Repeated 6205 bearing failure on a water pump
A municipal water pumping station has three horizontal centrifugal pumps, each rated 90 m³/h at 45 m head. One pump's drive-end bearing, a single-row deep-groove ball bearing 6205 (25 mm bore × 52 mm outside diameter × 15 mm wide), failed after 900 running hours. The inner ring shows a bright polished band, and the shaft seat has a wear step of about 0.03 mm; the outer ring shows fretting marks aligned with the split line of the housing. The log records that the shaft was reclaimed 18 months ago by welding and turning to 24.985 mm, and the housing bore now measures 52.040 mm. The pump handles clear water at 60 °C; the stuffing box is packed with PTFE gland packing and drips about 20 drops per minute. The station engineer must decide, within a two-day shutdown, whether to re-metal-spray and regrind the shaft, replace the housing, or change the bearing arrangement, and whether the sealing should be upgraded at the same time.
1. Check the shaft seat and housing bore against the fits normally required for this bearing (rotating inner ring, stationary outer ring, light-to-moderate load). What do you conclude about the cause of failure?
Read the worked answer
A 6205 bore is nominally 25 mm with a bore tolerance of roughly 0 to −10 µm, so the ring bore is about 24.990–25.000 mm.
Shaft. With a rotating inner ring under load, the seat needs interference: k5 for 25 mm is about +0.002 to +0.011 mm, i.e. 25.002–25.011 mm. The shaft at 24.985 mm gives a clearance of 0.005–0.015 mm instead. The ring creeps, polishes the seat and wears a step - exactly the bright band and 0.03 mm step seen.
Housing. 52.040 mm bore on a 52.000 mm ring OD gives 0.040 mm clearance. H7 for 52 mm is 0.000–0.030 mm, so the bore is 0.010 mm outside even the loosest standard fit. Together with fretting marks along the split, this indicates outer-ring creep and a housing that is out of round or has a loose cap.
Conclusion: both fits are wrong. This is a fit and geometry failure, not a lubrication or load problem, and fitting another bearing without correcting the seats will repeat it.
2. Set out the failure chain and choose the repair for a two-day shutdown, comparing metal spraying, a new shaft and a retaining compound. What measurements must be taken before the decision is fixed?
Read the worked answer
Failure chain: undersize seat → inner-ring creep → wear step → growing clearance → vibration → outer-ring fretting in the housing; the 20 drops/min gland drip may also have washed grease out if the flinger or labyrinth is missing.
Repair options.
- Metal spray and grind to about 25.005 mm, plus boring or sleeving the housing and fitting a new bearing: achievable in 24–36 hours with an outside grinder, and durable. Preferred.
- New shaft: technically the best answer if the wear step exceeds ~0.05 mm or the shaft is bent, but normal lead time rules it out for a two-day shutdown.
- Loctite 638: legitimate as an emergency measure (fills up to about 0.25 mm diametral gap) but it will not restore a stepped seat and does nothing for the housing. Use only to keep the station running until the planned outage.
Measurements first: micrometer readings at three axial sections and two diameters at 90° on the shaft seat (taper and ovality), bore gauge readings on the housing at several points and across the split, and shaft runout on V-blocks. Decide on these numbers, not on the shutdown calendar.
3. Should the sealing arrangement be changed at the same time? Justify your answer with the running conditions.
Read the worked answer
Packed glands on clear water at 60 °C are acceptable, and 20 drops/min is a normal running drip for PTFE packing - the volume of drip is not the fault here. The real question is where those drops go: fit a proper flinger or labyrinth and a drain so water cannot reach the bearing, and check the packing is not over-tightened (which wears the sleeve and heats the shaft).
A single mechanical seal (carbon against silicon carbide, EPDM elastomer for 60 °C water) would eliminate the drip, cut sleeve wear and reduce routine attention. But it needs a sound sleeve surface (roughness around Ra 0.4 µm, no step under the faces) and shaft runout within about 0.05 mm. On the present worn shaft a mechanical seal would leak or fail early.
Decision: upgrade the seal, but only in the same job as the shaft repair, not instead of it. Look for the standard packing drip rate - a dry-running packed gland will destroy the sleeve, while a mechanical seal run dry will destroy itself, so if the pump can lose suction, keep packing or specify a seal with a quench.