Why a fitting leaks
even when the thread looks fine
Thread damage is the obvious suspect. But many leaks from otherwise correctly assembled fittings have little to do with the thread. The leak path is somewhere else — the sealing face, the gasket, the wrong seal type, the face geometry. The thread is fine. The joint is not.
Where the leak actually comes from
Most people check the thread first. Often, that is not where the leak is. The sealing face — the flat machined surface the gasket or washer compresses against — is where most leaks in correctly assembled fittings actually originate.
In a flat-face BSPP union, the parallel thread mainly supplies clamping and the separate washer or gasket seals at the face. Other BSPP designs may use an O-ring, bonded washer, cone, metal seat or specified thread sealant, so identify the actual joint before assigning the leak to the face. If the thread is intact and the fitting is correctly engaged, the thread is doing its job. If the joint leaks, the problem is at the seal, not at the thread.
Thread integrity and joint integrity are not the same thing. A fitting can have a perfect thread and a leaking joint simultaneously. The thread holds the fitting in place. The gasket or face seal stops the fluid. Both have to work.
The real causes — why a sound thread still leaks
The most common cause of repeat leaks on a fitting with an intact thread. The old gasket came out — or appeared to — but left a layer of compressed material bonded to the seating face. The new gasket sits on that residue. It does not compress against clean metal. The contact is uneven and the seal is inconsistent.
The thread looks fine. The fitting is tight. The joint weeps from the areas where the gasket never made proper contact.
The sealing face — the flat machined surface the gasket or washer compresses against — can be scored, pitted, raised or distorted by previous work. A screwdriver used to pry off the old gasket. An overtightened fitting that deformed a soft brass face. Corrosion pitting on an older cast iron flange. Scale deposits that were scraped off unevenly.
A damaged face prevents even compression across the full gasket seating area. The gasket bridges the high spots and leaves gaps at the low ones. The thread is engaged correctly. The seal is not.
A flat gasket where an O-ring belongs. A rubber washer where a fibre gasket is needed. A cellulose gasket on a joint that sees temperatures above its rating. The thread is not the issue. The wrong sealing component is doing a job it was not designed for.
Wrong seal type fails differently depending on what was used. A rubber seal on a flanged joint may extrude under bolt load. A fibre washer on a high-temperature joint may harden and crack. A flat gasket where an O-ring belongs will not compress correctly in the groove geometry it is sitting in.
The fitting is tight. Very tight. So tight that the gasket has been compressed past its sealing range — the material has extruded at the edges and the seating cross-section that was doing the sealing is now reduced. The joint feels solid. It leaks from the areas where the gasket is no longer in full contact.
A crushed gasket does not recover because you tighten it further. This is more common than most people expect, because the instinct when a joint leaks is to tighten it further — which makes the overtightening problem worse, not better.
On a flanged joint with multiple bolts, tightening in sequence around the flange — rather than in a cross pattern — creates uneven compression across the gasket face. The gasket is correctly fitted. The bolts are all tight. But the compression distribution is not uniform: highest near the last bolts tightened, lowest between them.
The joint leaks between bolt positions — not at the thread, not from the gasket material itself, but from the areas of lowest compression where the gasket never achieved full contact with both faces.
A gasket that has been under load and temperature in service has already been compressed. When the joint is opened and closed again, that gasket cannot achieve the same compression profile it had originally — the material has taken a permanent set. It may seat, appear correct, and hold initially. Under the first sustained load cycle, it will lose sealing contact.
This is a very common cause of leaks that appear shortly after a service visit. The thread is fine. The fitting was correctly reassembled. The gasket was reused.
Where to look — the leak path diagnostic
When a fitting leaks and the thread appears intact, the location and character of the leak usually tells you what you are dealing with.
Likely cause: old gasket residue, wrong gasket material, or reused gasket. Even seepage around the full circumference points to the sealing face, not the thread.
Likely cause: damaged or distorted mating face at that point, or localised gasket failure from uneven compression or a surface defect.
Likely cause: uneven bolt tightening. The leak is typically in the low-compression zones — between bolts, not at bolts.
Likely cause: overtightening. Extruded gasket at the outer edge means the material was forced beyond its sealing range.
Likely cause: reused gasket, or gasket material incorrect for the temperature. The seal holds cold but loses contact under thermal expansion and load.
Likely cause: thread damage, inadequate thread sealant, or a thread that was never fully engaged. This is the case where the thread is actually the problem — and it may be less common than face or gasket issues in flat-face connections.
In an identified flat-face union, a fine thread only draws the components together; the sealing interface is the gasketed face. Other BSP joint geometries may seal elsewhere, so identify the joint before choosing the sealing method.
When a fitting leaks and the thread looks fine, open the joint and look at what is actually happening — the face condition, the gasket condition, the compression pattern. The answer is there. It is rarely in the thread.
Clean face. Correct gasket. Even tightening. New seal every time. Those four things cover the majority of leaks from fittings with intact threads. None of them involve the thread.