Welding rim cracks

I've welded up quite a few rims over the decades.

I really don't like doing it.

To me, it indicates the rim has done it's work ( though, over the years, I've come across quite a few that have done very little work, and that's a worry), and you should be thankful it hasn't collapsed on you.

But, I get cornered by mates, a few times a year, and get, as I said, 'cornered' into helping them out. With the provision that the wheel really should not be used, and, if it is, to keep an eagle eye on it, after every practice / moto. Some crack straight away - some last for a long time onwards, despite my saying to hoik the rim as soon as possible. A few mates have re- welded rims I've done years ago, that are still going strong (bad pun) and I always give them a hard time about not replacing them. 

Near on 99.9% of the time, the crack is at the original weld point. Yes, they are welded, and ground back. It's the inherent weak  / brittle point, and a wheel flex's like crazy. In my work within the bicycle industry, seeing slo mo videos and real durability testing ( at DT Swiss in, of course, Switzerland ) on their Brutal test machines, it's quite the eye opener.

I check for how 'egged' the wheel is, then tighten the spokes adjacent to the crack, and, often loosen some on the opposite side, to try to remove said 'egging' before the welding. If it's significantly ( and by that, I mean if it has more than absolutely bugger all egging) egged, well, I won't even attempt the bodge job. And, if it's an old rim, with the usual internal 'rim cancer' , well, I'll not touch the job at all.

So, a re- weld can be done, but I don't recommend it. 

Yes I've done it and it just cracked again. It's cheaper to get a new rim than pay for a welding the.n later a new rim. 

TSR will restore the finish, they are polished then clear anodised.  No problem welding, you need a TIG.

Tried it twice, both resulted in a collapsed wheel. Makes for a real surprise when the wheel comes out from under the front fender during a jump landing. Don’t do it if the bike gets ridden, maybe only for a showpiece, and even then just don’t. 

Never seen it work well.  Always cracks again.  Not worth the risk.   In order to do it you have to remove the tire and really should de-tension the entire wheel.  If you're that far in, just put a new rim on.

They can be welded. However, there is substantial risk that the rims will break again. 

You can weld them BUT YOU SHOULD NOT. The problem isn't the alloy and there's not particularly special about aluminum rims but the problem comes from the fact they are heat treated AFTER they are hooped and welded at the factory. Welding previously heat treated structural parts without re-heat treating them makes the weld just put down to fix that crack the single weakest area of the rim, causing it to crack again. Sure, you could reheat treat them after repairs are made like they do with exotic and rare car wheels that aren't made anymore, but we can buy new rims for our bikes any day. So, it's really just best to play it safe and pick up a new set of rims.

It's sketchy, as many have cited it can result in catastrophic failure ... but I've done it, too ... and it worked - in a pinch (like to get home on a loooong offroad ride).

This falls into the category of:

"If I am questioning doing this, I probably shouldn't do it."

Because rims, at least modern rims, are in fact heat treated. I pass Pro-wheel going to/from my shop almost every day and have talked to them about why rims crack and why we can't weld them when we have brought in cracked rims for warranty replacements, to which they have told us that the heat treating is the problem and not to do it.     

Quite right. Aluminum in the as-welded state is fairly soft. At the edges of the weld, in the heat affected zone, the material will be brittle. The only way to get uniform mechanical properties after welding is heat treating

That's most likely because they're using resistance welding to fuse the ends. There is no filler wire, so the alloy in the welded zone is chemically identical to the alloy in the unwelded zone. When you introduce a filler wire, that filler wire will be different from the base metal chemically. You'll have pure filler wire alloy in the center of the weld, and where there's admixture (the base metal mixing with the filler metal), you'll have differing chemistry and mechanical properties, eventually getting to pure base metal as you move away from the welded area. Chemical aging or heat treating (normalization or stress relieving) is then used to change the mechanical properties (hardness, yield strength, modulus of elasticity) of the welded zone to more closely match that of the base material.