If I take two plates (or two sections of tubing) and do a butt joint (assume I grind down the plate edge/tubing edge to allow for full penetration of the weld), then section off a couple coupons, I'm aware of two tests for strength: one is where I bend the coupon to ensure it doesn't fail at the joint, the other is putting the coupon in tension until it fails. My understanding is any coupon failing at the weld is bad; is this wrong? Are there other tests for weld strength that are used?
This comes about because someone I know was told in a class on welding roll cages for NASCAR that the welds will be the weakest part, and that you always use a weaker metal as filler. This seems contrary to what I've read and heard elsewhere, so if one of us is off base, I'd like to know. I assume in the case of a roll cage or airframe, you would take the entire unit and heat treat it to help the HAZ. Is this not the case?
I see the American Welding Society has a publication, D1.1, which probably has all the answers; they want a couple hundred dollars for it, though. If you can point to any online resources which back up your positions, that would be great, too.