I didn't see a more appropriate board to post this on, so here it is.

I have become amazed at how easy it is to bring a bearing back from 'the dead'. I have had bearings completely siezed, unable to turn even a tiny bit, with shields on both sides and a seal on one side, so hardly any access to get at them, and after a few minutes in the dunk tank it spins like a brand new bearing. I read somewhere 'never evaluate a used bearing until after it has been cleaned' but I never dreamed that they could be pulled back from the brink of death like this. Mind you, I have seen bearings with physical damage that couldn't be salvaged, but I've even had a big double-row spherical roller bearing (6"-ish ID) come out of a machine where the seal had been blown through the bearing, and the springs from the seal where all wound around the rollers and the races were full of shredded plastic... it was a mess. A few minutes in the dunk tank and life is beautiful. It's been back in the machine for three weeks working just fine. Gotta love it.

So like the man said: 'never evaluate a used bearing until after it has been cleaned'.

At our plant, we are probably over-conservative, but we never reuse a rolling element bearing. We consider that just the trauma of installing and dismounting a bearing with an inner ring interference fit to the shaft is enough to "taint" the quality of the bearing, so we wouldn't reuse it no matter how good condition. Our motor repair specs say the same thing and I think it's a pretty widespread practice.

It makes sense to me for small bearings where the cost of replacement bearing is much less than the labor for installation. But I cringe when a large $5,000 bearing gets thrown in the trash just because the machine was disassembled/reassembled for reasons unrelated to bearings. Apparently it's a learned practice from some big problems in the old days. I've had no opportunity to judge the experience with the practice of cleaning and reinstalling bearings since we don't do it.

How has your experience been? Are they every bit as reliable as a new bearing?

I'm sure that there is some benefit to automatically replacing bearings, but I don't always have a new one on the shelf (in fact, I rarely do) so cleaning the one I have and putting it back in is often my only choice to get the machine going again. I then order a replacement and put it on the shelf for when the cleaned bearing bites the dust. I haven't had a cleaned bearing go bad yet, but then I've only been doing maintenance for a bit over a month (previously I was a fabricator) so it's not like I've got years of data to go by. The previous maintenance guy let the place go to the dogs, so I've been cleaning and replacing bearings on a regular basis, and aside from 'permanently lubricated sealed bearings' which I'm not fond of, I haven't had to throw out a single bearing yet.
Keep in mind that after cleaning I _do_ rotate them with just a very light oil on them to feel for any burrs they might have. If they're smooth (as they have all been so far) then they go back in.
In the case of that big grinder, the previous guy had put in a brand new bearing and seal in January. The seal alone was well over $100, and the bearing is fiercely spendy. The shaft surface that the seal is meant to seal against is all hogged out, and it's on the end of a shoulder so a speedy-sleeve is out of the question. I'll need to get the shaft welded up and re-turned. I used to do that work, myself, but in my shop at the new job I don't have machine tools other than a drill press with a bent spindle. Neither of my lathes at home will turn the 30+ inch diameter required for this grinder rotor, so I'll have to send it out. Anyway, I bought a new seal for it, but not the spendy viton one that I'm supposed to use. What's the point when it's got no surface to seal against? So I cleaned the bearing and put on a $35 nitrile seal and the machine has been purring for about a month, which is a third as long as the brand new bearing lasted before needing to be serviced. I know it's not the bearing's fault, it's the fault of other things, like the grinder screen and blades are in horrible shape, and the shaft is buggered, but that's all the more reason not to waste money on a new set of bearings for it until everything else can be addressed. That's the way I see it, anyway.
So yes, it takes around three hours to pull the grinder apart to remove and service the bearing, but so far that looks like a good deal to me. If it was a $20 bearing then I'd have to think a lot harder about it.
I will point out, though, that cleaning and re-packing bearings has a long and glorious history from back when we had a less 'disposable' mindset.

Is this a grinder for production purposes? What kind of grinder is it?

Whether or not we automatically replace bearings is largely driven by the cost of the bearing.

The big bearings on press rolls and the like, where the bearing can be $10k +, we clean them up and re-use them if they are in good nick.
Felt roll bearings that cost $50 get thrown away and a new one put in, because it's just not worth the risk.

Pete, we have had bearings in service for many years that have been cleaned and repacked on a number of occasions: 10 years plus, with the bearings cleaned/repacked every 12 to 18 months, so I think that with due care and attention they can be very reliable.

Mike, I'm not questioning your judgement, and I don't know the application, but I do get very nervous when I hear the words "weld repair", "re-turn", and "shaft" all in the same sentence.

Ian

The grinder in question is a 100HP unit that grinds up scrap plastic for re-use. It probably runs 2 hours per day on average. When the blades and screen are in good shape it grinds the plastic into chunks, but when the innards are in poor shape it tends to produce fine shreds which work their way past the labyrinth and into the bearings.

As for welding up shafts, I've done it dozens of times with not a single failure so far. The shaft does need to be re-turned afterwards, of course, and care has to be used in pre-heat and keeping the heat from welding even around the shaft so that it doesn't warp. Usually you put the rotor/shaft on the lathe first and turn down the area to be welded up. At this point you can either press on a heavy sleeve or weld it. If there is a keyway involved then I always weld, but sometimes a sleeve is easier. After welding or sleeving you re-turn to get your dimensions. It works.

That's we do when the budget is tight. So I guess the bearings achieve their L10 life?

Mike,

I certainly wasn't questioning your judgement or ability.

It's just that I've seen quite a lot of shafts fail after they have been weld repaired. The issue comes from the welding/thermal stresses that remain in the shaft, even after the machining work has been completed.

If the application has a substantial radial load, such as a belt driven shaft where there is a lot of belt tension, then weld repairing the shaft without the appropriate post weld stress relief is a dangerous practice, as the shaft will very likely suffer a fatigue failure (snap in two). Depending on the application, it might take a few years before the failure occurs. It's a function of the severity and number of the cyclic bending stresses that are induced in the shaft.

If there isn't much bending stress, then the chance of a fatigue failure reduces quite dramatically.

Without knowing the detail of your application, I'm in no position to judge. I just think that, as a general principle, you have to be a bit careful about weld repairing shafts.

Ian

Welding a shaft isn't _always_ a good idea, of course. It depends quite a bit on what the shaft is made of. Most of the shafts that I have encountered have been a mild steel of the 1018 variety or even milder. There are shafts out there, though, made from 1045, and that stuff is not something to toy with unless you know what you're doing. A very slow cooling is required, and a high tensile filler, as well, not your generic MIG/TIG filler. The approved filler would be something along the lines of 100ksi filler, like 10018 stick filler, or ER100 MIG or TIG filler. Slow heat, full pre-heat, maintain inter-pass temperatures, and slow slow cool-down. If you do all that, then you're golden... if you don't, you're going to have a very unhappy customer on your hands.
With mild steel (the most common case) it's a dawdle, really, if you know what you're doing. Much less work, and the welded shaft is much stronger than a sleeved shaft.
It's important to turn down a millimeter or so from the radius before you weld, because if the transition layer between old and new material ends up on the surface of your turned shaft it will look like hell.

We're kind of getting off topic, though I certainly don't mind. It's nice to be on the 'giving' end of things for a change, since we're into a topic that I know intimately.

Certainly, though, your point is a good one that welding a shaft is not to be done lightly by the inexperienced.

Mike,
From what I understand the reason these bearings are seized up is from the plastic in between the rollers. Have you ever disassembled one of the seized up bearings and inspected it for damage from the plastic shavings? The dunk tank must dissolve the plastic and not scar the race or rollers?

Roller bearing repair with repair potential indicated:
http://www.ahrinternational.com/bearing_repair_and_refurbishment_rollers.htm

Plain bearing repair:
http://www.ahrinternational.com/bearing_repair_and_refurbishment_rollers.htm

Don: After I clean them up and get the bits of detritus out of them I oil them and spin them by hand. If they spin smoothly then I figure they're good to go. If there was a burr or catch when I spun them, then that would indicate to me that either I missed some detritus, or there is a physical problem with a ball/roller.
So for me, basically, if it spins smooth it's good, and if it doesn't it gets cleaned again. If it's still chunky then it goes in the garbage.