Selecting the Right Semi-Rigid RF Cable Assemblies

Published on 7/29/2026 • Updated on 8/1/2026

The cable inside your Vector Network Analyzer that measures 40 GHz to 0.05 dB accuracy. The interconnect in a phased-array radar that has to hold phase across 200 channels. The RF distribution inside a satellite transponder that has to work in vacuum, at −40 °C to +85 °C, for 15 years. Every one of those is semi-rigid coaxial cable. And the reasons why come down to a single design choice: replace the braided shield with a solid metal tube.

Once you make that choice, everything else follows. The cable becomes electrically better in every measurable way: lower loss, higher shielding effectiveness, tighter impedance control, superior phase stability. It also becomes mechanically different: you can only bend it once. And it becomes expensive: two to five times what a comparable flexible cable costs. Whether that trade makes sense depends entirely on what the cable has to do.

This is the article most brochures should have been. Not a spec sheet with pretty numbers, but a proper look at how semi-rigid cable is built, why it performs the way it does, where it's the right answer, and how to spec one so it delivers what you paid for.

What semi-rigid actually is (and what it isn't)

Terminology in this corner of the RF world is genuinely confusing, and half the questions that get asked about semi-rigid are really terminology questions in disguise. Let's clear that up first.

Semi-rigid means the outer conductor is a seamless drawn metal tube (usually copper or tin-plated copper), typically with a solid PTFE dielectric and a solid silver-plated copper center conductor. The cable is formed to shape once, at manufacture, using precision bending tools with dies matched to the cable OD. After forming, the cable holds its shape permanently. You cannot flex it repeatedly without work-hardening the copper and eventually cracking it. Industry designators you'll see include UT-047, UT-085, UT-141, and UT-250 (numbers refer to nominal OD in thousandths of an inch), along with mil-spec equivalents like RG-402 (0.141") and RG-405 (0.086").

Hand-formable or conformable cables (Times Microwave TFlex, Micro-Coax UT-Flex, Belden 82XX series) are a distinct product. They use a corrugated or spirally-wrapped outer conductor that can be bent by hand without tooling. Shielding effectiveness is very good (typically 100 to 115 dB) though not quite semi-rigid class. They cost more than flexible but less than true semi-rigid, and they're what you reach for when you need semi-rigid-like performance in a location where you can't bring a bending fixture.

Semi-flexible is a marketing term that means different things to different manufacturers. Sometimes it means hand-formable. Sometimes it means a very stiff flexible cable. Read the datasheet.

Flexible coax (RG-58, RG-142, RG-400, LMR-series) uses a braided shield and can be bent repeatedly. This is what most of the coaxial cable in the world is. It's fine for the vast majority of applications. It is not what this article is about.

Anatomy of a semi-rigid cable

Four concentric layers, each with real engineering behind the material choice.

Why a solid outer conductor changes everything

The reason semi-rigid outperforms flexible cable on almost every electrical spec is worth understanding at first-principles level, because the physics is not that complicated and it explains a lot.

A braided shield, no matter how good, has gaps. Even a 95 percent optical coverage braid, which is excellent, leaves 5 percent of the outer conductor's surface as small holes in the weave. At low frequencies this doesn't matter; the wavelength is huge compared to the gap. At 10 GHz, wavelength in air is 30 mm. At 40 GHz it's 7.5 mm. Individual braid gaps become electrically significant. Some RF leaks out; some external interference leaks in. Signal integrity suffers in ways that don't show up on a DC continuity check.

A seamless drawn tube has no gaps. Zero. The outer conductor is a continuous, unbroken cylinder of metal, and RF current flows smoothly on its inner surface (skin depth in copper at 10 GHz is about 0.66 μm, so the current is confined to a thin surface layer regardless). What you get from this construction:

The specific numbers vary by exact cable and manufacturer, but the pattern is consistent: semi-rigid is roughly half the loss, two to four times the shielding effectiveness in dB, and dramatically better phase stability compared to the equivalent flexible cable.

Common sizes and what each is for

Semi-rigid cable size is named by outer diameter in thousandths of an inch. The four sizes you'll see most:

The pattern is straightforward. Larger cable diameter gives you more inner conductor surface area, lower loss, higher power handling, and better tolerance of manufacturing variation. It also gives you a bigger minimum bend radius (a rule of thumb is 3x the OD for a semi-rigid without cracking), heavier weight, and larger connector footprint at the ends. Design starts with "what's my frequency and my power?", moves to "what's my routing space?", and lands on a size.

Semi-rigid vs flexible: a real comparison

The right way to think about this comparison is: what job does the cable have to do?

There's no "better." There's fit-for-purpose. The engineering error most often made here is choosing semi-rigid because it "sounds premium" when a good flexible cable would do the same job for a fifth of the price. The mirror-image error is spec'ing flexible on a phased-array radar module because it "should be cheaper" and then chasing phase mismatch problems for six months.

Phase-matched cable pairs

This topic almost never appears in general RF cable articles, but it's one of the main reasons semi-rigid exists as a product category. Whenever you have a multi-channel system where the RF signals from different paths have to arrive at the same time (phased-array radar, beamforming, interferometric measurement, MIMO characterization), the electrical length of each cable in the set must match to within a very small tolerance.

"Electrical length" is not the same as physical length. It's how long the RF signal takes to propagate, which depends on the exact velocity of propagation of the dielectric, the exact center-conductor geometry, and the cable's temperature. Two flexible cables of the same nominal length can easily differ by 5 to 10 electrical degrees at 10 GHz. That's enough to ruin an interferometric measurement or degrade a phased-array's beam pattern.

Semi-rigid cables are inherently phase-matchable because the geometry is fixed. When SigmaRF (or any competent supplier) delivers a matched set, each cable in the set is measured on a VNA against a reference, and the physical length is trimmed by the supplier until all cables in the set fall within the specified phase tolerance. Common tolerances:

• ±1 degree at 10 GHz for demanding radar and interferometry
• ±5 degrees at 10 GHz for typical phased-array applications
• ±10 degrees at 10 GHz for MIMO and beamforming characterization

Tighter tolerances cost more because they require more measurement and iteration during manufacture. If you need phase matching, specify the tolerance up front. Adding it after the fact means the assemblies get re-worked, and re-worked semi-rigid is a compromised assembly.

The trade-offs you're accepting

Semi-rigid is not free performance. The construction has real costs, and any engineer specifying semi-rigid should be aware of what they're trading away.

Selecting the right semi-rigid cable

Six questions get you 90 percent of the way to a proper spec. Answer them before you look at any catalog.

What is the operating frequency range? This drives cable size and connector choice. Above 26.5 GHz you're in 2.92 mm connector territory (or precision 3.5 mm); above 40 GHz means 2.4 mm; above 50 GHz means 1.85 mm. Standard SMA is fine to 18 GHz but not above.

What is the loss budget? Total system insertion loss at the operating frequency. Longer runs need larger diameter cable, and vice versa. If you have 2 dB to spend on the cable at 20 GHz over a 30 cm run, 0.085 semi-rigid gets you there. Try to do the same job on 0.047 and you'll be over budget.

What is the power level? CW and peak. Semi-rigid handles more power than equivalent flexible because the solid outer conductor dissipates heat better. 0.085 is typically rated to 100 W CW at 1 GHz, 30 W at 10 GHz. 0.141 doubles that. 0.250 doubles it again. Peak power ratings are much higher, limited by dielectric strength.

What's the routing? Draw the actual path. Note every bend, its radius, and the distance between bends. Semi-rigid needs at least 3x OD for the bend radius (so 0.141 cable needs 0.42 inches minimum radius on any bend), and it needs a straight section on either side of every bend for the forming tool to grip. This is where a lot of amateur specs come apart: on paper it fits, in reality the forming die can't get to it.

What's the environment? Temperature range, atmosphere (air, vacuum, dry gas), humidity, corrosion exposure, vibration, radiation. Copper outer conductor for most work. Tin-plated copper for improved solderability and mild corrosion resistance. Stainless steel for high-temperature (aerospace engine bays, industrial). Aluminum for weight-critical space applications. PTFE dielectric handles −65 to +200 °C; if you need more, you need a specialty product.

Do you need phase matching? If yes, specify the tolerance and the reference frequency. If no, save yourself the cost.

Connectors for semi-rigid

Semi-rigid connectors attach by soldering the outer conductor tube directly to the connector body. This is where a lot of the price of a proper semi-rigid assembly goes, and where the difference between a good supplier and a bad one shows up.

Precision connectors commonly paired with semi-rigid:

• SMA: workhorse, works to 18 GHz with commercial versions, 26.5 GHz with precision. Solder-attach standard for semi-rigid.
• 3.5 mm: mates with SMA mechanically, precision electrical performance to 26.5 GHz.
• 2.92 mm (K): to 40 GHz, mates with SMA and 3.5 mm.
• 2.4 mm: to 50 GHz, does not mate with SMA (intentionally: prevents accidental damage to precision hardware).
• 1.85 mm (V): to 67 GHz, mates with 2.4 mm.
• SMP: subminiature push-on for board-to-cable interconnect, to 40 GHz.
• SMPM: smaller SMP variant for dense arrays, to 65 GHz.
• N-Type: for larger 0.141 or 0.250 semi-rigid where higher power and lower frequency (to 18 GHz) is the priority.

Which connector to use is driven by frequency, then by mating requirements, then by the mechanical constraints of what board or panel the connector interfaces with. Do not pair a $2 SMA with a $50 semi-rigid cable expecting the SMA won't be the weakest link. It will be.

How semi-rigid assemblies are actually built

A proper semi-rigid manufacturing process is nine to eleven steps, and each one affects the final performance. Skip any of them and the cable is compromised.

What proper testing looks like

A semi-rigid assembly should arrive with test data. The minimum you should expect:

• Insertion loss measured across the specified frequency range on a calibrated VNA, with the trace or the tabulated data included in the test report
• Return loss / VSWR at both ports across the same range
• Phase length when phase-matched pairs are ordered, with the measured phase difference relative to the reference documented at the specified frequency
• Visual pass / fail confirming the assembly passed microscope inspection
• Serial number tying the specific assembly to its test data
• Cable date code and manufacturing traceability information for aerospace and defense use

If your supplier will not provide this documentation, they either don't test every assembly (in which case some fraction of what they ship is out of spec) or they don't want you to know what tolerances they're actually holding. Both are bad signs for anything above commercial-grade cable.

Failure modes and how they happen

Semi-rigid assemblies have long service lives when they're built and installed properly. When they fail, the usual suspects are:

Notice how many of these are traceable to manufacture rather than to the cable design itself. This is why choosing a competent supplier matters more than choosing between UT-085 and UT-141.

Questions engineers actually ask

When semi-rigid is the right answer

Semi-rigid is not for every RF system. It is for the systems where the RF interconnect has to be as close to invisible as physically possible: no measurable phase drift over the operating temperature range, no shielding leakage that shows up as receiver noise, no impedance perturbation that pushes VSWR beyond spec. When those requirements are real (radar, satellite, mmWave test, aerospace, precision measurement), the seamless outer conductor is not a nice-to-have. It is the reason the system works.

And when those requirements are not real, semi-rigid is expensive theater. There is no engineering virtue in overspec'ing a jumper cable that just needs to carry a 900 MHz signal from a radio to an antenna two feet away. Use the right cable for the job.

The judgment lives in that assessment. Understand what semi-rigid actually does, know how it's built, spec the trade-offs honestly, choose a supplier who tests every assembly and ships the data, and it becomes the boring, reliable part of your system. Which, at these frequencies, is a rare and valuable thing.