Penetration of weld and weld build up?

[CGCINC]

Just wondering....

Most people say that you need a build up of weld material approximately the same thickness as the material your welding.
If you get full penatration through the material your welding is this still the rule?
I know you can weld then grind off the welds to look good, which I do a lot, but is the strength still there if you get full penatration with the weld?

[Rambozo]

It all depends on the joint and how you prep it. If it's a butt joint, you can vee both sides to 37 degrees and leave a proper gap between the plates, then fill that with weld back to the top and the joint will be as strong as the filler metal you use. Other joints will require a fillet to be full strength.

[jakeru]

The thickness of the weld, or the "fusion zone" (the metal that has been melted and resolidified, to form the weld) would indeed include both the portion beneath the surface (also known as the "depth of fusion", but more commonly referred to as simply "penetration") as well as the face reinforcement (the part of the weld that rises above the front surface.) Sometimes, the fusion zone can even be "sunken" below the back surface as well, and that counts too.

Assuming your filler rod is the same or greater tensile strength as your parent metals, there is little advantage to thickening the fusion zone by much more than the thickness of the parent metals. Unless you need something like increased ridigity, etc. But to allow for a little porisity, contamination, or other discontinuity, building the weld a little thicker (like say, 120% of the thickness of the parent metals) is not a bad idea, IMO. Failure right down the middle of a weld is not a good scenario, and is also often avoidable with more thoughtful welding technique.

The above assumption may also not be true. Often times, (such as with many aluminum alloys) the parent metal will weaken in the heat affected zone (if it's been work-hardened and/or heat treated, which is the case for many metals.) Depositing a thicker weld won't repair the loss of tensile strength the parent metal will encounter that is next to the weld.

Penetrating fully is better for fatigue resistance, and avoiding a stress riser that can occur in a root that is not fully penetrated when put in tension.

There's a lot that goes into weld joint design, these rules of thumbs assume many things. But the post would be too long to cover all aspects of weld joint design. There are books that discuss recommended dimensions of some more common weld joints, which would serve as a good starting point for destructive testing on welded sample coupons.

[Slartybartfa]

That's an interesting point, something I've wondered for a long time. Say you're welding 6061 T6. Because of the heat of the welding process, the metal on either side of the bead will be annealed back to the T0 condition. So about how far from the weld is the metal's strength/hardness affected? Is that really an issue for everyday projects, or is it only critical where high strength/weight applications are vital such as in bicycle frames?

[jakeru]

6061-T6 loses dramatic strength when annealed. Some of it is restored from letting it age, but nowhere near the original strength. There is no solution, aside from re-heat treating the parts (often not practical), or else doing something alternative to welding in the first place (like bolting parts together or maybe even soldering, provided the metals are kept below their stress relieving temperatures, which in the case of aluminum, rules out brazing.)

Even non-heat treated metals are often delivered in a "cold worked" condition, which raises their tensile strength. For example, copper can be dramatically strengthened by cold working. But this extra strength is lost when they are annealed by nearby welding. Metal formers will often need to anneal a panel being shaped, because it gets too hard partway through the metal forming process. Annealing it lets it become further worked. The good news with annealing is that it generally increases ductility. Which means that at least if/when it does bend, it will "give", rather than suddenly fracture.

Cutting a section, followed by polishing and chemical etching (typically using specific acids applied for specific lenths of time) can reveal the depth of the heat affected zone in the parent metals - as well as the extent of the fusion zone. Sometimes these different areas can even be visible to the naked eye without doing any etching:
Attachment 7267