If you can't push them out from the back side as it appears this is the case from the pic if I can see it right. Get a dremmel tool and use a grinding wheel to cut out a 1/4" piece, just be careful not to dig into the case. Once you relese the tension, it should fall out or require vey little force.

OK, that is damned cool!

Yup that made me smile... I'd not heard of that one...real clever.
Mind ya gotta be a pretty smart welder to do it.
Del

Even if you cut very slightly into the casing with the Dremel, it should not cause a problem.

So, here is another way to do the job, one I learned while working in a large truck and idustrial equipment repair shop. I ran into the sam incident on a forklift, and the shop owner, a real genius, told me to run a bead of weld with an arc welder around the center of the race, let it cool and the race fell out. No damage to the housing as long as you have a steady hand at welding. I have used this for many years now and it always works, even on races that are really bound up.

Wow, if ever I need to do this myself, I will try just that. If its as good as you say, many, many thanks....

this is a very old trick , with your arc welder run a bead all the way arround the face of the race let cool and they will fall out nicely.

you can take the housing and heat it up with a torch. the housing will expand before the race and should drive out pretty easy. that is if don't over heat it and the race gets hot. on some aircraft hubs we had to heat in an oven and then use dry ice to cool the race down. the race would just fall out and we did in reverse to install a new on. the oven we use got hot than the one in you kitchen, but i don't remember how hot. it is bad getting old. but i know on my car's hub i just used a torch to get it out and to install i cooled the race in the freezer and reheated the hub up to install. i hope this helps.

When I did a lot of this stuff, I'd work in this order:

1. Use a slide hammer with adapters made for this purpose. The adapter is a cylinder, roughly, with protruding knife edges at the extreme end, which reach under even very small edges (such as that left by the radius ground into the bearing race). The cylinder expands to lock under the edge. Fast, easy, works 80% of the time.

2. Carefully heat the aluminum housing, but to no more than about 300 f (150c) to avoid changing the aluminum temper. If necessary use ice, dry ice, a wet rag, or liquid nitrogen on the race. Sometimes the race will come out simply because of the differences in expansion rate alone (even if the race and hub are the same, elevated temperature).

3. Weld around the inside of the race, as the article suggests. This was a little tricky before the days of LCD welding lenses, (especially if you tried to do it with a stick welder [where you needed to "strike" an arc] but is now simple. I've never done this (or had it done, when I felt my bubble-gum-style skills were not quite up to the task) and had any outcome other than the bearing race falling out by gravity alone.

4. I've never had to go this far, but one could weld a bar across the bearing race (after finding that the fit was too tight for the circumferential weld to work) and whack it out from behind, or from in front with a hook on a slide hammer.

One has to be very careful, the earth should not be connected to bearing housing. This may lead the current passes through the interface between the housing and bearing and may lead to arcing and pitting.

Is it possible to drill/mill two or more holes on the backside and push/press/force the bearing race through these holes? And these holes can remain there for ever for this purpose...

vesyen, That was my first thought, but looking at the bearing housing, that doesn't seem to be a viable solution in this case. Although having a 2nd picture showing the side view would help. In this pic, it looks as there is an inner race and wall thickness that is to shallow for drilling these said holes. Another profile shot would have helped.

Hi Dstrohl,
"By some careful measuring, the parts guy at the local NAPA figured out which races to order."
What is the brand and bearing number of this subject bearing?
This could help in figuring out the thickness of the outer race.

yesyen

do you have trainning material for bearing types its assembly dismentling and analysis, the way you have for pump and mech seal.

I have done the welding to get races out for many years and it does work very well. This method is common practice in the machine shop at the paper mill I work in on bearing races from a couple of inches to 24" OD or larger. One thing I would like to correct is that in the picture the welder is not using a TIG (GTAW) welding but is very clearly using a MIG (GMAW) welder. I think the Invertig 201 welding machine is a constant current welding (TIG or Stick) only machine and it will not weld MIG very well. MIG (other than some flux core MIG, FCAW) is always done on a constant voltage power source. While either will work I would prefer to use a TIG welder for race removal. I say this because a welder has no way of controlling the amount of welding filler metal and spatter he is depositing with MIG and one wrong move and the whole thing is ruined. Steel weld spatter will many times damage aluminum on contact. With TIG welding you put the filler metal in via a separate rod and it is very easy to control everything.IMO One thing I would like to add is that please be careful trying this if you are not a very skilled welder I have seen many ruined hubs from inexperienced people trying this method. A few times I actually had to remove races that was mistakenly welded to the hub while someone was trying this method.

The NDT way, turn on the hot water faucet and let it get hot, run into a bucket. Drop housing in the bucket of hot water; the housing will expand at a much greater rate than the bushing, just keep pushing until the housing expands enough to release the bushing.

Please excuse me if this question is off the subject topic:

When heating a 'hollow cylinder', we all know it expands. My questions are:

1. Does both OD and ID expand?

2. Does it expands about the 'mean diameter'?

3. Does the OD/ID ration has a play in the pattern of expansion? If the ration is high, close to 1, both OD and ID expand. And if ratio is relatively less, does the OD expands and ID contracts? If later is also true, where is the break-even diameter likely?

We have studied of the hollow cylinders expansions, but in practice take granted that heating a housings ease the removal of bearings. Is true in all the cases practically?

Yes

If yes, will the profile of expanded surface be somewhat like shown?

If yes, most of the bearing housings have thicker section at the entry, aren't they?

Is heating/worming some of such bearing housings/brackets might hold the bearings tight rather than easing them out?

Has any one experienced so?

Often it has been my habit to use hot water for fitting steel into aluminium and it's alloys due that the heat is less than the blue wrench and more capable to handle w/o tongs etc., as the heat of water will quickly expand the aluminium though not so much the steel bushings or pins. The bushing or pin will fall out

If a retention compound a temperature of 150°-300° will aid removal