The paradox: a clean sweep and a weak carrier
A return-loss sweep comes back inside the acceptance mask. The antenna is not condemned. The LNA bias is nominal. The modem reports lock. And yet the received carrier is 3 dB low, plus or minus 0.5 dB depending on which day and which reference you trust. The natural response is to start swapping boxes. That is usually the wrong first move, because swapping changes several variables at once and destroys the before/after comparison you will need later.
The more useful reading is this: a return-loss pass is an aggregate reflection metric over a band. It tells you that the total reflected energy at the reference plane stayed below some threshold. It does not tell you where along the run that reflection originated, and it does not tell you whether the loss is reflective or dissipative. A corroded flange, a partially seated waveguide adapter, a dented run, or moisture ingress can all produce a reflection that is small enough to pass a coarse mask while the through-loss quietly eats several dB of carrier. The pass/fail number is a weak localization tool. Distance-to-fault is a stronger one.
This article is a diagnostic discipline, not a case report. The scenario in the title — a corroded WR-75 flange near 28 feet — is a plausible field pattern, and I will use it as a worked hypothetical. The measurement record that would make it a reported case is not in hand, so treat the specific distance and the specific 3 dB as illustrative, not as testimony.
What return loss and DTF each actually measure
Line sweeping, as Keysight’s field-measurement application note describes it, produces attenuation and return loss as a function of frequency along a transmission line. That is the aggregate view. It is excellent for confirming that a system is inside a mask and poor for telling you which mating surface is responsible.
Distance-to-fault works differently. The analyzer sends a signal down the line, analyzes reflections caused by impedance mismatches, and converts those reflections into distance measurements. Keysight’s note is explicit that DTF can help estimate the physical location of a fault or damage along a transmission line, and that it is used to pinpoint issues such as damaged connectors, splices, or crushed cables. The same note makes two points that matter enormously for waveguide work: waveguide systems can be characterized once an appropriate adapter is installed between the waveguide and the analyzer’s coaxial interface, and accurate DTF results depend on entering the correct velocity factor for the line type.
That last point is where most waveguide DTF arguments fall apart. A coaxial cable has a velocity factor that is a property of its dielectric. A hollow waveguide has a propagation constant that depends on frequency and on the guide’s dimensions, and it is dispersive. If you enter a number that belongs to a coaxial cable, or a single number that ignores the guide’s dispersion across the sweep, the distance axis will be wrong — and a wrong distance axis is worse than no distance axis, because it sends a technician to the wrong flange with a torque wrench and a confident expression.
The measurement-first workflow
The order below is deliberate. Each step protects the validity of the next.
1. Verify calibration and reference plane
Before any DTF run, confirm the calibration state and write down the date. A cable cal from last month’s sweep is a starting point, not a guarantee. The reference plane must be the one you intend to quote distances from — typically the analyzer port or the waveguide adapter flange. If the reference plane moves between the baseline and the repair measurement, the before/after comparison is void. State the uncertainty you are carrying: a few tenths of a dB in return loss and a distance uncertainty that depends on how well the propagation constant is known.
2. Confirm the propagation constant for the guide
For WR-75, do not guess. Obtain the guide’s propagation constant or the analyzer’s waveguide mode settings from the instrument documentation or the guide manufacturer’s data, and confirm that the sweep is set up in the correct waveguide mode. If the analyzer supports a waveguide calibration kit, use it; if you are adapting from coax, account for the adapter in the reference plane. The distance axis is only as trustworthy as this step.
3. Sweep the operating band and record the aggregate
Run the return-loss sweep across the band of interest and save the trace. This is your baseline. Note the marker values, the resolution bandwidth, and the number of points. A sweep with too few points across a dispersive guide can alias the very reflection you are hunting.
4. Run DTF with a deliberate window and distance span
Choose the window and span to match the question. A wide window sharpens distance resolution at the cost of amplitude fidelity; a narrow window does the reverse. Set the distance span to cover the full run plus a margin, and place a distance marker at the far end (the antenna flange) so you can see whether reflections cluster before or after it. Record the window type, the span, and the marker positions in the log. These are the settings that make the evidence reviewable six months later.
5. Interpret the DTF trace
A discrete reflection at a distance is a candidate mismatch: a flange, an adapter, a bend, a dent, or moisture ingress. A distributed slope suggests loss rather than a point fault. This is reasoning, not a quoted rule, and it should be checked against the physical route. If the DTF shows a reflection at roughly 28 feet and the route drawing shows a flange pair at that distance, you have a candidate. You do not yet have a diagnosis.
Reconciling model versus measured
The waveguide run has an expected through-loss and an expected return loss. Compute both from the guide data and the route length, and compare them with the measured sweep. If the discrepancy is a few dB, prioritize localization over component swapping. The point of the model is not to be right; it is to tell you which measurement is inconsistent with the others. A run that models at, say, 0.8 dB of through-loss but measures 3.5 dB is telling you something the return-loss mask cannot.
Be careful with the masking effect. Keysight’s note states that high insertion loss in the cable between analyzer and antenna can mask the true return loss of the antenna. The same logic applies to a lossy waveguide run: the reflection from a distant fault is attenuated on the way back, so a real mismatch can look smaller at the reference plane than it is at the fault. That is one reason a return-loss pass is not evidence of a healthy run.
Field verification and repair discipline
When DTF points to a flange, inspect it before you replace anything. Check the mating surfaces for corrosion, pitting, or moisture. Check that the flanges are properly seated and that the hardware is at the specified torque — use the flange manufacturer’s torque value, not a remembered number. Look at the dehydrator status and the waveguide pressurization log; a run that has been breathing moist air is a candidate for exactly this failure mode. If the flange is corroded, clean or replace it, reassemble to specification, and re-measure with the same setup: same calibration, same reference plane, same window, same span, same markers. Only then is the before/after comparison valid.
If the re-measurement does not recover the expected carrier, do not assume the repair failed. Re-examine the reference plane and the propagation constant first. A distance that moved by a few feet between runs is a sign that something in the setup changed, not that the fault migrated.
What to record so the evidence survives review
- Calibration date, kit, and reference plane.
- Waveguide mode and propagation constant source.
- Sweep settings: start, stop, points, IF bandwidth.
- DTF settings: window type, span, distance markers.
- The DTF trace and the return-loss trace, saved as files, not photographs of a screen.
- The physical route drawing with the indicated distance marked.
- The pre-repair and post-repair carrier or C/N readings, with the measurement uncertainty stated.
- The dehydrator and pressurization status at the time of the fault.
None of this is glamorous. All of it is what makes a maintenance log defensible when someone asks, a year later, why the flange at 28 feet was opened.
FAQ
Can a return-loss sweep ever be sufficient on its own?
For a go/no-go acceptance test against a defined mask, yes. For localizing a fault when the carrier is low, no. The sweep is an aggregate; DTF is a localization tool. Use each for what it measures.
Why does the distance axis disagree with the route drawing?
The most common causes are an incorrect propagation constant for the guide, a reference plane that is not where you think it is, and a sweep that is too coarse for a dispersive guide. Check those before you distrust the drawing.
Does a corroded flange always show up as a large DTF reflection?
No. The reflection amplitude depends on the geometry and the degree of degradation, and it is attenuated by the loss between the fault and the reference plane. A small reflection at the right distance is still a candidate worth inspecting.
How do I know the repair actually fixed the problem?
Re-measure with the identical setup and compare the carrier or C/N against the pre-repair baseline, with the uncertainty stated. A repair that is not re-measured is an anecdote.
Sources
Keysight Technologies, Techniques for Precise Cable and Antenna Measurements in the Field, application note 5991-0419. Used for the definitions of line sweeping, DTF operation, waveguide characterization via adapters, velocity-factor dependence, and the masking of antenna return loss by insertion loss. https://www.keysight.com/br/pt/assets/7018-03477/application-notes/5991-0419.pdf
Rohde & Schwarz, The Importance of Accurate S-Parameters for PAM-4 Applications, application note GFM355. Used for the description of S-parameters as capturing amplitude and phase response, loss and reflectivity, and coupling between network elements. https://scdn.rohde-schwarz.com/ur/pws/dl_downloads/dl_application/application_notes/gfm355/GFM355_1e_THE_IMPORTANCE_OF_ACCURATE_S-PARAMETERS_FOR_PAM-4_APPLICATIONS.pdf