In-Service Cable Testing: Continuity, Insulation and Flex Checks That Catch Cable Before It Fails
Quick Answer: Cable announces its decline long before it fails, but only if someone measures. A short in-service test routine, continuity at access points, insulation resistance trended against commissioning values, and flex-section inspection on a schedule, converts cable from a surprise failure into a planned replacement, and the readings themselves justify the budget line.
Introduction
The commissioning tests that a new cable installation receives are the subject of our note on insulation resistance testing, and they are necessary. They are also the last measurements many harnesses ever see, which is why cable failures arrive as surprises even though the cable spent months announcing them. In-service testing closes that gap: a modest routine, run at sensible intervals, that watches the trend instead of betting on the endpoint.
This guide is written for the buyer as much as the technician, because the readings drive purchase decisions. A trend is the difference between a spare-harness shelf stocked on faith and one stocked on evidence, and between a replacement ordered at a convenient shutdown and one ordered at three in the morning.
What to Test, and What Each Reading Means
Continuity is the cheapest and most misunderstood check. Every conductor through every access point, at rest, catches complete opens quickly, but the value is in comparing against the record: a resistance that has crept up over successive visits at one termination is a crimp degrading, and the connector-side mechanics are covered in our note on why robot connectors fail. Continuity taken only when a fault exists proves little; continuity taken twice a year against a saved baseline is a monitoring programme.
Insulation resistance is the trend that matters most on power runs. Measured conductor to conductor and conductor to shield at the manufacturer’s stated test voltage, then logged, it degrades gradually as moisture, contamination and thermal ageing accumulate. A single reading means little without history; the same reading halved over eighteen months is a replacement decision arriving in writing. Temperature correction matters, because warm cable reads lower, which is why every entry should carry the measurement temperature beside the value.
Flex sections get a physical check rather than an electrical one: jacket surface, diameter at the high-work section, and bend set, inspected on the interval the duty justifies. The evidence of what failure looks like in these sections is catalogued in our note on cable wear patterns, and a fifteen-minute inspection with that catalogue open catches corkscrewing and abrasion at the stage where the harness still has months of safe life.
The Decision Table: In-Service Checks, Intervals and What They Justify
| Check | Interval | Tooling and limits | What a bad trend justifies | Cost of skipping |
|---|---|---|---|---|
| Continuity at access points | Semi-annual | Multimeter, saved baseline | Termination refurbishment or pigtail replacement | Intermittent faults discovered by production |
| Insulation resistance, power runs | Annual, more in wet or hot zones | Insulation tester at stated voltage, temperature logged | Planned harness replacement at a convenient window | Ground fault on the shop's schedule, not yours |
| Flex-section inspection | Quarterly on high-duty axes | Visual with wear-pattern reference, caliper | Higher-flex construction on the next order | Conductor fatigue found as an axis fault |
| Termination and relief check | Semi-annual | Visual, seating, relief geometry | Crimp specification review with the supplier | Cable replaced for a crimp's failure |
| Carrier and routing condition | Annual | Visual along the full travel | Reroute, sleeve or guide work | Jacket wear against a fixture nobody logged |
| Records review | Annual | Trend sheets per axis | Spare stock and RFQ updates | Buying spares on faith |
Setting Limits and Escalation Rules
The routine as a whole costs a few hours per cell per year, most of it folded into maintenance windows that already exist. What it buys is sequencing: the harness that trends badly is replaced on your calendar with your chosen construction and a validated lead time, instead of failing into whatever harness the nearest supplier can ship at panic freight.
Turning the Programme Into Purchase Leverage
A trend without thresholds is a diary. Set, in writing, what each reading triggers: an insulation value below a stated multiple of the commissioning figure schedules a replacement order; a further fall below a second threshold schedules the replacement, not the order. Continuity drift beyond a stated percentage at one termination triggers refurbishment at the next window. The limits do not need to be perfect, they need to exist, because an engineer applying a written rule defends the downtime, and an engineer applying judgement defends the judgement.
The electrical safety envelope deserves its own line: test voltages for in-service work are capped well below commissioning levels on older cable, and instrumentation cores carry their own limits, per our note on terminating robot harnesses and the circuit limits the supplier states. A test programme that damages healthy cores has failed at its purpose, and the caps belong on the same sheet as the thresholds.
Lamination of the routine matters more than its depth. Six checks that fit inside an existing maintenance window will run for years; a programme that needs its own dedicated shift will run until the first busy quarter, and then quietly stop. Design for the busy quarter from the start.
Environment Changes the Programme
The trend sheets have a second life as procurement documents. A replacement justified by three years of insulation readings is approved without argument, and the RFQ attached to it quotes the actual duty the old harness survived, which is how the replacement gets specified for reality rather than for the drawing. The mechanism is the standing robot cable RFQ, and the cycle from trend to specification to order is what makes in-service testing a reliability investment rather than a maintenance chore.
For fleets, aggregate the sheets annually. If six of twenty similar axes show the same flex-section trend, the problem is the specification, not six harnesses, and the aggregate view is what turns that observation into a construction change across the whole fleet at the next order. The spare-stock logic follows the same evidence, as our note on spare cable programmes sets out: the axes with the worst trends hold the spares, and the healthy ones hold none.
There is a quiet negotiation benefit as well. A buyer who arrives at a supplier conversation with three years of trend data, thresholds agreed in advance and a failure history per axis is a buyer nobody needs to second-guess, and quotations arrive priced for a competent operation rather than padded for an unknown one. Records change how the market prices you, which is an outcome no amount of haggling replicates.
Before the First Test: What to Freeze
The same robot in three buildings does not need the same test calendar. A washdown food cell washes insulation values down with its shifts, and its intervals belong at the short end with moisture contamination high on the watch list, as the hygiene-duty constraints in our note on food grade machinery cable describe. A foundry-side cell adds heat and abrasive dust, where jacket surface inspection earns its frequency and the temperature column in the record does real work. A cleanroom cell runs the gentlest duty of the three, and its harnesses usually age on the calendar rather than on the duty, which argues for longer intervals and a sharper eye at the termination points instead.
Season and process changes deserve a note in the file too. A plant that shuts for two winter weeks, or a process that added a hot new machine beside the cell last summer, will show up in the trend sheet as a step change that is not cable ageing at all. Annotating the sheet with such events keeps the trend honest, and it is the difference between a record that supports decisions and one that generates arguments.
Where conditions are genuinely harsh, the programme’s findings should feed the specification directly. A cell whose insulation readings sag every summer is telling you the cable’s temperature band was set optimistically, and that finding belongs in the RFQ for the replacement, not only in the maintenance log. The correction cycle is the whole point: environment shapes the test programme, and the test programme reshapes the specification.
When More Testing Is Not the Answer
| Item | What to state | Evidence to attach | Cost of leaving it open |
|---|---|---|---|
| Test voltages | Per circuit type, in-service caps stated | Supplier limits on file | Healthy cores damaged by testing |
| Baselines | Commissioning values per axis | Commissioning records retrieved | Trends with no reference point |
| Intervals | Check calendar per axis class | A schedule in the CMMS | Checks that happen after failures |
| Thresholds | Values that trigger order and replace | A written escalation rule | Judgement calls at the worst hour |
| Temperature logging | Value and temperature per entry | A record template | Seasonal readings mistaken for decay |
| Access method | Where each check connects, per model | Photos in the procedure | An hour lost per cell per visit |
| Safety envelope | Lockout and discharge steps | A permit routine | Testing that creates the fault it hunts |
| Record owner | Where trend sheets live, who maintains | A named owner | History lost with staff turnover |
| Escalation contact | Who receives a threshold breach | A notification path | Red readings read by nobody |
| Review cadence | Annual fleet aggregation | A calendar owner | Fleet-wide specification faults missed |
RFQ Checklist
When the cable is already condemned. A harness with a known ground fault does not need a trend, it needs a replacement, and every additional test on a condemned section is downtime spent re-proving the obvious. Testing is for the decision window before failure, not after.
When the reading cannot be compared to anything. Without baselines, an insulation value is a number, not a trend. Where commissioning records are gone, establish a baseline now and read trends from the next visit onward; do not let anyone sell urgency off a single unanchored measurement.
When the duty is too light to justify the calendar. A robot that runs one shift of gentle pick-and-place does not need quarterly flex inspections. Match the interval to the duty, and spend the saved hours on the axes that actually work their harnesses, along the lines of our note on robot cabling serviceability.
When testing replaces specification. A monitoring programme that finds the same flex-section trend every year on the same duty is documenting a specification error with admirable persistence. Two annual cycles of the same finding should end in a construction change in the RFQ, not a third year of readings.
Conclusion
- In-service test voltage limits per circuit type requested from the supplier in writing
- Commissioning insulation and continuity readings required as a delivery deliverable
- Recommended test intervals for the duty profile stated by the supplier
- Threshold guidance for insulation and continuity trends requested per cable type
- Flex-section inspection points and wear signatures documented in the harness file
- Temperature correction guidance supplied with the test limits
- Replacement lead time stated so threshold breaches can be planned, not reacted to
- Construction change commitments tied to agreed trend findings
- Spare harness quantities justified by trend data, reviewed annually
- Trend records retained as warranty evidence for the cable’s whole service life
