Ordered by depth, not by product
Six ways a closed end opens again, by exposure band
Put the part that failed back on the scale before you read further: two of these six happen to an end that never gets wet, two at a crown wetted and dried in cycles, and two under permanent backfill or standing water.
What this page can and cannot tell you
No failure rate is published for any of these parts, so nothing below is a frequency and nothing below is a field-return statistic. Each entry is reconstructed from two things that are published: what a given cap's datasheet claims to prevent, and the mechanics of how a lining bridges to a surface. Where the reconstruction runs out, the entry says where it stops.
The six run downward and never double back: two in air, two at a crown in the splash zone where end grain is open to the sky, and two under backfill or water, where the lining is bonded across its whole contact face or it is not a seal at all.
Each entry closes on the decision that would have prevented it, and on the one page carrying that decision in full. Start at cable end caps if the failed part was on a drum.
The six, from a dry compound down to standing water
The band is in the meta line. Read the two that match where the part actually lived, not where it sat when the order was raised.
Gas builds behind a cap that cannot vent
A drum closed at both ends and left standing in a hot compound keeps making gas inside the cable, and a plain sealed cap holds it there for as long as the drum stands. Two costs follow. The drum sits longer in the heating chamber before anyone can joint it, and trapped methane is named on the manufacturer's sheet as a fire risk at the moment the cap comes off, which is when a blade and often a flame are already at the end. The decision: at 33 kV and above, specify the relief-valved variant, which lifts above 3 psi and reseats.
A cover was ordered where a seal was required
The published claim for the push-on PVC cap is protection against weathering, moisture and contamination. That describes a cover over the end, not a seal bonded into it: no adhesive, no mastic, no ingress rating anywhere on the sheet. Read at speed it sounds like sealed. The end then goes somewhere it was never rated for and gets wet with the cap still firmly in place, so the fitting is rarely suspected first. The decision: if the end will ever leave the in-air band, the lined cap is the part, claimed to IP68 wherever that lining is present.
Birds and ice work on an open crown
A sawn crown is end grain pointing at the sky. Water stands in the checks, birds perch and woodpeckers open them wider, and in a freezing climate each cycle of freeze and thaw levers the split further apart. A shape that only sits over the top reaches into none of it. The lined cap is published to keep the top dry, stop perching and nesting and protect against freeze-thaw, and its adhesive runs into cracks and cavities as it shrinks, reinforcing the top rather than hiding it. The decision: a lined, torch-applied cap sized on the measured top diameter, on a power pole crown or a timber pile in the tide.
The ageing row did not cover the bond
The ageing row on the cable end cap sheet is 500 hours at 120 °C, after which the body still pulls 11 N/mm² and stretches 300 per cent. What was measured afterwards was the polyolefin, pulled: not the adhesive, not the mastic, and not wetting, drying, salt or sunlight, which between them are the whole duty in this band. The row that earns its place in a cold marina is the low-temperature one, no cracking after four hours at -40 °C (-40 °F). The decision: read ageing as evidence about the body, and check the band against the continuous temperature limit instead.
The lead wire works its way out of the anode
An anode bed rarely fails at the anode. It fails where the lead wire leaves it: that joint is buried, permanently damp and carrying current at once, and corrosion at the exit walks the wire out of the anode body until the circuit opens. The manufacturer names it as the premature-failure mode the cap exists to answer. The cap's rubber-based mastic bonds to two dissimilar surfaces at once, the anode material and the wire insulation, sealing to IP68, insulating, and taking the bending stress off the exit. The decision: the mastic-lined anode cap, sized at both ends, remembering its operating band is -55 °C to +100 °C and not the cable figures. Ordering variants are on the manufacturer's anode cap page.
The cap never pulled down onto the sheath
The lining makes this seal, not the fit: it softens under heat, flows onto the sheath and sets as the part cools. That matters most where the sheath is neither round nor smooth, since a jacket deformed on the drum, raised over the armour bedding or scored during stripping leaves a channel running the length of the contact face. A cap chosen near the top of its recovery range has too little shrink left to drive the lining into those low spots, so the bridge forms in patches. In air the end survives that. In a duct that fills, it does not. The decision: size on the measured jacket diameter against the supplied and recovered figures for the code, never on the nominal cable size. The codes are on the size charts; the part is Insulcap.
None of the six is a material failure
Every entry above resolves to a decision taken before anyone reached the end: a band assessed wrongly, a cover read as a seal, a code taken off the nominal size, or a valve left off a drum that then stood through a summer. The published body figures are identical whether the cap held or not, at 12 N/mm² tensile, 350 per cent elongation and 12 kV/mm dielectric strength. Worth settling before a batch goes back as defective.
What each published row is evidence of, and what it is not
Standards references appear in this column and nowhere else on the page. Every method listed gets its explanation on what an end-cap datasheet proves.
| Row on the datasheet | Published value | Test method | Not evidence of |
|---|---|---|---|
| Accelerated ageing, 120 °C for 500 hours | Tensile 11 N/mm² min, elongation 300% min | ASTM D2671, pulled to ASTM D638 | The adhesive or mastic bond. Only the body is pulled afterwards. |
| Low-temperature flexibility, -40 °C for 4 hours | No cracking | ASTM D2671 | Repeated freezing and thawing of a cap already wet and fitted. |
| Heat shock, 250 °C for 30 minutes | No cracking or flowing | ESI 09-11 | Torch technique. An oven result is not an installation result. |
| Water absorption | 0.2 % max | ASTM D570 | Ingress past the lining. This is uptake by the material itself. |
| Continuous temperature limit, cable end caps | -40 °C to +110 °C | IEC 216 | The anode cap, which is published at -55 °C to +100 °C. |
| Dielectric strength | 12 kV/mm min | ASTM D149 | A voltage rating for the cap once it is fitted to an end. |
| Ingress protection, lined constructions | IP68 where the lining is present | Not a test row on the sheet | The push-on cap, which carries no ingress rating at all. |
Asked after something has already leaked
The cap was still on the end and the end was still wet. Is that a fitting fault?
Can a lined cap be re-shrunk or patched once it has let water in?
Would a larger cap have sealed better?
We cannot use a torch here. Which of the six are we exposed to?
Our pile caps are twelve years old. Is that near the end of them?
Each part, filed under the failure it answers
In band order, the same way the page above runs.
- Terminal insulation capsIn air. The second failure inverted: contact, not water
- Live end capsIn air. An energised end the next person has to see coming
- De-gassing end capsIn air. The first failure, and the only one a valve answers
- Tamper-proof end sealsIn air. A drum arriving short, and a seal that can argue back
- PVC push-on end capsIn air, and no deeper. The second failure starts by ordering this
- Pole capsA crown in air. Perching, nesting and flash-over
- Dock pile capsSplash zone. Freeze-thaw and impact on a sawn timber top
- Cable gland entry sealsSplash zone at an enclosure, where the entry is the leak path
- Insulcap cable end capsBuried. The sixth failure is a sizing error on this part
- End cap with pulling eyeBuried, then winched. One moulding doing both jobs
- Anode capsBuried and submerged. The fifth failure, at the lead-wire exit
Send the failure and the dimension together
Four lines make a useful first reply: what the part was, which band it actually lived in rather than the one it was ordered for, the measured diameter over the thing being closed, and what the failed cap looked like as it came off. Nothing here is stocked or priced; Gala Thermo Shrink Pvt. Ltd. of Mumbai makes these parts.