The hose is on, it is tight, and it comes off again on the next pull. Sometimes it stays for a week and lets go at the worst possible moment. Repeated ejection from the same joint is not bad luck, it is a joint that was never mechanically located and has been relying on friction alone.
A coupler on a pipe is held by three things: the friction of the rubber on the metal, the bead rolled into the pipe that gives the clamp something to pull the hose against, and the clamp's ability to squeeze the hose onto that bead. Lose any one of them and the joint will hold at low pressure and fail at high pressure.
The bead is the small raised ring near the pipe end. Its job is not to seal, it is to stop the hose sliding off. The clamp must sit just behind the bead, on the hose, so that when the hose tries to move the bead blocks it. A clamp sitting on top of the bead squeezes the hose onto a ridge and grips almost nothing.
Look at the witness mark on the hose when it comes off. A clean band of rubber at the very end means the hose was never pushed past the bead. A mark set back from the end means it was seated and then moved. That single observation separates an assembly error from a retention problem.
A hose whose bore is too large for the pipe will not grip even when the clamp is tight, because the clamp runs out of travel before the rubber closes onto the metal. This is common when a nominal size is used across parts made to different standards.
With the clamp loose, push the hose onto the pipe and note the effort. It should need lubricant and a firm push. If it slides on with no resistance the bore is wrong. Then tighten the clamp and look at how much thread or band remains; a clamp near the end of its travel is telling you the joint is oversize.
An engine moves on its mounts under load. If the charge pipe run has no give, that movement is fed into the coupler as shear and peel, and the joint creeps off the bead a little at a time. The hose that keeps coming off is often the one at the end of the most rigid run with the least compliance.
Watch the joint with the bonnet open while someone loads the engine against the brakes, in a safe place and with the car stationary. If the gap at the joint opens and closes, the run needs more compliance rather than a tighter clamp. Adding clamps to a joint being pulled apart only delays the failure.
Oil on silicone is a lubricant. Breather mist coating the inside of the tract, or oil that has run down the outside from a leaking cover, lets the hose slide under load at a fraction of the force it would otherwise take. The joint then looks correctly assembled and still fails, which is why this cause is often missed.
Wipe the pipe and the inside of the hose with a solvent that leaves no residue, let both dry, and refit them dry. If the joint then holds and fails again later, you have found a contaminant source to chase rather than a clamping problem.
With the car stationary, pressurise the tract and watch the joint. A brush of soapy water around the hose end shows a leak as bubbles. Then load the engine briefly and recheck. A joint that passes static pressure and fails under movement has a compliance problem, not a clamping one.
When you replace the coupler, use one with a wall thick enough to resist cold flow and a bore that matches the pipe, and seat it past a proper bead. The intercooler end tanks and outlets differ in how the bead is formed, so a coupler chosen for the pipe it actually meets is more reliable than one chosen for the nominal size. Where the joint sits near heat or sees constant movement, the heavier construction in the silicone hose range is the better starting point.
