Hi Luis. Your query stated BALUN. However, if an UNUN, then a simple R pad, either PI or T with a Z value that matches the UNUN transform ratio is required.
The method in either case requires a measure of only s11 or the return as this is directly tagged back to the insertion loss as loss= 1/(1-s11).
If it is a large transformer ratio, the prior post of measurement of short and open s11 is a method to try.
