Manufacturing, Testing & Quality
TDR cable fault location: set time zero before calculating distance
Summary: The distance on a TDR screen starts at the agreed electrical reference plane and follows the signal path. It can help maintenance staff narrow the inspection area. To map the cursor position to a particular connection in a wire harness, they also need the propagation velocity, the test adapter segment and the actual wiring. Treating the distance as a straight-line position inside a cabinet can lead to dismantling the wrong area.
This article is for bench inspection of cables that have been isolated, confirmed de-energized and checked against the instrument's connection conditions. Its purpose is to turn a reflection reading into a location record that another engineer can check. It provides no method for testing energized systems or acceptance limits for any component.
What the instrument measures first
A conventional time-domain reflectometer sends a rapidly changing signal into the path and observes the returning waveform. Changes in impedance along the path produce reflections. The return time includes both the outward and return journeys. Keysight’s FieldFox time-domain documentation distinguishes another implementation: a network analyzer first measures the frequency response, then transforms it into a time-domain response. The record must therefore identify the equipment, mode and reference plane, rather than contain only a “TDR value”.[1]
For a path segment with approximately uniform propagation velocity, using the same reference plane and a cursor that actually represents the round-trip time for that segment:
Distance L = c × VF × Δt ÷ 2
c is the speed of light in a vacuum; the examples here use 0.299792458 m/ns. VF is the dimensionless velocity factor. Δt is the time from the reference plane to the reflection location and back, in nanoseconds. If the instrument has already converted the horizontal axis to one-way distance, read it directly in that mode without dividing by 2 again.
The Tektronix 1502C manual uses a cable of known length and the same type to determine an unknown propagation velocity. It also warns that an incorrect velocity setting affects the distance reading.[2] An insulation material name or the setting for another test jumper can help establish an initial assumption, but cannot replace the parameters of the cable under test.
One cursor, two locations
The following is a teaching example, not a product measurement. Assume that the adapter segment between the instrument port and the cable start contributes 10 ns of round-trip delay. The target reflection appears at 40 ns from the instrument port, and the target cable has VF = 0.66.
| Treatment | Round-trip time used in the calculation | Calculated result | Meaning of the result |
|---|---|---|---|
| Convert the entire delay using 0.66 | 40 ns | 3.96 m | Equivalent length obtained by treating the adapter segment as the same medium |
| Confirm the adapter contribution, subtract it, then calculate | 30 ns | 2.97 m | Estimate from the target cable start to the reflection location along the cable path |
The difference is approximately 0.99 m. It comes from the treatment of the reference plane and does not mean that the cable length has changed. For a path containing different media, it is better to check each segment in the time domain first. Under the approximation of a single distinguishable path, the total round-trip delay is the sum of 2L/(c×VF) for the individual segments. Subtract the delay of known segments before converting the segment to be located, rather than applying one VF to all adapters.
Reference-plane schematic, not a wiring diagram:
- Instrument port
- Adapter segment
- Target cable start
- Reflection location
40 ns represents the complete round-trip path. After subtracting the known adapter segment’s 10 ns, the remaining 30 ns represents the target cable portion. If calibration has already moved the reference plane to the target cable start, do not subtract the same segment again.
Keysight defines the calibration reference plane as the zero point of the time-domain horizontal axis.[1] Moving the displayed cursor or adding an axis offset is different from removing the adapter segment’s losses and reflections. Record whether calibration or de-embedding was performed according to the functions of the actual instrument. This example demonstrates delay accounting only.
Judge distance accuracy and separation of two locations separately
Uncertainty in velocity factor changes the distance result proportionally. Whether two adjacent reflections can be separated also depends on the edge reaching the test point, system bandwidth and path loss. Tektronix’s TDR documentation distinguishes “seeing a small discontinuity” from “separating two adjacent discontinuities”. The resolution for one cannot be applied directly to the other.[3]
A network analyzer’s transform mode also has specific setting constraints. Keysight explains that increasing the frequency span can improve response resolution, while changing only the time-domain display range cannot improve measurement resolution.[1] Save these settings with the screenshot so that later comparisons can distinguish changes in the specimen from changes in measurement conditions.
Consider a separate teaching example: Δt = 100 ns, VF = 0.66, giving a distance of approximately 9.89 m. If the assumed ranges under discussion are limited to VF from 0.64 to 0.68 and cursor time from 98 to 102 ns, substituting the two endpoints directly gives a distance range of approximately 9.40 to 10.40 m.
This interval reflects only the stated input ranges. It is neither an instrument accuracy specification nor a statistical confidence interval. It also excludes reference-plane error, dispersion, multiple reflections and the rules for selecting the cursor point. Giving maintenance staff a range and its basis usually makes inspection planning easier than retaining many decimal places.
From waveform to the location to dismantle and inspect
The following bench interpretation sequence was compiled for this article. It follows the propagation and measurement relationships above. Use the instrument and component documents for the actual connection, calibration and permitted signal levels.
- Draw the current path clearly. Mark the connectors, adapter segments and branches. Record the termination state and the two ends available for connection. Confirm that the system is isolated, and record whether any electronic devices remain connected. Stop if the instrument’s connection conditions are not met.
- Fix time zero and the settings. Save the mode, reference plane and VF source. Also save the edge or frequency span, cursor rule and raw waveform. If a specimen of known length and the same cable type is available, use it to check the distance scale. Record segments already compensated for to avoid subtracting them twice.
- Explain obvious structural changes first. A peak may correspond to a normal connector or an expected termination. Compare a reference specimen of the same configuration with the specimen under test under the same conditions, then list the new or changed features. Do not directly compare different termination states.
- Map candidate locations back to the physical path. Check markings and drawings along the cable’s laid-out length. Coils, bends and routing around a cabinet do not represent straight-line spatial distance. Where it is safe to connect from the other end, cross-check using the same point-selection rule and retain the differences between the two reference planes.
- Confirm the cause with the corresponding checks. Visual inspection, pinout/continuity, low-resistance and signal-integrity checks answer different questions. A reflection change alone cannot prove that a particular crimp is defective or that insulation performance meets its requirements.
| Observation | Check first | Judgment that can currently be delivered |
|---|---|---|
| Changing VF changes the cursor distance while the time feature remains largely unchanged | Parameter source and horizontal-axis units | The distance conversion has changed; there is no evidence yet of a physical change |
| Changing the adapter cable changes the near-end waveform | Reference plane, compensation and the adapter cable itself | Rule out changes in the measurement path before judging the specimen |
| Two connection locations appear as one combined response | Actual response resolution and the spacing between the two locations | The response cannot yet be assigned to one particular connection |
| Multiple branches, multiple reflections or strong attenuation make the location ambiguous | Current topology, terminations and segments that can be isolated | Retain the candidate sections and proceed with inspection of defined individual segments |
The work is complete when another engineer can use the record to restore the same wiring and parameters, recalculate the distance and identify the section to inspect next. A handoff that says only “there is a problem at about three metres” still lacks the reference plane, path and cause verification.
Reusable record
The accompanying CSV provides a blank record row and a clearly labeled teaching example. It can be edited in ordinary spreadsheet software. Its 13 columns are:
TDR location record: field guide
- Record type
- Specimen/revision
- Test mode
- Reference plane
- Termination
- Raw round-trip time
- Known adapter round-trip delay
- VF
- VF source
- Estimated distance
- Candidate location
- Verification result
- Limitation
The CSV does not control an instrument or automatically generate an acceptance conclusion. The example has been recalculated using this article’s formula; enter field data in the blank row.
The CSV has no dedicated port column. It also omits instrument identity, the raw waveform, edge or frequency span, and cursor-selection rules. Record these details, the actual wiring and the compensation applied to each segment in separate companion notes at the time of measurement. Link the notes by specimen/revision and the corresponding measurement record. The port record must identify the instrument port, connection end and cable port. reference_plane records time zero, while termination records the termination state; neither replaces port identity. Together, the CSV and companion notes form the record described above for restoring the wiring and parameters.
If a low-resistance anomaly requires a separate measurement, continue with this column’s “Low-resistance connector measurement: what four-wire connections, dry-circuit testing and offset compensation each resolve”. That article addresses measurement boundaries and error sources. It cannot replace propagation-based location here.
Sources and versions
[1] Keysight FieldFox: Time Domain, implementation, Calibration for Time Domain and Measurement Response Resolution. The public help page has no stated revision date; accessed 2026-10-13.
[2] Tektronix 1502C Metallic Time-Domain Reflectometer User Manual, 070-7169-05, pages 1-12 to 1-15. This historical instrument manual is used for the relationship between propagation velocity and distance, rather than as a current procurement, interface or safe-operation specification.
[3] Tektronix: TDR Test, Measurement Resolution and Rise Time. This public explanation is used to distinguish detection from resolution of two points; accessed 2026-10-13.
This article was researched and compiled with AI assistance. The numerical examples and interpretation sequence were developed through research and compilation; no physical MTTJ testing was performed. Check the corresponding records again when the instrument mode, connection topology or cable parameters change.
Related resources and manufacturing scope
TDR location record: field guide
TDR location record: field guide
Fill a TDR location record
Use tdr-location-record.csv to retain the measurement context, time reading, distance estimate and follow-up check in one row. Keep the 13 column names unchanged. The file contains a blank field-record row and a separate teaching example; it can be edited in ordinary spreadsheet software. It does not control an instrument, calculate results automatically or determine whether a component passes.
Use the record for a cable that has been isolated, confirmed de-energized and checked against the instrument’s connection conditions. Follow the instrument and component documents for connection, calibration and permitted signal levels. The distance conversion needs a defined reference plane, a suitable velocity factor and an identifiable reflection along the path.
Distinguish the two supplied rows
record_type token |
How to use the row |
|---|---|
BLANK_FIELD_RECORD |
Blank row for actual field or bench data. Apart from the row label, all supplied values are empty. Fill the fields from the actual setup and retain missing values as blanks. |
ILLUSTRATIVE_NOT_MEASURED |
Constructed teaching example. Its times, VF and distance describe the article’s assumptions, not measured cable or instrument results. Keep this label with the example. |
A blank field is an unfilled or missing value. It is not a zero reading, a confirmed absence of delay or a passed check. Do not replace blank fields with the example values. To record an actual measurement, fill the blank row or copy it into another record row and retain enough specimen and setup information to identify that measurement.
The 13 columns
| Column | Meaning and recording condition |
|---|---|
record_type |
Distinguishes the blank field-record template from the illustrative row. The supplied tokens are BLANK_FIELD_RECORD and ILLUSTRATIVE_NOT_MEASURED. |
specimen_and_revision |
Specimen identifier and the drawing, cable or assembly revision governing the recorded setup. The teaching row uses teaching_example_A. |
test_mode |
Instrument mode and the basis of the horizontal-axis reading. The teaching row uses round_trip_time. State whether the actual instrument reports round-trip time or has already converted it to one-way distance. |
reference_plane |
Electrical origin from which the reading is taken. The teaching row uses instrument_port_with_known_adapter_delay. Record the actual origin and any calibration or de-embedding already performed. |
termination |
Termination state and the reflection interpretation used for this record. The teaching row uses assumed_single_identifiable_reflection; this is an assumption, not a measured termination or confirmed fault. |
raw_round_trip_ns |
Raw round-trip time in ns from the recorded reference plane to the selected reflection and back. The example contains 40. Keep the raw reading separate from the compensated time used for calculation. |
known_adapter_round_trip_ns |
Independently known round-trip delay in ns of the adapter segment to be removed from the raw reading. The example contains 10. Record its basis and subtract it only if the same segment has not already been compensated for at the reference plane. |
velocity_factor |
Dimensionless VF used to convert the remaining time into distance along the target cable. The example contains 0.66. Use the value applicable to the target cable, rather than assuming every adapter and cable segment has the same VF. |
velocity_factor_source |
Basis for VF, such as the relevant cable information or a check using a known length of the same cable type. The teaching row uses assumed_for_example. Preserve an assumption as an assumption. |
estimated_distance_m |
Manually recorded calculated distance in m from the defined cable start or reference plane along the relevant signal path. The example contains 2.96794533. A numerical result does not validate the measurement assumptions or confirm a fault. |
candidate_location |
Physical cable section corresponding to the estimate, with its reference origin and route. The example contains 2.97 m along cable from cable start. Map along the cable length rather than to a straight-line position in a cabinet. |
verification_result |
Outcome of the follow-up check used to assess the candidate location or cause. The teaching row contains arithmetic_only, meaning only the calculation is demonstrated. Record the actual visual, pinout/continuity, low-resistance or signal-integrity check as applicable. |
limitation |
Assumptions, unresolved conditions or limits needed to interpret the row. The example contains No instrument control; no automatic acceptance; assumptions not measured. |
The CSV has no dedicated fields for instrument identity, ports or waveform files. It also omits edge or frequency-span settings and cursor-selection rules. Record these details, the actual wiring and the compensation applied to each segment in companion notes at the time of measurement. Link the notes to the CSV by specimen/revision and the corresponding measurement record. Identify the instrument port, connection end and cable port in the notes. reference_plane records time zero and termination records the termination state; neither replaces port identity. Together, the CSV and these notes allow another engineer to restore the wiring and parameters.
Read the 40 / 10 / 30 ns example
The illustrative row assumes a raw reflection time of 40 ns from the instrument port and a known adapter contribution of 10 ns. The round-trip time assigned to the target cable is therefore:
Δt = 40 ns − 10 ns = 30 ns
Using the article’s vacuum light speed and assumed target-cable VF:
L = 0.299792458 m/ns × 0.66 × 30 ns ÷ 2 = 2.9679453342 m
The CSV retains this as 2.96794533 in estimated_distance_m and as 2.97 m along cable from cable start in candidate_location. These decimal places represent arithmetic, not an instrument accuracy specification. Converting all 40 ns with the same assumed VF would instead give 3.9572604456 m, approximately 3.96 m. The approximately 0.99 m difference comes from reference-plane treatment.
Subtract the 10 ns only when it is the known contribution still present in the raw reading. If calibration has already placed time zero at the target cable start, subtracting it again would remove the same segment twice. If the instrument reports one-way distance, read that mode directly rather than dividing its distance by 2.
Preserve the assumptions and the next check
The article’s formula assumes that the selected time represents the round trip from the same reference plane through the target segment, with approximately uniform propagation velocity. For mixed media, check known segments in the time domain before converting the remaining target segment. Multiple branches, multiple reflections or strong attenuation may leave more than one candidate section.
Retain the VF basis and any uncertainty in the time or reference plane. Distance accuracy and the ability to separate two nearby reflections are different judgments. A reflection change does not by itself identify a defective crimp or confirm insulation performance; use the check that answers the specific cause question.
The article cites Keysight FieldFox’s Time Domain documentation for implementation, reference-plane and response-resolution conditions, the historical Tektronix 1502C manual 070-7169-05 for the propagation-velocity/distance relationship, and Tektronix’s TDR Test explanation for detection and two-point resolution. Use those sources within the scopes stated in the article. The CSV and this field guide do not supply current instrument-specific operating instructions or component acceptance limits.