Published on 10/13/2026
A cable assembly can look perfect and still wreck your link budget. The braid can be broken under the jacket, the crimp can be loose on the dielectric, the center contact can sit 0.3 mm proud. None of that shows up on a visual check. It all shows up on a vector network analyzer in about four seconds, which is why serious cable shops treat the VNA sweep as the final gate before anything ships.
This article explains how VNA verification works for coaxial cable assemblies: what the instrument measures, how a typical production test procedure is structured, and what you should expect to see on a test report. The procedure described here is a typical industry flow, presented for education. It is an illustrative walkthrough, not a copy of any one company's internal work instruction. What is specific to SigmaRF: every assembly we ship is VNA-swept for VSWR and insertion loss, across 100 MHz to 40 GHz.
On This Page
A VNA launches a known swept signal into the assembly and measures what comes back and what passes through, as complex ratios called S-parameters. For a two-port cable assembly, two of them carry almost all the information you need:
| Measurement | S-parameter | What it tells you | Common buyer limit |
|---|---|---|---|
| Return loss / VSWR | S11 | How much signal reflects back from connectors and impedance bumps along the cable | VSWR 1.25 to 1.5 max in-band, application dependent |
| Insertion loss | S21 | How much power the assembly absorbs, in dB, versus frequency | Cable datasheet loss plus a small connector allowance |
| Phase / electrical length | S21 angle | Electrical length; critical for phase-matched sets | Match tolerance between paired assemblies |
VSWR and return loss are the same measurement in different clothes: VSWR 1.5 corresponds to about 14 dB return loss, VSWR 1.25 to about 19 dB. A spec written in one unit converts directly to the other.
Key idea: The VNA does not test whether the cable conducts. A multimeter does that. The VNA tests whether the cable conducts cleanly at frequency, which is the thing that actually matters for RF.
Every VNA measurement includes the test cables, adapters and fixture between the instrument and the device under test. Calibration moves the measurement reference plane to the ends of the test cables, mathematically removing everything upstream. Skip it, or calibrate yesterday and measure today after the bench warmed up, and you are measuring your test setup, not the assembly.
A production station typically calibrates with a mechanical cal kit (open, short, load, through) at the start of a shift and verifies with a known check standard. When the product's connectors do not match the test port gender or family, quality adapters get added after calibration and their effect is either accepted within the measurement uncertainty budget or removed with adapter-characterization techniques. For deeper bench practice, see our article on building a reliable VNA test bench.
The exact document varies by shop, but a sound production VNA procedure for cable assemblies runs through the same stages:
Stage 01
Set the sweep
Frequency range covers the customer band with margin on both sides; point count high enough to catch narrow suck-outs (801 to 1601 points is common); IF bandwidth low enough for a clean trace but fast enough for production, typically 1 kHz.
Stage 02
Calibrate and verify
Cal at the test cable ends, then measure a check standard with a known result. If the check standard reads wrong, stop and fix the station before testing product.
Stage 03
Connect with discipline
Correct torque with a torque wrench (see our RF connector torque table), clean interfaces per our connector cleaning guide, and support the assembly so the test cables are not stressed at the port.
Stage 04
Sweep and judge against limits
The limit lines on screen encode the customer spec: VSWR or return loss max, insertion loss max. Pass means every point in-band is inside the lines, not just the band-average.
Stage 05
Record the result
Save the trace with the serial number, date, station ID and limits applied. The record is what makes the result auditable later.
Every SigmaRF assembly ships with a VNA sweep behind it. Send your frequency range and limits.
Three patterns cover most of what you will see on a sweep. A slow upward drift in insertion loss with frequency is normal cable behavior. Sharp narrow spikes in return loss usually mean a connector interface problem: wrong pin depth, a gap, or contamination. A periodic ripple that repeats along frequency points to two mismatches a fixed electrical distance apart, echoing against each other; on a cable assembly that often means both connectors have small defects, or the cable itself has an impedance irregularity.
A credible report identifies the assembly (serial or lot), the station and operator, the date, the calibration or verification status, the frequency range and limits applied, and the measured result: worst-case VSWR or return loss and worst-case insertion loss in-band, or the saved trace itself. A report that says only 'tested OK' with no numbers and no limits is a sticker, not a test.
Testing mistakes that hide bad assemblies
What is a good VSWR for a cable assembly?
It depends on the application. Many commercial specs accept 1.5 max in-band; precision test and antenna feed work often demands 1.25 or better. The limit belongs in the purchase spec, not assumed.
Can a multimeter test a coax cable instead of a VNA?
A multimeter catches opens and shorts only (see our field troubleshooting guide). It cannot see impedance bumps, connector defects or excessive loss. For RF performance, the VNA sweep is the meaningful test.
How often does a VNA station need calibration?
Common practice is a full cal each shift or after any change to test cables, adapters or environment, plus a check-standard verification. High-mix benches recalibrate more often.
What is the difference between insertion loss and return loss?
Insertion loss (S21) is how much power the assembly absorbs. Return loss (S11) is how much it reflects back toward the source. Both matter; a cable can have low loss and still reflect badly if a connector is defective.
Does every cable assembly need a VNA sweep?
For DC or audio use, continuity may be enough. For anything carrying RF above a few MHz, a sweep is the only common test that verifies the assembly actually performs. That is why SigmaRF sweeps every assembly before shipment.
What frequency range can be tested?
It depends on the VNA and test cables. SigmaRF verifies assemblies across 100 MHz to 40 GHz, covering everything from HF through common mmWave bands.
VNA verification is what separates a tested cable assembly from a hopeful one. The recipe is simple: current calibration, clean torqued connections, real limit lines, and a saved result tied to a serial number. When you buy assemblies, ask what the final test measures, at what frequencies, against which limits, and whether the trace is kept. The answers tell you most of what you need to know about the shop.
Need assemblies built and verified this way?
SigmaRF builds standard and custom coaxial cable assemblies at our Vadodara, India volume facility and our Cleveland-area precision facility in the USA. Every assembly is VNA-swept for VSWR and insertion loss across 100 MHz to 40 GHz before it ships, and lead times run as short as 2 days.