Can Flexible Waveguide Be Used for Continuous Motion?

Yes, flexible waveguide can accommodate repeated bending or limited back-and-forth twisting, but only when the specific assembly is qualified for that motion and service life. It cannot replace a rotary joint for endless rotation in one direction. The deciding factors are the construction, dynamic bend radius, maximum twist from the relaxed position, cycle count, and RF performance during and after movement. MIL-DTL-287J leaves the flexure angle, bending planes, and number of cycles to the procurement document; it does not give every flexible waveguide a universal continuous-motion rating [1].

A tracking antenna can reverse within a bounded travel range or keep rotating through successive revolutions. Those are different jobs. Use a suitably qualified flexible section for the first; use an RF rotary joint at an interface that must turn indefinitely. Flexible sections may still absorb alignment or vibration elsewhere in that system.

Which motion does your system need?

Describe the movement of the waveguide flanges, not just the antenna pointing angle. An elevation tilt may bend a routed section rather than twist it about its axis. A 180-degree antenna sweep therefore does not automatically impose 180 degrees of axial twist.

Why endless rotation needs a rotary joint

A flexible section accommodates a finite deformation. If one flange keeps rotating while the other stays fixed, the relative twist continues to increase. Returning toward the relaxed position reduces that twist, but it does not erase fatigue from previous cycles. The maximum twist angle and the number of allowable cycles are separate limits.

Electrical performance also changes with deformation. MIL-HDBK-216 describes increasing VSWR as flexible elements approach the extremes of twisting, shearing, bending, or extension. For the interlocked construction, it reports attenuation about twice that of rigid silver tubing. These are historical construction-specific observations, not guaranteed ratings for a modern assembly [2].

Not every flexible construction can twist

The same handbook distinguishes interlocked and unsoldered convolute constructions that accommodate twist from soldered convolute and seamless corrugated constructions that do not. A product labelled “flexible” is not necessarily twistable [2]. For a fixed change in flange orientation, a rigid waveguide twist may provide the required angle without using a moving flexible section.

What catalogue twist limits tell you

The Dolph twistable flexible waveguide matrix lists the following values. They belong to the listed constructions and sizes, not to every flexible waveguide with the same WR designation. The final column is a calculation, not a recommended rotating installation.

Each calculated length is 360 degrees divided by the listed degrees-per-metre limit, rounded to two decimals. The calculation assumes uniform axial twist over the active flexible length, with no allowance for simultaneous bending, installation preload, or design margin. It extrapolates the per-metre value and does not establish that the manufacturer offers or approves that length.

For a 1.5 m active length, the same multiplication gives 69 degrees at WR284, 462 at WR90, 540 at WR75, 654 at WR62, and 765 at WR28. These are calculated angular limits, not lifetime totals. Reversing movement does not “spend” a fixed allowance of degrees; it adds fatigue cycles. The matrix does not establish a repeated-motion service life.

A worked WR90 motion check

Consider a hypothetical WR90 installation with 0.50 m of active flexible length and a flange that twists 90 degrees from relaxed before returning. Using the listed 308 degrees/m gives an angular limit of 0.50 × 308 = 154 degrees. The proposed 90-degree movement passes that arithmetic screen; a 180-degree twist from relaxed does not.

The corresponding length calculations are 90 ÷ 308 = 0.292 m and 180 ÷ 308 = 0.584 m. Even rounding those upward to 0.30 m and 0.59 m would not qualify either assembly for repeated service. The supplier still needs the actual routing, bend radius, twist waveform, rate, and cycle count. For a sweep from −90 to +90 degrees, specify both the 90-degree extremes from relaxed and the 180-degree end-to-end travel rather than calling both quantities “twist.”

This is a calculation example, not a bench test or a cycle-life prediction. Its useful result is the question to put to the supplier: does the offered assembly remain within its mechanical and RF limits throughout this particular motion?

What the qualification tests prove

MIL-DTL-287J covers twistable, non-twistable, and extra-flexible twistable assemblies that connect to rigid rectangular waveguide components, such as straight waveguide sections. Its movement tests distinguish a short bending or twist test from repeated-motion qualification [1].

• Bending, paragraph 4.7.14: five complete bend-and-return cycles in each specified plane, at the specified centreline radius.
• Flexure, paragraph 4.7.15: bending in both directions, half the specified angle on each side of relaxed, at a minimum of 30 cycles per minute. The procurement document sets the planes, angle, and cycle count.
• Axial twist, paragraph 4.7.16: twistable types undergo five complete cycles of alternating clockwise and counterclockwise twisting, at approximately one cycle per ten seconds. The specified angle must suit the length under test.
• Repeated twist, paragraph 4.7.17: the referenced twist procedure applies with the procurement-specified number of cycles; the clause is designated for twistable type one-way twisting. The agreed test procedure must define the required excursion and cycling sequence [1].

For flexure and repeated-twist testing, the specification permits nominal representative samples of 610 mm for assemblies operating up to 12.4 GHz inclusive, 305 mm above 12.4 GHz through 26.5 GHz, and 51 mm above 26.5 GHz. These are permitted test samples, not mandatory installed lengths [1].

At exactly 30 cycles per minute, 1,000 cycles take about 33 minutes and 10,000 take about 5.6 hours, excluding setup and measurements. That is test duration, not service life. The specification sets a minimum test rate, not a maximum operating speed.

After the movement tests, the assembly must have no mechanical damage and must meet the specified insertion-loss stability and VSWR requirements. A post-test measurement does not, by itself, demonstrate uninterrupted RF performance while the section moves. If the link must remain live throughout the stroke, include in-motion RF measurements in the acceptance plan [1].

Match the test to the actual motion

The Dolph flexible waveguide selection guide lists different static and dynamic radii for WR90 Flex-Twist: approximately 64 mm static versus 254 mm dynamic in the E-plane, and 127 mm static versus 381 mm dynamic in the H-plane. Treat those as manufacturer guidance for that construction, not as proof of endurance. Confirm the radii and any combined bend-and-twist restrictions on the offered assembly drawing.

A mechanism’s speed also needs its own specification. ETSI EN 302 977 V2.1.1, informative Annex C, lists typical Class B vehicle-mounted earth-station scan rates of 400 degrees per second on each axis. That is an antenna-system figure, not a waveguide rating or a definition of continuous motion [3].

At a constant 400 degrees per second, one revolution would take 0.9 seconds. That calculation assumes a full revolution at constant speed; it does not show that a terminal actually operates that way. Revolutions per minute cannot be used as an endurance comparison with the standard’s bending cycles per minute.

What published antenna designs show

Rotary joints also appear in limited-travel mechanisms. The CTS ground-terminal feed at Rosman received at 11.7–12.2 GHz and transmitted at 14.12 GHz ±250 MHz. Two rotary joints connected its rotating feed paths to stationary waveguide connections. The layout limited feed excursion to about 200 degrees, so this is evidence of a rotating interface, not of unlimited travel [4].

A 1996 gimballed spacecraft-antenna design integrated an RF rotary joint into the azimuth actuator. Its dual-reflector arrangement eliminated the elevation-axis RF rotary joint, reducing feed loss and removing a component estimated at 10 per cent of production cost [5]. JPL’s account of the 26 m DSS-11 antenna similarly describes a 960 MHz feed using rigid WR975 waveguide and a polariser with two rotary joints [6].

A 2015 Radioengineering paper describes rotary joints for azimuth, elevation, and polarisation in satellite news gathering vehicles. It reports unusual loss after wear where flexible waveguide is used for polarisation correction. Its proposed joint covers 10.7–14.5 GHz, with input reflection (S11) below −23 dB, insertion loss below 0.4 dB, and a stated 1000 W design power-handling target. Those figures describe that design, not all rotary joints [7].

At higher frequencies, a 2025 IEEE accepted paper reports a contact-free gap-waveguide rotary joint operating from 35 to 55 GHz with insertion loss below 0.68 dB. It feeds a mechanically rotating antenna with 360-degree beam steering. The absence of sliding RF contact does not change its function: it is still a rotary joint across the rotating interface [8].

Flexible sections and rotary joints can coexist

A 1970 NASA contractor study provides a useful bounded-motion example. Its conceptual antenna design used WR90 flexible sections totalling 12 inches per channel for roll and yaw movements below ±10 degrees, arranged as E-plane bending. A separate four-channel TEM rotary joint served continuous rotation at 360 degrees per day. The study separates small reversing movement from continuous rotation rather than asking one flexible section to do both [9].

What to put on the drawing

For a moving flexible section or a rotary interface, give the supplier the following acceptance conditions:

• Frequency band, WR size, flange interfaces, and installed route.
• Construction and motion type: bending, axial twist, combined motion, or continuous rotation.
• Static preload, maximum angle from relaxed, reversal range, and minimum dynamic bend radius in each plane.
• Cycle definition, rate, acceleration, operating schedule, and required total service cycles.
• Insertion loss and VSWR limits at rest, during movement where required, and after endurance testing.
• Average and peak RF power, duty cycle, pressure, temperature, vibration, and the required test environment.
• Qualification procedure, sample configuration, and the test evidence applicable to the supplied assembly.

For continuous rotation, the Dolph rotary joint matrix lists models across 2.6–40 GHz with model-dependent maximum insertion loss of 0.25–0.50 dB and average power ratings of 30–600 W. That is a family range, not one joint’s coverage. Check the selected model’s speed, endurance, band, and power ratings; the system must stay within every component’s limits, not just the rotary joint’s.

The practical answer

Choose flexible waveguide for a defined, reversing movement only when the offered assembly has suitable dynamic radii, twist limits, and endurance evidence. Choose a rotary joint where the RF interface must rotate indefinitely. A catalogue angle calculation can screen a layout, but it cannot establish cycle life or RF stability during motion.

References

1. US Department of Defense. MIL-DTL-287J: Waveguide Assemblies, Flexible, Twistable and Non-Twistable, General Specification for, 10 April 2012, paragraphs 3.5.13–3.5.16 and 4.7.14–4.7.17. DoD document hosted by EverySpec. Accessed 27 October 2026; relevant clauses rechecked 31 October 2026.

2. US Department of Defense. MIL-HDBK-216: RF Transmission Lines and Fittings, Section 6B, 12 June 1975, section 6.4.2. DoD document hosted by EverySpec. Accessed 27 October 2026.

3. ETSI. EN 302 977 V2.1.1: Vehicle-Mounted Earth Stations operating in the 14/12 GHz frequency bands, June 2016, informative Annex C, Table C.1. Accessed 27 October 2026.

4. G. J. Gill. Dual frequency feed system for the CTS communications link characterization experiment, NASA-CR-144755, 1976. Accessed 27 October 2026.

5. M. Leckie and D. Laidig. Development of a Gimballed, dual frequency, space-based, microwave antenna for volume production, 30th Aerospace Mechanisms Symposium, 1996. Accessed 27 October 2026.

6. JPL. Large Antennas of the Deep Space Network, DESCANSO Monograph 4, Chapter 2, section 2.3. Accessed 27 October 2026.

7. H. Torpi and S. M. Bostan. Ku Band Rotary Joint Design for SNG Vehicles, Radioengineering, 24(4), 2015. Accessed 27 October 2026.

8. A. Farahbakhsh, D. Zarifi and M. Mrozowski. Design of mmWave Broadband Rotary Joint and 360-degree Beam-Steering Rotenna Based on Gap Waveguide Technology, IEEE Transactions on Antennas and Propagation, accepted version, 2025. DOI: 10.1109/TAP.2025.3552222. Accessed 27 October 2026.

9. D. W. Power. Antenna Pattern Shaping, Sensing, and Steering Study, NASA CR-72628, 15 April 1970, section 2.2.3. Accessed 31 October 2026.