A thread from a candy-packaging line where the gearbox drip was measured, not guessed, and the cure turned out to be a material swap and a shaft polishing step, not a tighter spring. Names blotted for the supplier.
From: J. Osei, Maintenance — To: R. Almeida, Design; cc line supervisor
Subject: conveyor gearbox CB-07 input seal leaking again, 6-week cycle
Ricardo, we are on the third NBR lip seal in four months on the input shaft of the CB-07 geared motor, and the leak is clockwork, not drama. We fit a new seal, it runs dry for about five weeks, then somewhere in the sixth week the oil starts weeping down the shaft and on to the belt pulley, about 30 ml a shift by the drip tin, and we change it again at the next PM. Input shaft 55 mm, 1750 rpm through a belt, oil is the same ISO VG 220 the box is filled with, oil bath temperature reads 82 to 86 C on the recent probe, which I think is the story, because the seal flank that comes out is hard and cracked at the lip. We install with the plastic sleeve you get in the box, shaft finish is a turned and burnished groove, measured Ra 0.9, and there is no scoring I can see with a boroscope, the contact band is even. I am about to order the same seal again because that is what the drawing says, but the drawing is five years old and I would rather ask you before I repeat a 6-week pattern.
From: R. Almeida, Design — To: K. Ivanova, Seal Supplier Applications; cc J. Osei
Subject: RE: lip seal life 6 weeks at 85 C — what is the honest service life here
Katerina, applied for your seal-choice review. Conveyor gearbox input, 55 mm shaft, 1750 rpm, ISO VG 220 oil, 82-86 C oil bath measured at the probe, current seal is a standard NBR with a wire garter spring, failing at roughly six weeks, hard cracked lip, weep past the spring side. Shaft Ra 0.9, burnished, no scratch, installed with the sleeve. I am told the catalogue says NBR works to 100 C, but our reality is six weeks, so either the catalogue number is honest only at the shaft, or our temperature is not the number we think it is, because lip temperature runs hotter than the oil bath, and nobody here has measured lip temperature at all. I want your calculation of actual lip temperature for this size at 1750 rpm, and your recommendation on material, because if this is temperature-driven I would rather fix the material than invent a heavier spring and call it an improvement.
From: K. Ivanova, Seal Supplier Applications — To: R. Almeida; cc J. Osei
Subject: RE: lip seal life 6 weeks — your catalogue number is honest, your lip is hotter than it
Ricardo, the catalogue 100 C for NBR is a continuous-service figure at the material, and your lip is not at the catalogue temperature, it is at the shaft-contact zone, which for a 55 mm seal at 1750 rpm with VG 220 runs somewhere between 18 and 25 C above the oil bath in a location like yours, so 82-86 bath pushes the lip to roughly 104 to 111 C. NBR continuous rating is 100, and a lip sitting 10 C over its rating for six weeks does exactly what you saw, the accelerator ages out, the lip loses its elasticity, the garter spring cannot pull a stiff lip back to the shaft, the wedge opens, and the drip starts. Six weeks is not a coincidence, it is the NBR lifetime at that temperature, logged by every seals bench I know.
Material answer: FKM, and I will stop there on the polymer, because the second half of the fix is the shaft. If you put FKM on Ra 0.9, you trade a six-week crack for a different failure, a lip that wears a groove into a rough shaft and then leaks through the worn track at the same six-week cadence, because FKM has a harder edge and a rougher track chews it faster. The shaft under a lip seal of this class should be ground or polished to Ra 0.2 to 0.6, preferably 0.3 to 0.5 with a lead in the direction of rotation, and hardness of the shaft surface over about 45 HRC so a foreign particle does not plough a groove that kills the seal anyway. Roughness holds the oil film, but only up to a point, and 0.9 is past the point for a 1750 rpm lip. There is no controversy here, the drawing date is your clue: a five-year-old drawing carries the five-year-old Ra, run surface with a hand polisher and 320 grit paper, then 400, along the axis, fifteen minutes, and the FKM will keep up with the rebuild interval of the gearbox instead of the PM interval of a seal.
One nuance before you buy: keep the garter spring, do not switch to a springless lip, at 1750 with VG 220 the spring is what holds the lip on the shaft across the thermal slack; and check the shaft eccentricity, total runout over 0.15 mm at this size and the lip will follow the shaft and inch a crescent wear pattern even in FKM. I will send the FKM part number with the 55 mm bore dimension and the current spring rate so you can order the correct interference.
From: J. Osei, Maintenance — To: R. Almeida; cc line supervisor
Subject: RE: FKM price and lead time — is there a cheaper first step
FKM seal is roughly three times the NBR price and the first stock arrives next week at best, the line supervisor is asking if we can do one more NBR cycle with an air nozzle aimed at the housing to buy a month while FKM clears customs. My honest answer to him was no, because we would be repeating the same lip at the same temperature and expecting different life, and the nozzle would cool the housing, not the lip, the shaft side is the hot side wrapped in 220 oil. So I will not do that. But I do want to skip a hidden step if there is one: you said check the eccentricity, before I pull the seal, what number am I measuring to, and if it is high, does the FKM fix survive a tilted shaft with a crescent wear, or are we grinding the pedestal too?
From: R. Almeida, Design — To: J. Osei, K. Ivanova; cc line supervisor
Subject: RE: eccentricity budget — measure it with the collar, not the caliper
Measure it the correct way, a dial indicator against the machined seal land, not against the belt pulley, you want the land that the lip actually rides, total indicator runout including any tilt of the housing face. Budget is 0.15 mm TIR, our drawing holds 0.10. You have a stopwatch, not a job, so do it before ordering FKM, because if the runout comes back over 0.15 we also shim the motor foot or bore the end cover true, and then the seal order is the same but the installation is honest. Katerina confirmed radial runout sensitivity of this seal family at this speed, and she says the deviation limit is exactly the crescent-wear threshold she mentioned, so treat it as a hard floor, not a recommendation.
If the reading is in range, then the only parts we change this cycle are the seal, the shaft finish below it and the grease on the sleeve, and the drawing gets a note that the input seal is FKM and the land is Ra 0.3-0.5, with a date and a reason attached so the next five-year drawing review does not quietly revert it to NBR.
From: J. Osei, Maintenance — To: R. Almeida; cc K. Ivanova
Subject: RE: eccentricity reading — 0.12, in budget, FKM done
Dial on the seal land gave 0.12 mm TIR, inside the 0.15 floor, so no grinding, and I fitted the FKM seal on the polished land, went over the track with 400 grit wet along the axis, then cleaned with solvent and a lint-free cloth, sleeve install, filled the housing to the level plug. Data so far: ran three hours at 1750, cover temperature 88 C at the probe, no weep at the drip tin, surface of the seal body feels cooler than the old NBR did, which matches what you said about FKM carrying the temperature. I will log it weekly instead of the old six-week ritual, and if it goes dry past the next oil change interval, the drawing note earns its keep.
From: R. Almeida, Design — To: J. Osei, K. Ivanova; cc line supervisor
Subject: RE: closing — the numbers the thread changed
Closed and logged as DRG-CB07-2212, input seal now FKM, land Ra 0.3-0.5, runout verified 0.12, and the drawing carries a one-line history so the next revision does not go backwards. The thread changed exactly three numbers on the drawing, and none of them is the seal bore: material, roughness and the runout budget. Two questions answered along the way worth keeping. First, lip temperature is never the oil bath temperature, the contact zone earns 18 to 25 C extra for the privilege of spinning at 1750 in VG 220, and the catalogue rating is the room temperature of the argument. Second, seal life is decided at the interface, so roughness and runout are the other half of the material choice.
| Item | Original spec | Changed spec |
|---|---|---|
| Seal material | NBR | FKM |
| Shaft land finish | Ra 0.9 | Ra 0.3-0.5 |
| Runout budget | not stated | 0.15 mm TIR |
| Typical life | 6 weeks | to oil change interval |
Reference data — shaft 55 mm, 1750 rpm, ISO VG 220, oil bath 82-86 C, lip zone +18-25 C; NBR rating 100 C continuous; FKM raised; land 0.9 to 0.3-0.5 Ra; runout 0.12 measured vs 0.15 budget; original leak 30 ml/shift at 6-week cadence.
Glossary — radial lip seal: a circular elastomer seal whose lip presses radially on a shaft, against the oil and the dirt side; garter spring: the coil spring that maintains lip load as the elastomer relaxes; lip temperature: the true contact-zone temperature, above the oil bath; FKM: fluoroelastomer, continuous rating about 200 C, oil- and heat-stable; burntished turn: a smooth turned finish, often too smooth or too fibrous depending on the tool and feed.
Field note — the seal that dripped on schedule
A lip seal that fails on a calendar is rarely a bad seal and almost always a temperature or an interface problem. The oil bath said 85 C and the catalogue said NBR is fine to 100, so the maintenance team did everything right and changed the same seal every six weeks. The missing number was the contact-zone temperature, about twenty degrees above the probe, and once that is on the table the material choice stops being a guess. Then check the two things the drawing usually forgets, surface finish of the land and the runout budget, because FKM on a rough or tilted shaft just moves the six-week failure to a longer, noisier one. Fix the material, the finish and the runout, and the seal becomes a rebuild-interval part instead of a PM consumable.
From: K. Ivanova, Seal Supplier — To: R. Almeida, J. Osei
Subject: RE: FKM draw number and grease note
Attached is the FKM drawing with the 55 mm bore, interference band and spring rate marked, and one practical note from the bench: on install, grease the lip with the same VG 220, never the assembly grease from the kit, because on the very first start a film of clean oil is what the lip needs to lift off the dry land, and a thick assembly grease holds the lip against the shaft for the first thirty seconds of full-speed rubbing, which is precisely when a polymer edge is most vulnerable. That single habit is worth more than half the material advantage in a factory that changes seals in a hurry.
