Bottom line up front: The goal is not to pick “the best cable,” but the cable that matches the scenario. Feeder and jumper are functional layers, not material categories — the feeder handles long-distance, low-loss backbone transmission; the jumper handles short, high-flexibility transitions between equipment ports. The same 1/2-inch cable is called a feeder on the tower and a jumper in the cabinet. Ultra-flex cable is one implementation of a jumper, trading some loss performance for a smaller bend radius. In a typical DAS link, jumpers appear far more often than feeders — and jumpers are the more common source of field PIM problems. A jumper that has been repeatedly bent and re-tightened can degrade from -160dBc to -130dBc.
Get the Three Concepts Straight: Cable, Feeder, and Jumper Are Not on the Same Logical Level
Many selection mistakes come from conflating these terms. Cable is the material-level catch-all — any 50Ω coaxial cable carrying RF signals, whether semi-rigid, semi-flex, or braided, is a cable. Feeder and jumper are functional classifications describing the role the cable plays in the system.
Feeder connects the antenna to the transceiver. The engineering definition has three conditions: long transmission distance (tens to hundreds of meters), very low loss requirement, and relatively large physical size (1/2-inch, 7/8-inch, 1-5/8-inch). A 7/8-inch corrugated copper feeder has about 0.04 dB/m attenuation at 1GHz; 1/2-inch is about 0.07 dB/m. At 2.6GHz, 1/2-inch braided cable loss is about 7–8 dB/100m, and 7/8-inch corrugated copper is about 4–5 dB/100m. Over 100 meters, the difference between 1/2-inch and 7/8-inch is 3–4 dB — which is why long backbone runs must use larger diameter.
Jumper is a cable assembly with connectors on both ends, used for short connections between equipment, or between equipment and feeder, typically 0.5m to 3m. A jumper must be a factory-terminated assembly, not bare cable with field-installed connectors. Its core mechanical constraint is bend radius: antenna ports and cabinet interiors have very limited space. A 7/8-inch feeder typically has a minimum bend radius over 200mm and simply will not fit. A 1/2-inch ultra-flex jumper can achieve under 50mm.
Ultra-flex cable is one physical implementation of a jumper. Even at the same nominal 1/2-inch size, an ultra-flex jumper uses a finer braided inner conductor and softer dielectric, giving a smaller bend radius, but its loss and shielding performance differ from 1/2-inch corrugated feeder. Not all jumpers must be ultra-flex; short connections with adequate space outside the cabinet can use semi-flex or standard braided jumpers.
Loss: Decide with Numbers, Not “Feel”
Cable attenuation grows with the square root of frequency and is inversely proportional to cable diameter. Two consequences: higher frequency means higher loss; larger diameter means lower loss. Here are reference magnitudes for the Sub-6GHz band (50Ω system, 20°C, excluding connectors):
| Type de câble | 100MHz | 1GHz | 2.6GHz | 3,5 GHz |
|---|---|---|---|---|
| 7/8-inch corrugated copper | ~0.02 dB/m | ~0.04 dB/m | ~0.05 dB/m | ~0.06 dB/m |
| 1/2-inch corrugated copper | ~0.04 dB/m | ~0.07 dB/m | ~0.09 dB/m | ~0.11 dB/m |
| 1/2-inch ultra-flex jumper | ~0.06 dB/m | ~0.12 dB/m | ~0.18 dB/m | ~0.22 dB/m |
Note that “1/2-inch” from different manufacturers has different internal structures — annular corrugated, helical corrugated, and ultra-flex braided all have different losses. The table above is an order-of-magnitude engineering reference; exact values must come from the specific model’s datasheet.
An easily overlooked calculation: A 10-meter 1/2-inch ultra-flex jumper at 3.5GHz has about 2.2dB loss. 2.2dB sounds small, but in a link budget it is equivalent to cutting transmit power to about 60%. If the jumper length is shortened from 10m to 2m, loss drops to 0.44dB, directly “recovering” 1.76dB. For high-band DAS jumpers, use 0.5m if you can instead of 1m, and run straight instead of looping — this is not about neat installation, it is about link budget.
Connecteur loss cannot be ignored either. Connector loss is approximately proportional to the square root of frequency; each connector at 3.5GHz typically contributes 0.1–0.2dB. A jumper has two connectors, plus equipment ports and feeder ports — connector losses along one link can easily exceed 1dB. The cable loss calculation can be summarized as: total insertion loss = attenuation per unit length × length + connector loss coefficient × √f + mismatch loss.
Bend Radius: Mechanical Constraints Determine Electrical Fate
Bend radius is not an “installation convenience” option — it is a hard constraint on electrical performance.
When the cable bend radius is smaller than specified, the center conductor and outer conductor are no longer strictly concentric. The characteristic impedance at that point deviates from 50Ω, creating an impedance discontinuity. PIC’s technical documentation describes this clearly: bending, stretching, and kinking push the center conductor to one side, so conductor and shield are no longer concentric, appearing on a time-domain reflectometer as an impedance “bump” or “dip.” The consequences are increased loss, abnormal reflections, and higher VSWR. More critically, this damage is often invisible to the eye — the outer jacket looks fine while the internal structure is already deformed.
Some verifiable engineering reference values:
1/2-inch ultra-flex jumper: minimum installed bend radius about 35mm, operating bend radius about 70mm (Molex corrugated jumper data). Huawei’s installation guideline requires: 1/2-inch ultra-flex jumper bend radius greater than 50mm, 1/2-inch standard jumper greater than 127mm, 1/4-inch jumper greater than 35mm. PolyPhaser’s 1/2-inch low-PIM jumper specifies a minimum bend radius of 1.25 inches (about 32mm).
7/8-inch feeder: bend radius requirement greater than 250mm. This is why feeder cannot go directly into equipment ports — the space near the port simply does not allow a 250mm bend radius.
The most dangerous location is at the connector root. PIC points out that the most vulnerable part of the cable is the connector — where the rigidity of the ferrule or clamp meets the flexibility of the cable, forming a “stress riser.” Pulling the cable perpendicular to the connector axis can directly damage the termination or the connector itself. Engineering practice: leave a straight section behind the connector so the bend occurs away from the termination; where space is extremely tight, use right-angle or angled connectors.
Connector-to-Cable Matching: Not Every Combination Works
The first principle of connector selection is not “which performs better,” but “which one fits this cable.”
Common pairings: 7/16 DIN typically mates with 1/2-inch corrugated copper jumper or feeder for high-power scenarios; 4.3-10 is the mainstream interface for 5G DAS, paired with 1/2-inch ultra-flex or 1/2-inch helical cable — many manufacturers offer 4.3-10 male to 1/2-inch ultra-flex assemblies; Type N has a wider fit range, compatible with 1/2-inch ultra-flex, 3/8-inch, and 1/4-inch jumpers, and is common in test and low-power connections.
A common field configuration: 4.3-10 male jumper at the RRU end, 7/16 DIN male or 4.3-10 male jumper at the antenna end, and 7/8-inch corrugated copper feeder for the main run. The connector types on both jumper ends must match the equipment port and feeder port. A mismatch either will not physically mate, or will mate with an incompatible contact interface — both VSWR and PIM will suffer.
PIM: Jumpers Are a More Common Weak Point Than Feeders
In multi-carrier DAS, both feeder and jumper affect PIM, but jumpers are the more common PIM source. The reason is not that jumper electrical design is worse, but that the mechanical history is more complex: jumpers are repeatedly bent and re-tightened during installation and maintenance, and the mechanical stability of the connector interface is more easily compromised. A jumper that left the factory at -160dBc can degrade to -130dBc after multiple disconnects and re-tightenings.
This means jumper selection cannot rely only on the “factory PIM specification.” Factory numbers are measured under ideal torque, clean interfaces, and single mating. Under field conditions, interface cleanliness, torque repeatability, and whether the cable was over-bent during installation all cause actual PIM to deviate from nominal. Choosing a low-PIM jumper is necessary, but low-PIM jumper + disciplined torque and cleaning procedure is what makes it sufficient.
Scenario-Based Selection
| Scenario | Recommended Configuration | Key Reason |
|---|---|---|
| RRU/equipment port to surge arrester or combiner | 1/2-inch ultra-flex jumper, 0.5–1m, matching equipment interface | Limited space, small bend radius needed; shorter length means lower high-band loss |
| Surge arrester/combiner to main feeder | 1/2-inch ultra-flex jumper, 1–2m | Transition connection, flexibility first |
| Floor-to-floor or equipment room to antenna backbone | 7/8-inch corrugated copper feeder (preferred when >30m or >2.6GHz) | Low loss over long distance; 3–4dB less than 1/2-inch over 100m |
| Indoor DAS ceiling horizontal runs | 1/2-inch corrugated feeder or 1/2-inch ultra-flex | Constrained bending, acceptable loss |
| Tower-top antenna port | 1/2-inch ultra-flex jumper + weatherproofing | Must bend, must seal against moisture |
| Test bench/lab | 3/8-inch or 1/4-inch ultra-flex jumper | Highest flexibility requirement, very short distance, loss is not the primary concern |
Feeder selection logic: Below 2.6GHz and within 100 meters, 1/2-inch is usually sufficient; beyond 100 meters or at higher frequency, consider 7/8-inch.
Jumper selection logic: First check whether the bend radius fits, then whether the connector mates, and finally whether PIM meets the system noise floor requirement. These three cannot be reordered — a jumper that will not fit is meaningless no matter how good its PIM is.
Common Mistakes
Mistake #1: Using feeder as jumper. 7/8-inch feeder has low loss, but its bend radius exceeds 200mm and it simply will not bend at an equipment port. Forcing the bend damages the internal structure and raises VSWR — and the damage is often invisible.
Mistake #2: “Close enough” jumper length. At high bands, per-meter jumper loss is not negligible. At 3.5GHz, a 10-meter ultra-flex jumper has over 2dB loss — equivalent to cutting effective radiated power to about 60%. Jumper length should be calculated with the link budget, not grabbed off the shelf during installation.
Mistake #3: Repeatedly disconnecting and reconnecting jumpers without checking PIM. Jumpers are disconnected and reconnected during commissioning and maintenance, and the connector interface mechanical state changes each time. A jumper that was once compliant may no longer be a low-PIM jumper after multiple mate cycles. After maintenance, re-measure system PIM instead of assuming “it’s still the same cable.”
Mistake #4: Bending directly at the connector root. The connector-to-cable interface is a stress riser; bending from the root applies mechanical stress directly to the weakest termination area. Correct practice: leave a straight section and move the bend point back; where space does not allow, use a right-angle connector.
MS TELERF’s low-PIM product line — couplers, power splitters, combiners, and dummy loads — undergoes 100% PIM testing at 2×43dBm before shipment. But the connection quality between components depends on cable assembly selection and installation. A -160dBc power splitter connected to an over-bent, torque-uncontrolled jumper results in system PIM determined by the jumper, not the splitter. Cable assemblies are not “accessories” — they are part of the RF link.
Article connexe
Bottom line up front: A low-PIM component that measures -160dBc in the lab can drop to -145dBc on site — and the …
An RF dummy load is a passive termination device that absorbs RF power at a 50Ω matched impedance and dissipates it as …


