increasingly sensitive at high impedance. Calibration at the measurement plane, open/short compensation and comparison against known standards establish where the result remains trustworthy.
Series-Through S21: Absolute Impedance in the Right Topology
For an ideal series element between two equal matched ports:
Zseries = 2Z0(1 − S21) / S21
Keysight lists the port-1-to-port-2 series method for middle-to-high impedances, with calibration or fixture compensation at the DUT terminals. The method can yield absolute complex impedance when the topology fits the equation and parasitic transfer paths are controlled.
Y21 becomes particularly attractive when separate shunts from each terminal to the fixture reference spoil the simple S21 series model. The full admittance matrix separates those local shunts while retaining the transfer branch.
IEC 61000-4-6 Is Not a Choke-Impedance Standard
IEC 61000-4-6 defines conducted RF immunity testing of electrical and electronic equipment from 150 kHz to 80 MHz. Its objective is repeatable evaluation of how equipment behaves when specified RF disturbances are coupled onto connected cables.
That is an important EMC test, but it does not standardise Y21 choke extraction or declare a universal “true choke attenuation.” A CDN, EM clamp or current-injection arrangement tests a defined equipment-and-cable system. Comparing its result directly with standalone ZCM requires an explicit circuit model and a defined measurand.
Any numerical accuracy claim needs a documented comparison. State the uncertainty contributors, reference artefacts, fixture details, calibration method, frequency range and measured data. The VNA brand or the Y21 equation alone cannot supply a percentage.
Can a NanoVNA Produce the Required Data?
Many NanoVNA-class instruments measure S11 and forward S21 in one orientation; they are not automatically full two-port VNAs that source and receive in both directions. A Touchstone file populated with four S-parameter entries does not establish that S12 and S22 were independently measured.
For a reciprocal, stable passive choke fixture, one practical approach is to make a second calibrated measurement with the DUT/fixture ports reversed and carefully assemble the four complex terms. A true full two-port VNA is more convenient. Either way:
- verify that all four complex terms contain real measurements rather than zeros or duplicated placeholders;
- keep the same reference planes and calibration state for both orientations;
- check reciprocity by comparing S21 and S12;
- use sufficient averaging and a narrow enough IF bandwidth for high impedance; and
- do not exceed the instrument’s trustworthy transmission dynamic range.
A Practical Y21 Validation Routine
- Calibrate at the fixture planes. A full two-port calibration is preferred when the instrument supports it.
- Characterise the empty fixture. Record open, short and through behaviour so direct leakage and residual delay are visible.
- Reverse the DUT. A reciprocal passive choke should not change materially after accounting for fixture asymmetry.
- Change fixture spacing deliberately. A large result change reveals electric or magnetic bypass coupling.
- Choke the analyser leads. If added ferrite changes the result substantially, the test cables are part of the measured path.
- Test known impedances. Resistors and R-L/R-C standards near the expected range expose magnitude and phase error.
- Compare methods in their overlap region. Agreement between calibrated S11, series-through and Y21 is stronger evidence than loyalty to one equation.
- Check installation current separately. Component ZCM and station current reduction are different measurements.
Report the Measurand and Validation Evidence
Publication-grade choke data should include:
- RCM, XCM and |ZCM| versus frequency;
- fixture diagram, conductor geometry and calibration planes;
- the complete extraction equation and assumed equivalent circuit;
- instrument, sweep settings, averaging and dynamic-range checks;
- open/short/through or fixture-leakage evidence;
- repeatability after reversal and reconstruction; and
- separate differential insertion-loss/return-loss data if the choke carries a wanted transmission-line mode.
A single smooth |Z| curve is not enough to show whether the impedance is resistive or reactive. A single “40 dB” curve says even less unless the source, load and definition of attenuation are stated.
Engineering conclusion: Y21 is especially useful for extracting a high series choke impedance when a valid π fixture model contains separate local shunt parasitics. Direct transfer paths, model breakdown and calibration residuals still require checks, while installed current suppression remains a separate system measurement.
Technical references
Mini-FAQ
- Why is Y21 useful? In a valid π model, it isolates the transfer branch from separate shunts at the two ports.
- Does Y21 de-embed the entire fixture? No. Direct port-to-port coupling, cable common mode and calibration errors can remain in Y21.
- When is one-port S11 useful? Within a verified accuracy range; sensitivity to small calibration and phase errors increases as impedance rises.
- Can S21 produce absolute impedance? Yes, in a calibrated series-through topology whose parasitics fit the model.
- Does IEC 61000-4-6 define choke measurement? No. It is a conducted RF immunity standard for equipment and connected cables.
- Can any NanoVNA export valid Y21? Only if the required complex two-port data are genuinely measured or carefully reconstructed; placeholder S12/S22 values are not sufficient.