Ultrasonic testing can be performed using shear wave and L-wave methods. I am assuming here that you are attempting to examine the entire volume of a butt weld, which would be performed using shear wave. If this is correct, your only limitations are the OD of the pipe, stand-off distance from the weld and the diameter of the transducer for the shear wave process.
If you are trying to inspect thin material, you are going to need to work beyond half node. That will eliminate using angle longitudinal waves which will mode convert to indecipherable L and S waves on the ID surface. As far shear waves are concerned, if the thickness of the material approaches a wavelength of so, the generation of plate or lamb waves is a concern. My own personal rule of thumb is 1/4 inch or so as the thinnest weld I will consider testing with angle beam. I don't have any technical source behind that number, it is just what I wok with. As Metengr stated, the OD of the pipe must also be considered. Beam divergence will occur as you couple through the OD surface.
In the nuclear industry, procedures for stainless steel piping go down to about 0.237" and we don't count any bounces. If we don't have two sided access we get no credit for anything on the other side of the weld because we can't be sure the wave is even getting through the weld. Typically all welds are flat topped so we can go over the top of them to get coverage on the upper weld if required. If we are limited to one side we we use an angled longitudinal wave just to get a little more information.
