Kexingyu E-Power Group

Robot Joint Cable Routing: What to Specify for Bend, Twist and Service Loops

Flat infographic comparing five cable routing paths through a robot joint: a static run fixed inside a link, a bend only service loop, a torsion section through a rotating wrist, a hollow shaft or through arm route and an external dress pack

Quick Answer: A joint cable rarely dies because the cable was the wrong grade. It dies because of how it was routed. Bend radius, torsion angle, loop length and clamp position decide what the cable experiences, and left to the assembler those four numbers get decided by whatever fit at the time. The fix is cheap: put the routing on a drawing, name the bend radius and torsion rate, give the service loop a length and a tolerance, and require evidence at those values.

Introduction

Every axis on a robot asks the cable for something different. A shoulder or elbow sweeps through a bend. A wrist rotates and puts the bundle into torsion. A tool axis may do both at once. The cable you buy and the way you route it are two halves of one decision, and buying well while routing badly is the most common way a joint cable fails early.

The harness overview in our note on robot harnesses and connectors places the cable among the other elements. This guide stays on the routing itself: what the joint does to a cable, which path to choose, how to size a service loop, and what to write into the order.

What a Joint Actually Does to a Cable

A robot joint produces three kinds of motion, and each damages a cable differently. Bending stretches the outer layers of the bundle and compresses the inner ones, working the jacket, the screen and the conductor strands at once. Torsion rotates the layers against each other, since each is laid at its own pitch, and a long twist opens the screen and lets the cores migrate. Combined motion is worst of the three, because the cable is bent while it is twisted and the two stresses add at the same point.

The point for a buyer is that the cable does not see the joint angle. It sees the bend radius and the torsion rate the routing gives it. A joint that sweeps 180 degrees can be routed as a generous loop or as a tight fold, and the difference in life is set by a routing drawing.

The Three Motions and What Each Demands

A static run only has to reach, so length and clamping are the whole specification. A bend-only axis is governed by the minimum bend radius and the bending cycle count, which is why a continuous flex construction is bought by cycle count rather than by a general flex claim. A torsion axis is governed by the angle turned per unit of length, and a layered build with short lay lengths handles that far better than a standard cable. Where both happen at once, the routing should keep them in separate parts of the run, so the bend section is not also the twist section.

For the bend case, our notes on minimum bend radius and on continuous flex versus drag chain cable set out how the radius and the cycle count interact. For the twist case, torsion rated cable construction covers what changes inside the cable when the axis rotates rather than bends.

The Decision Table: Routing Paths and What Each One Costs

Routing Paths Through a Joint: Cable to Buy, What to Specify and Where Each Fails
Routing Approach Cable to Buy What to Specify Evidence to Demand Where It Fails in Service
Static run inside a link Standard flex construction Length tolerance, clamp positions and the clearance to the bore wall A routing drawing with clamp points marked Chafing at the clamp edge where the jacket is held hard
Bend-only service loop Continuous flex, rated by cycle count Minimum bend radius, bending cycles and the loop diameter A flex test run at the specified radius and cycle count A loop that is too tight, or one that migrates until it is
Twist axis through the joint Torsion rated, layered construction Torsion angle per metre and the number of twist cycles A torsion test at the specified angle and cycle count Screen opening and core migration after a few hundred thousand cycles
Hollow shaft or through-arm route Small diameter, low friction jacket Outside diameter clearance, bore radius at the entry and slip fit A fit sample passed through the actual bore Abrasion against the bore edge, then a screen fault
External dress pack Abrasion resistant, UV stable jacket Support spacing, loop shape and the entry transition at each end An abrasion test plus the sag and clearance drawing Sag that snags on the fixture, and wear where it rubs

Service Loops: Length, Position and Why Millimeters Matter

A service loop is not spare cable. It has three jobs: absorb the length tolerance of the assembly, give a technician something to work with during a service swap, and keep the point where the cable is clamped away from the point where it has to move. A loop that does none of those is just slack.

Length is where most of the trouble hides. Too short, and the loop pulls straight at the end of travel, putting the bundle in tension and loading the connector. Too long, and it swings with the axis, rubs against the link and catches on the fixture. The useful range is narrow, set by the axis travel, the loop diameter that keeps the bend radius above the cable’s limit, and the clearance around the joint.

Position matters as much as length. A loop should sit in one plane. If it has to bridge two axes, the cable is bent in one direction and twisted in another at the same time, which is the combined motion that wears a bundle fastest. Clamping the bundle on both sides of the loop, with the loop free to move in between, gives the cable a defined place to bend instead of letting it find its own. Where a loop ends at a connector, the transition belongs to a strain relief rather than to the loop itself, and the entry mechanics are set out in our note on strain relief at the connector interface.

Millimeters matter because the loop shape is set by the distance between two fixed points. With a loose length tolerance and drifting clamp positions, a loop designed as a generous arc can be built as a tight fold on one unit and a slack tangle on the next, from the same drawing. A tolerance on the loop length and a position on the clamp are what make the drawing reproducible on the line.

What to Freeze Before the Order

Before the Order: Eight Routing Decisions and What Leaving Them Open Costs
Item What to State Evidence to Attach Cost of Leaving It Open
Minimum bend radius The radius the routing actually imposes, not the catalogue minimum A routing drawing with the loop diameter dimensioned A loop drawn tighter than the cable can take, failing in months
Torsion rate The angle turned per metre of bundle, with the cycle count A torsion test at that angle and length Buying a flex cable for a twist axis and losing the screen
Cycle count The bending or twisting cycles over the service life, stated per axis A flex or torsion test run to that count A cable rated for a duty it never sees, or one that is over-specified
Clamp positions Where the bundle is fixed, on both sides of every loop A drawing showing each clamp and its hardware A bundle that migrates until it rubs on a link edge
Service loop length The loop length and its tolerance, plus the diameter it must keep A first-article measurement on the built harness The same drawing building tight loops on some units and slack on others
Entry strain relief The relief method at every connector, with the bend start point A drawing showing the bend starting clear of the shell Bending concentrated at the connector, cracking the seal
Jacket and friction Jacket material and its friction against the bore or guide it runs in A sample dragged through the actual guide or bore Wear that opens the screen before the conductors fail
Routing drawing One drawing per axis, issued with the order and kept current The drawing revision tied to the part number Field routing decided at build, invisible in every warranty claim

When a Service Loop Is Not the Answer

Where the joint rotates continuously. A loop cannot survive a full turn, let alone thousands, because it will wind up and break. If an axis rotates without a hard stop, the answer is a slip ring or a rotary union rather than a longer loop.

Where there is no room for the correct radius. If the link geometry cannot fit the bend radius the cable needs, the answer is a different route or a different construction, not a softer jacket folded tighter. A soft jacket folds more easily and fails sooner, because jacket flexibility has nothing to do with the strain limit of the conductors inside it.

Where the torsion is beyond the construction. If the axis turns further per metre than the cable is built for, no amount of looping fixes it. The options are a torsion rated build with a shorter lay length, a reroute that spreads the twist over a longer section, or a change to the motion itself. Routing cannot rescue a cable asked to twist past what its build allows.

Where the order has no routing drawing. Routing left to the build floor gets decided by whoever is assembling that unit, and it will not be the same twice. Where the harness leaves the robot is covered in our note on dress pack cable, and the choice between routing inside the arm and running outside it is the subject of our note on internal wiring versus a dress pack.

RFQ Checklist

  • Bend radius stated as the value the routing imposes, with the loop diameter dimensioned on a drawing
  • Torsion angle per metre and twist cycle count named for every rotating axis
  • Bending and twisting cycle counts stated per axis, not as one figure for the machine
  • Clamp positions shown on both sides of every service loop, with the hardware called out
  • Service loop length and tolerance stated, with a first-article measurement required
  • Strain relief named at every connector entry, with the bend starting clear of the shell
  • Jacket material and its friction against the bore or guide it passes through
  • One routing drawing per axis, tied to the part number and issued with the order
  • Flex and torsion evidence requested at the specified radius, angle and cycle count
  • Continuous rotation axes identified at enquiry, so a slip ring can be priced instead of a loop

Conclusion

Joint cable routing is the cheapest reliability decision on the machine and the one most often left to the build floor. The bend radius, the torsion rate, the loop length and the clamp positions decide what the cable experiences, and each can be written down in a sentence on a drawing. Buy the construction that suits the motion, put the numbers in the order, and ask for flex and torsion evidence at your values rather than at the supplier’s favourite test setting.

Kexingyu Cable Group (KXYE) supplies the cable side of robot joint routing: continuous flex and torsion rated constructions, including the robot composite cable built for moving axes, with stranding and lay lengths chosen for the bend and twist the routing imposes. Send us the axis motion, the routing sketch and the cycle counts, and we will return constructions and sample lengths that fit; the fastest route is a request for quotation.

Specify the radius your routing actually imposes, not the catalogue minimum. Measure the tightest arc the loop has to follow, state it on the drawing, and ask for the flex rating at that radius and your cycle count. Catalogue figures are usually static or occasional-flex numbers, and a joint bending thousands of times a day is neither. If the geometry cannot give you a radius inside the continuous flex rating, change the route or the loop diameter rather than accepting a tight fold.
Ask what the cable does, not what the joint is called. If the axis sweeps back and forth through a bend, it is a bending duty and a continuous flex construction is the starting point. If the joint rotates about the axis the bundle runs along, the cable is being twisted and each layer rotates against the next; that needs a layered, short lay length construction rather than a flex build. Many wrists do some of both, and then the answer is to split the motions so no single point takes both stresses at once.
Long enough to keep the bend radius inside the cable's limit at full travel, plus enough slack for a technician, and no longer. The useful range is narrow: too short pulls the bundle into tension at the end of travel and loads the connector, while too long swings, rubs and snags. The loop length is set by the axis travel, the loop diameter needed to hold the bend radius, and the clearance around the joint. State the length and a tolerance, and ask for a first-article measurement, or the same drawing will build differently on different units.
Yes, because the loop shape is a function of the distance between two fixed points. If the cable arrives long and the clamps sit where they were drawn, the loop is a slack tangle that rubs on something. If it arrives short, the loop is pulled straight and the bundle is in tension. Both cases come from the same drawing, and both are cured by a length tolerance and a clamp position rather than by a better cable. Cutting and re-terminating a finished harness in the field is the expensive way to discover this.
No. Continuous rotation winds the bundle up until the conductors and the screen fail, and no flexible jacket changes that. Where an axis has no hard stop, the answer is a slip ring or a rotary union that passes power and signal across the rotating joint, with a short fixed cable on each side. Identifying a continuous rotation axis at enquiry lets the slip ring be priced into the build instead of appearing as a field failure later.
Routing, in one of four forms: a loop tighter than the rated radius, a twist axis fed with a cable built for bending, a bundle never clamped so it migrated onto a sharp edge, or a service loop with a length nobody controlled. In each case the cable met a duty the drawing never described. That is why the routing drawing and the length tolerance belong in the order.