Originally Posted by seeker333

It's not the balls that wear so much as the cup and cone in Shimano hubs.

Water intrusion due to immersion or cleaning with a pressurized hose will result in rapid corrosion, in the form of pitting of all bearing components. Shimano's labyrinth seal keeps out dirt/foreign matter OK but doesn't stop water - in fact this "seal" will trap water in the bearing cavity where it emulsifies the grease, greatly shortening bearing life. The only remedy is to disassemble, clean and repack.

Normal wear will erode a rounded track into the cup, and to lesser extent the cone, and the bearings wear too although it is easier to observe the track formation. This process takes thousands of miles/hundreds of hours depending on load and hardness of bearing parts.

The cup is pressed in and cannot be replaced on Shimano hubs, spares are not made or sold. Replacing the balls and grease will allow slightly smoother operation even after the wear track becomes evident as suggested in post no. 12. Once you can see the wear track it is time to start looking for a deal on a replacement hub. You can still use the old hub for hundreds/thousands of miles, but a wheel rebuild is definitely coming.

The advantage of a hub which employs cartridge (aka sealed) bearings is that they can be replaced and the hub's life extended infinitely - in theory at least. Many hubs require specialized tools to permit bearing replacement, and the tools may cost as much as the hub, so as a result only a few LBSs might have the tools to facilitate a bearing swap, and the bearings may be an oddball size so that special ordering is required and the cost may not be trivial. DT hubs used to be this way and probably still are. A few hubs do not require special tools and bearings may not be too unusual or expensive (PW, Hope, VO come to mind).

Another part which might fail is the freewheel mechanism, which again may require special tools and an expensive replacement part.

In the end one must accept the fact that hubs are simply another part subject to failure and eventually requiring replacement - one of hundreds on a bicycle. This is why Shimano hubs have historically been a smart choice - they are not the longest lasting but they are readily available and inexpensive. Most of them made in the past ~20 years are designed with the same flange diameter (45mm PCD) and almost same spacing, so that one can simply swap a new hub into an existing wheel and reuse the spokes (probably good idea to rebuild with new nipples).

WRT building wheels just for the experience, it's hard to justify since quite a few wheel suppliers sell wheels for only a little bit more than the price of parts. QBP's "Quality Wheels" are a good deal in economical wheels - you can get a 36h set suitable for a LHT frameset buildup for as little as $150, for a set where the rims and spokes would cost $167 ordered from the cheapest QBP online store (Velocity NoBS rims and WS SG spokes+nipples, which doesn't include hubs and substitutes less costly WS spokes for DT sold in 100 packs)!

It is not too hard to lace a wheel and true it well laterally. It gets more complicated when you then attempt to true the wheel radially (take out the high/lows in the "hop"), then recover the lateral true. It yet again gets more complicated when you move from the easier front wheel build to a dished, asymmetrically-tensioned rear wheel, which is basically all rear wheels with a cassette/freewheel. After you do all this, you need to use a tensiometer to check for under/over tensioned spokes and try to reduce these outliers, which can be 1/3 of the spokes and requires a complete redo of the aforementioned lateral/radial truing steps. Additionally, if you wish to prevent the spokes from their eventual, inexorable detensioning and concomitant de-truing of the wheel, you need to use some kind of threadlocker compound like Spokeprep, Loctite blue or boiled linseed oil (also an excellent steel frame rustproofing agent). I used penetrating Loctite green on my last wheel build, adding it after build was complete to avoid messy handling associated with all other threadlockers.

Building a good durable wheel is not a trivial matter, and in some cases it is not economically sensible.

Sorry this is likely TL/DR for most of you. It could be much longer, but my tall lawn beckons.