[jakeru]

If you attempt to TIG it, consider that the thickness of that tubing could be only .020" thick. That was the thickness of the tubing of an intercooler of similar (welded tube) construction to that radiator that I worked on. Better practice welding on some pop cans! (and get good at them too!)

I would use a low melting temperature (700 F or so...) zinc-aluminum brazing/soldering rod for that application, heated with an oxy fuel torch, if you have access to it, if I were you. It will be much more reliable than an epoxy repair. But if you were to go with an epoxy repair, I would go with one of the ones with metal filler in it, such as the grey marine tex, or JB weld. Any epoxy will only hold well if you can roughen the surfaces before application with an appropriate "tooth" beforehand, such as by sanding.

Although TIG might be fun to try just to see if you can do it. Practice on some aluminum cans first... welding .020" material is a matter of fine control of the welding current on the low end. Use a small electrode like .040" and maybe 1-2 dozen amps. If you have trouble making a fine adjustment of the current knob in the very low range, you can make a fine adjustment of heat into the part by adjusting the AC balance, if you machine has that setting (turning up DCEP / cleaning setting will put less heat into the part, kind of like turning the current down by a little bit). And also rather than reaching for a common 1/16" TIG filler rod, you may find using a smaller diameter MIG wire more suitable for precision dipping. Hope this helps and good luck!

Also for TIG, yes aim your torch at the thicker header plate / end tank, as it will soak more heat away than the thin tubing. Intercooler I worked with of welded tube construction had .050" header plate, with .020" tubing wall thickness. This radiator looks of similar construction. And it is likely that the braize material holding the tubing to the header plate melts only a few dozen degrees lower than the parent metals, which you would not notice any difference between when TIG welding.

 

[jakeru]

Dave powelson - I've got a late model, aluminum automotive radiator core in the garage with its plastic end tanks removed. (Original, used radiator from a 2001 Ford Focus, which had inexplicably developed a tiny pinhole leak.) I measured the tubes that were sticking about 3/8" or so through the header plate using digital calipers, and saw measurements varying from .016" on the thinnest tube (the "outlier"), to a more common .018-.020", and up to about .022" thick on the thicker spots.