Views: 0 Author: Site Editor Publish Time: 2026-10-24 Origin: Site
A cartridge seal reaches the plant in a plain cardboard box. Nobody gives the box a second look until it starts weeping on the third shift. By then the box is in the skip and the only thing left to argue with is memory. The part that decides whether that seal lasts a year or a month got settled on the shop floor, weeks before the box was ever sealed.
It begins with a pump data sheet, not a shelf. A seal is matched to a fluid, a temperature, a pressure, a speed. Get that match wrong on paper and the machine shop can do everything right and still ship a seal that fails. A refinery once took delivery of a run matched to the wrong service line. The turning was clean, the lapping was flawless. The seal was still wrong, because the drawing was wrong. The drawing is the first thing that has to be right.
Stage | What happens | What it costs if skipped |
|---|---|---|
1 Engineer | Config from pump datasheet, drawing and BOM controlled | Wrong fluid matched, seal doomed before cutting |
2 Machine | CNC machining of sleeves, glands, seal-ring holders | Runout at speed, whole stack off |
3 Inspect | Selected critical dimensions and face flatness checked | Drift built in, leaks in the field |
4 Assemble and test | Complete cartridge built, air-leak tested per unit | Site assembles blind, no proof to show |
5 Release | Documentation, traceability, crating | No trail, seal not trusted twice |
The parts that carry the load get cut on CNC. Sleeves, glands, the seal ring seats. Bores have to run true, gland faces have to sit flat, the numbers have to hold. A sleeve that wanders a few microns throws the whole stack off once it spins. Good shops hold that across the hundredth order, not just the first. A shop that nails the prototype and drifts by the tenth has not got control, it got lucky once.
The faces are the reason everything else exists. They get lapped flat to a couple of light bands, under a micron across the face. One helium light band reads about 0.3 micrometres off an optical flat under monochromatic light. That flatness is what lets a film roughly that thick hold back the process fluid. Miss it by a fraction and the seal leaks, and no assembly trick covers it. API 682 keeps this near two light bands for a reason. Flatness under load is the check most shops skip. A face has looked perfect on the bench and wept within three weeks because that check never happened. The leak shows in the field, never on the bench.
Critical points get measured, sleeves, glands, both faces, on calibrated kit. Not every number on every part, only the ones that kill a seal if they drift. A face that goes out of flat only after it is clamped in its seat is not the same fault as one that was never flat, and the check has to catch which before assembly. The records ride with the order.
It goes together as one cartridge, not a bag of bits the site has to assemble blind. Faces, springs, sleeve, gland, O-rings, drive gear, set into a pre-set unit. Clips hold the spring at its running compression so handling does not knock it out. Then the whole unit gets an air test, every single one, and the result ties to a serial number with a report in the box. That record is the only thing standing between a tested seal and a guess. A shop that cannot produce that paper when a seal comes back has already lost.
The sealed unit ships with its identity. Material certs, dimensional records, the test report, all on one serial. Years later a repeat order lands or a seal returns and that trail says what actually left the building. The crate is the last hand the seal passes through before the site, desiccant in for the long hauls, labels straight so the right seal meets the right pump. A seal that survives the bench and dies in the box throws away every step before it.
Engineering, machining, inspection, assembly and test, release. Five gates. Each one either earns the rating or gives it away. A cartridge seal is not reliable because any step is clever, it is reliable because all five line up on every unit. FBU builds them that way, air tested as one before they ship with the records to back it. The seal that runs for years is the one that proved it could before the crate was closed.
What does the air-leak test actually catch?
It catches a unit that cannot hold across its primary and secondary seals before it leaves the building. The result sits against a serial number, so a field leak gets checked against the shipped condition instead of guessed at. A seal with no test behind it is a seal nobody can defend when it comes back.
Why does face flatness matter more than the rest?
The seal runs on a fluid film about a micron thick. Flatness at that scale is the only thing keeping the process in. A face finished on a machine but out of flat on the optical flat leaks in service, and the best assembly in the world will not fix it.
Why ship one pre-set cartridge instead of loose parts?
Because the setting step is the one installers most often get wrong from loose pieces. Clips lock the spring at running compression so the seal arrives already set. A complete cartridge cannot be assembled incorrectly on site. That is the whole point of the type.