Diagnostics
Why a recovered tube fails: five faults and what caused each
Recovered wall against its ±10% chart figure, the line a split runs along, what shows at the cut ends. Three measurements, five faults, and the part still carries the evidence.
Read the part, not the job report
A recovered tube records its own installation. The wall it finished at, the line a split follows, whether anything moved at the cut ends — each points somewhere different, and all three survive the fault.
The manufacturer publishes no failure guide. The sheets carry laboratory exposure of unrecovered material, so every figure below is a threshold borrowed for a purpose it was not written for. Where no figure exists, this page says so.
The five faults
1 · A split running with the bore
Found: a straight split along the extrusion axis, edges still matching. Cause: heat held on one line instead of travelling. The 250 °C for thirty minutes that GMW and GHW carry to ESI 09-11 is a uniform oven exposure. Nothing on these sheets describes a gradient across a wall under a stationary tip. Direction separates the two cases. A split forced by a crimp shoulder underneath takes the shoulder's shape, and that one was decided at sizing.
2 · A void under a wall that looks finished
Found: nothing from outside, until water turns up in a splice whose surface is intact. Cause: working inwards from both ends, so the bore closes at each before the air between can leave. Absorption through the wall is 0.5% maximum on GMW and GHW, 0.15% on the thin wall and dual wall sheets, both to ASTM D570. The wall is not the route in. No tube sheet carries a void criterion; this one is prevented, not inspected.
3 · The tube recovered, the lining never flowed
Found: a GDW length gripping hard, cut ends clean and dry. Cause: heat enough to close the outer wall, not enough to move the hot melt beneath it. Read the chart before the part. Its wall column is printed including adhesive: GDW 6/2's 1.1 mm is 0.7 mm of polyolefin over 0.4 mm of lining, GDW 50/17's 2.4 mm is 1.5 over 0.9. The same fault reaches GDWH, GMWR and GHWR.
4 · Recovery stopped short, or never gripped
Found: a wall under the chart figure, or a bore standing off the cable. Cause: the size more often than the gun. Two of these charts head their columns D min. and d max. — supplied is a floor, recovered a ceiling. A tube halting at d has done what was published, and where d sits above the substrate the specification failed before the heat did. Check the medium and heavy wall sheet, then size again on the size chart.
5 · Service above the continuous row
Found: not a puddle. A wall that cracks when the cable is next disturbed. Cause: specifying on the shrink temperature rather than the row beneath it. 125 °C recovers GMW and GHW; the service band is −40 °C to +110 °C, and the GDLT sheet stops at +100 °C. The ageing row names what long heat takes: 12 N/mm² and 350% elongation new, 10 N/mm² and 300% after 150 °C for 168 hours. The loss is budgeted in advance.
Not published: any acceptance criterion for a recovered tube
Every thermal row on these sheets tests unrecovered material in an oven — heat shock, low temperature flexibility, accelerated ageing. None states what a recovered part must be. There is no void limit, no bead requirement, no minimum wall on a constrained substrate, no criterion for a split. So the thresholds above are borrowed ones. Where an inspection has to hold up commercially, agree the criterion with the manufacturer in writing before the tube goes on.
Four measurements, in this order
Take them before anyone reconstructs the job from memory.
- Wall, with a vernier. Against the code's published T, allowing ±10% and remembering the chart figure is free recovery onto nothing. GMW and GHW share code numbers and not walls: 105/30 is 3.0 mm as GMW, 4.0 mm as GHW.
- Split direction. Photograph it before anything is cut open. The line it runs along, not its length, is the evidence.
- Cut ends. Both of them, right round. One clean end says nothing about the other.
- Where it sat. Buried, in air or in sunlight, and whether the assembly runs at the continuous row or above it. Fault five is decided here.
Where the rest of this is written
Surface tracking and dry-band arcing are compound and creepage faults, not installation ones. Each is written once on this site.
Send the code off the failed part and what it sat on
The code, the diameter it recovered onto, the largest obstruction it had to pass, the service temperature. Those four turn a post-mortem into a size. Recovered ID is a reference and manufactures nothing; the tube is made by Gala Thermo Shrink Pvt. Ltd.