Tech Journal · #22

edp

90Ω or 100Ω: Let the Protocol Pick Your Differential Impedance

EDPcable Engineering Team2026-06-28
90Ω or 100Ω: Let the Protocol Pick Your Differential Impedance
ARTICLE · #222026-06-28

Summary

90Ω and 100Ω are familiar differential-impedance targets, not the only possible values and not a protocol-identification shortcut. Confirm the protocol version, endpoint requirements, channel definition, tolerance, test reference plane and acceptance method before specifying a cable. The source matrix distinguishes characteristic impedance, receiver input impedance and S-parameter reference impedance, then assigns each requirement to the PCB, connector, cable or complete channel. A six-step checklist covers source and sink identification, revision control, topology, construction data, fixture/calibration or de-embedding, and the separate roles of qualification, process monitoring and release testing.

What Differential Impedance Actually Is

Differential characteristic impedance describes the relationship between differential voltage and current for a propagating mode on a transmission structure. It is not the DC resistance measured with a multimeter. Geometry, conductor and dielectric properties, coupling and nearby reference conductors all matter.

The target and tolerance come from the controlling protocol or component specification for the exact generation and channel. Do not confuse characteristic impedance with receiver input impedance or the reference impedance used to normalize S-parameters.

Use a Source Matrix, Not a 90/100 Shortcut

Decision fieldQuestionControlling evidence
Protocol identityWhich interface and revision does the source/sink implement?Protocol specification and endpoint data sheets
QuantityIs the value differential characteristic impedance, common-mode impedance, input impedance, or S-parameter reference?Definition in the named test section
LocationDoes the limit apply to PCB, connector, cable, fixture, or complete channel?Compliance topology and reference planes
ToleranceWhat range and frequency/time-domain conditions apply?Revision-specific electrical limits
MeasurementTDR, VNA/S-parameters, fixture characterization, or system compliance?Prescribed method and calibration/de-embedding procedure
CoverageQualification sample, process monitor, or release test?Customer validation and quality plan

The USB-IF publishes the USB 2.0 specification and related documents; MIPI publishes a versioned D-PHY overview; VESA publishes versioned eDP material. Consult the relevant revision instead of extending a number from one generation or topology to every implementation.

Why 10Ω Is Worth Fighting Over

An impedance discontinuity can reflect energy. Its system impact depends on magnitude, electrical length, location, transition shape, spectral content and the receiver's available margin. A nominal ten-ohm difference does not predict pass or fail without that context.

Map the complete path before assigning a cable requirement:

source package → PCB launch → board receptacle → cable plug → cable → cable plug → board/panel receptacle → sink

The connector transitions and fixtures may dominate a local discontinuity even when the cable's uniform section meets its target.

How a Cable Hits Its Target Impedance

Differential impedance is influenced by conductor geometry and spacing, dielectric properties, coupling, shield or reference-conductor geometry and transitions into the termination. The approved cable and connector specifications should define the permissible construction; assembly processing must preserve pair geometry and shield/ground treatment.

TDR and VNA measurements answer different questions. TDR is useful for locating time-domain impedance changes. Frequency-domain S-parameters can describe insertion loss, return loss and crosstalk under defined reference conditions. Neither method is meaningful as a release claim until fixtures, reference planes, calibration/de-embedding, limits and coverage are agreed.

The One Rule for Selection

  1. Identify the exact source, sink, protocol revision, lane mode and data rate.
  2. Copy the target, tolerance and test definition from the controlling source with its revision.
  3. Mark the compliance topology and reference planes across PCB, connectors and cable.
  4. Select a connector/cable construction whose manufacturer data supports the application.
  5. Define fixture, calibration or de-embedding, measurement outputs and acceptance limits.
  6. Separate qualification, process monitoring and release-test coverage in the validation plan.
FAQ04

Frequently asked questions

  • Can I mix 90Ω and 100Ω cable in the same link?

    Do not treat the labels as interchangeable. Evaluate the discontinuity against the controlling channel specification and loss budget; the outcome depends on transition geometry, electrical length, data rate and receiver margin.

  • How do I know what impedance my cable needs?

    Read the exact protocol version, endpoint data sheets and connector/cable specifications. Record whether the number is characteristic impedance, an endpoint requirement or a measurement-system reference, together with its tolerance and reference plane.

  • Are single-ended 50Ω and differential 100Ω the same thing?

    No. Differential and common-mode behavior depends on coupling and geometry. Use the quantities and limits named by the controlling specification rather than deriving one from a remembered two-to-one rule.

  • Can impedance be measured on the finished cable?

    A TDR can locate impedance changes versus propagation time, but the fixture, reference plane, rise time, de-embedding, gating and acceptance window must be defined. Coverage and reporting are project-specific.

Last updated: 2026-08-05
SEC · 03Related applications

Related applications

SEC · 02Related Articles

Related Articles