I just put a three element up yesterday at two wave lengths in height for 10/11 meters.
Used an old mosley beam as a starter, split element driver with a 1:1 balun at the feed point.
Used the standard dipole formula for the driven element and then 10% bigger for the reflector and 10% smaller for the director.
the three element went together very well and gives a broad bandwidth according to the mfj analyzer, 2:1 at 26.300 to 28.800.
The F/B Seems to be roughly 24 DB and forward gain is better than expected.
Used .15 WL spacing from driven to reflector and .12 from driven to first director, standard 65" and 56" respectively.
Build the beam with all elements in place then tune the driven for lowest SWR at desired freq, then re adjust the reflector and director for the 10% diference in lenght then re-check VSWR.
This is old school practice and as stated before it does take patience to get it correct.
Once correct you will be amazed at the side null of the beam as no lousy matching gamma to deter the beam pattern.
Side nulls fall off the chart when at a 90 degree angle from recieving station.
The balun, reccomend a CURRENT type to force equal amounts of currnet into each leg of the dipole/driven element for equal distribution of current.
Also keep into consideration the balun current rating, I used a 2KW but it is rated in SSB, if you are an AM user suggest you go four times the rating on the balun as to the amount of watts you intend to use so there will not be any heat problems at the feed point balun.
Use some type of sealant on the conections at the dipole, I use liquid electrical tape as it seals everything.
I have not found any baluns that are 100% weather proof no matter what they are advertised as so pay attention to water proofing the balun also.
Antennas are alot of fun to build and experiment with, good luck with your yagi experiment and let us know how it turns out.