The phasing stub is actually a ½ wavelength section, collapsed over itself, so the total length of phasing stub (i.e. upper horizontal+bend+lower horizontal) is equal to ½ wavelength exactly. So the length of wire before collapsing it into hairpin should be 138mm.OK thanks. That clears up my confusion. I can start on the Franklin now.
For matching stub, horizontal distance from start of stub to tip of bend should be ¼ wavelength. So the total length of wire = ¼ wavelength + length in bend+ ¼ wavelength = ½ wavelength + length in bend i.e. 138mm+length in bend. That means the length of wire in matching stub is 138mm + center-to-center gap, i.e few mm more than ½ wavelength.