What the tests on an end-cap datasheet actually prove

IN AIR, SPLASH ZONE, BURIED or SUBMERGED: settle the band first, because that is what decides which half of a cap datasheet is evidence and which half is furniture.

Then read the size code

Place the end, then open the sheet

An end IN AIR is asked for very little, and almost any closed shape satisfies it. An end in the SPLASH ZONE is wetted and dried in cycles. A BURIED end lives in whatever the backfill holds against it, and a SUBMERGED one lives under a standing head. Only the lower two bands make an ingress claim load-bearing.

Three limits, before a single figure is quoted. No end cap in this published range carries a type-test report of its own: the withstand sequence below belongs to dip-moulded accessory components, and it qualifies a design rather than the part in your carton. The push-on PVC cap and the heat-shrinkable gland publish no ingress code at all, so nothing here reads across to them. And a certificate held by the factory is not a certificate held by the cap, which the note further down marks out precisely.

What survives those limits is still a great deal. A cap datasheet carries three separable classes of evidence, worth quite different amounts. An ingress code is a claim about an assembled joint between two materials. A property set is measured on the compound and says nothing about how the part was fitted. A type-test sequence is an electrical qualification of an accessory design. Readers trade one for another constantly, and the engineer reading 12 kV/mm as proof of a voltage rating has just done it.

IP68, digit by digit

IP: Ingress protection, defined in IEC 60529. Two digits in fixed order, solids then liquids, each graded on its own scale. Neither implies anything about the other.
First digit, 6: Dust-tight, the top of the solids scale, and on a closed end almost incidental. Nobody has the case where an end must exclude dust but not water.
Second digit, 8: Continuous immersion, and the digit that gets over-read. It is the one liquid grade whose severity the standard does not fix: depth and duration are agreed between manufacturer and user. Quoted with neither condition beside it, the code is an unfinished sentence.
What the code is actually about, on a lined cap: An interface. Cross-linked polyolefin is not porous and a flat sheet of it would pass a water test unaided; the leak path is the join between cap and substrate. The grade therefore describes hot-melt adhesive or rubber-based mastic bridging the whole circumference and staying bridged once cold, which no laboratory grade can transfer to a badly fitted part.
Where no code is published at all: The push-on PVC cap is published as protecting against weathering, moisture and contamination, with no grade anywhere on its sheet. The gland assembly is published as watertight and fume-tight, pressure-sealed to 25 psi (1.7 bar), also without one. Absence is information: read it as a narrower claim, not as an omission to fill in from a neighbouring product.

IEC 60502-4 and HD 629.1: two documents, one table

The guaranteed technical particulars reproduced below are published against two documents at once. IEC 60502-4 sets test requirements for accessories on extruded-insulation cable up to rated 30 kV, which is a highest equipment voltage of 36 kV. CENELEC HD 629.1 is the harmonisation document covering that ground for European practice, and its scope runs one designation further, to 20.8/36 (42) kV. They are not alternatives, and naming one is not naming the other.

That settles the column headings, which are the most misread part of the whole table. The figures 12, 24 and 36 are Um, the highest voltage for equipment, and not the rated voltage of the cable. So the 12 kV column is the accessory for a 6/10 kV system, 24 kV is 12/20 kV, and 36 kV is 18/30 kV. A tender naming 33 kV sits inside that last column. A tender naming 36 kV as the conductor-to-conductor rated voltage sits above it, and answering the second from the third is a real error rather than a rounding.

Why both documents are named matters at handover rather than at design. A European utility clause is normally drafted against HD 629.1 and an Indian or export clause against IEC 60502-4, and a sheet declaring compliance with both is saying its sequence satisfies whichever the buyer cites. It is not saying a certificate exists against either.

The published type-test sequence, in the order it is run

Test	Parameter for evaluation	12 kV Um	24 kV Um	36 kV Um	Result
DC high voltage, dry	15 min	38	76	114	No breakdown, no flashover
AC high voltage, dry	5 min	29	57	86	No breakdown, no flashover
AC high voltage, wet (outdoor terminations only)	1 min	25.4	51	76	No breakdown, no flashover
Partial discharge	—	11	22	33	≤ 5 pC
Insulation resistance, before impact	1 min at 500 V DC	—	—	—	Greater than 10³ Ω as printed; unit damaged in the source
Impact test (straight-through joints)	—	—	—	—	Pass, no visual damage
Insulation resistance after impact, immersed	1 min at 500 V DC	—	—	—	Greater than 10³ Ω as printed
Impulse withstand at elevated temperature	10 shots each polarity, conductor held at 95 to 100°C	95	125	170	No breakdown, no flashover
Heating cycle in air	3 cycles, 5 hrs heating and 3 hrs cooling	16 AC	30 AC	45 AC	No breakdown, no flashover
Partial discharge at elevated temperature	Conductor 95 to 100°C	11	22	33	≤ 5 pC
Partial discharge at ambient temperature	—	11	22	33	≤ 5 pC
Heating cycle in air (indoor and outdoor terminations)	60 cycles, 5 hrs heating and 3 hrs cooling	16 AC	30 AC	45 AC	No breakdown, no flashover
Heating cycle in water (straight-through joints)	63 cycles of 8 hrs, 5 hrs heating and 3 hrs cooling	16 AC	30 AC	45 AC	No breakdown, no flashover
Immersion test (outdoor terminations)	10 cycles, 5 hrs heating and 3 hrs cooling	—	—	—	No breakdown, no flashover
Partial discharge at elevated temperature	5 hrs heating and 3 hrs cooling, 95 to 100°C	11	22	33	≤ 5 pC
Partial discharge at ambient temperature	—	11	22	33	≤ 5 pC
Impulse withstand at ambient temperature	10 shots each polarity	95 peak	125 peak	170 peak	No breakdown, no flashover
AC high voltage, dry	15 min	16	30	45	No breakdown, no flashover
Examination	—	—	—	—	Pass: no cracking in the filling, no moisture path across the primary seal, no corrosion or tracking
DC high voltage, dry	15 min	38	76	114	No breakdown, no flashover
AC high voltage, dry	5 min	29	57	86	No breakdown, no flashover
Thermal short-circuit test	Two short circuits raising the conductor to the cable's short-circuit temperature; symbol illegible in the source	—	—	—	No visible damage
Impulse withstand at ambient temperature	10 shots each polarity	95 peak	125 peak	170 peak	No breakdown, no flashover
AC high voltage, dry	15 min	16	30	45	No breakdown, no flashover
Examination	—	—	—	—	Pass: no cracking in the filling, no moisture path across the primary seal, no corrosion or tracking
Humidity test (indoor terminations)	300 hour spray, water conductivity 70 ± 0.1 mS/m	8 AC	16 AC	24 AC	Pass: no breakdown or flashover, no tracking, erosion or mechanical damage
Salt fog test (outdoor terminations)	1000 hours salt spray at 1.25 × U0	7.94	15.87	23.75	Pass: no breakdown or flashover, no tracking, erosion or mechanical damage

Transcribed from the manufacturer's dip moulded components page, which prints it as one uninterrupted run declared against IEC 60502-4 and CENELEC HD 629.1. The order is kept, repeats included: an identical withstand run before and after an endurance block is the design of a type test. Two transcription notes. The discharge limit is printed as "= 5 pC" where ≤ is plainly meant. The insulation-resistance value and the short-circuit temperature symbol are both damaged in the source, and neither has been guessed at.

Three readings that decide whether the sequence means anything

Partial discharge is the figure a design lives or dies on, and it is quoted properly here, because the voltage it was measured at is printed beside the limit. A clause naming picocoulombs without kilovolts cannot be answered by anybody.

It appears three times, and one of those runs is far harder than the others. At ambient the interfaces are cold and tight. Held at a conductor temperature of 95 to 100°C the components have expanded away from one another, the lining has gone soft, and a void that was pressed shut cold now has somewhere to open into. A design meeting 5 pC only at room temperature discharges in service, because service is the hot condition.

The endurance block between the two impulse runs is where the calendar goes. Terminations take 60 heating cycles in air; straight-through joints take 63 in water, each an 8 hour cycle split 5 hours heating and 3 hours cooling. An outdoor termination then takes 10 immersion cycles and 1000 hours of salt fog. Nothing about the polymer is being asked there. The question is whether the interface survives being expanded and contracted sixty times without letting go.

The last row of each block is the one to read first. Examination names three specific outcomes: cracking in the filling, a moisture path across the primary seal, and corrosion or tracking. A sample can hold every kilovolt in the table and still fail there. Of everything in the sequence, that row is the closest thing to what a cap is for.

The polyolefin property set, with its test methods

Property	Value	Test method
Tensile strength	12 N/mm² (MPa) min.	ASTM D638
Ultimate elongation	350% min.	ASTM D638
Density	1.05 ± 0.2 g/cm³	ASTM D792
Hardness	45 ± 10 Shore D	ASTM D2240
Water absorption	0.2% max.	ASTM D570
Accelerated ageing	120°C for 500 hrs	ASTM D2671
Tensile strength, after ageing	11 N/mm² (MPa) min.	ASTM D638
Ultimate elongation, after ageing	300% min.	ASTM D638
Low temperature flexibility, -40°C for 4 hrs	No cracking	ASTM D2671
Heat shock, 250°C for 30 min	No cracking or flowing	ESI 09-11
Shrink temperature	125°C	IEC 216
Continuous temperature limit	-40°C to +110°C	IEC 216
Dielectric strength	12 kV/mm min.	ASTM D149
Volume resistivity	1 × 10¹⁴ ohm.cm min.	ASTM D257
Dielectric constant	5 max.	ASTM D150

Published for the cross-linked polyolefin used across the GEC end cap range and set out once here, so eleven product pages can quote a figure and point at this table instead of restating it. The methods in the third column are named and deliberately left undecoded on this site: what each ASTM procedure does, method by method, belongs to the shrink tube standards page. The values are the manufacturer's, from the end cap datasheet.

Four things that move underneath a shared standard number

Two sheets both naming ASTM D638 are rarely making the same claim, and the difference is almost never in the method.

Wall thickness. Dielectric strength is published per millimetre, 12 kV/mm minimum on this compound, so multiplying it by a wall does not produce a withstand rating. Two caps printing that identical figure behave differently if their walls differ, and the wall of a recovered cap is not the wall of the supplied part.
Condition. Tensile is 12 N/mm² as supplied and 11 N/mm² after 500 hours at 120°C; elongation moves 350% to 300% over the same ageing. A sheet printing only the unaged pair is publishing less of the test, not a better compound.
Which specimen. These properties are measured on the compound, not on the finished cap. They travel with the polymer rather than the geometry, which is why a thin-wall and a heavy-wall part from one batch print the same table and behave nothing alike on a drum.
Units, before any comparison at all. Volume resistivity is published as 1 × 10¹⁴ ohm.cm; the identical property in ohm.m is a hundred times smaller in figures and unchanged in fact. Density reads 1.05 ± 0.2 g/cm³ on the cable caps and 1.1 g/cm³ maximum on the anode cap, which is two compounds and not a contradiction.

What the certificates cover, and exactly where they stop

The third-party evidence is real and none of it is evidence about a cap. The CPRI certificate, dated August 2012, covers heat-shrink joints and terminations at 33 kV. ERDA's test report is for the busbar sleeve. MSEDCL approvals run first to 11 kV and then 11–33 kV, with further approvals from PGCIL, a State Electricity Board and EDL. UL lists heat-shrink tubes under File E328538 and the Busboot shroud under File E335936. Not one of those documents names an end cap, a pole cap or an anode cap. Management-system certification — ISO 9001:2015 for quality, ISO 14001:2015 for environment, ISO 45001:2018 for health and safety — and the in-house partial-discharge, high-voltage and product-testing labs sit behind the factory, not behind a part number. So where a cap must be evidenced, ask for the report number, the issuing laboratory, the date and the size code it was run on, and expect the accessory sequence above rather than a cap-specific document.

Questions a specification writer asks here

Does IP68 mean the cap will hold at any depth?

No. The second digit is the one grade in IEC 60529 whose immersion depth and duration are agreed between manufacturer and user instead of fixed by the standard, so the code carries no depth until both are attached. Ask what was agreed, in writing, wherever the end will sit under a permanent head.

Is that sequence run on every cap that ships?

No. A type test qualifies a design on samples, once. Nothing in it is repeated on the parts in a delivered carton, and it is neither a routine nor an acceptance test. Per-batch evidence has to be specified separately and will not come out of this table.

Can 12 kV/mm be turned into a voltage rating for the closed end?

No, and it is the commonest misreading of a cap sheet. It is a material property to ASTM D149, and these caps are sealing parts rather than insulating ones. The only item on this site published with a voltage figure is the gland assembly at 600 V, on cable gland entry seals.

The sheet says IP68 and my cap leaked. What failed?

The bond, nearly always, because the bond is what the grade describes. A code taken at the top of its recovery span, a sheath left wet or dirty, or a torch run too quickly each leave a stretch of circumference with no adhesive bead behind it. Six ways a closed end opens again sorts those by band; the torch procedure covers the fitting.

Why do the pole and pile pages carry no property table?

Because the manufacturer publishes none for them. Those products are described against named standards, with the pile cap given as five sizes spanning 140 to 460 mm and nothing finer. Borrowing the cable-cap values across would look like data and would be invention, so pole caps and dock pile caps say so instead.

Which of these figures belong in a tender clause?

Those carrying a method beside them and a condition under them: tensile, elongation, water absorption, dielectric strength, volume resistivity and the continuous temperature limit. The ingress grade goes in only with its immersion condition named. The withstand figures go in only if the clause also says they qualify the accessory design and not the cap.

Where each figure on this page is put to work

Every product page on this site quotes from the two tables above and explains neither, which is what keeps eleven pages from repeating one paragraph of standards.