First I have to correct myself, MJK's model allows for placement of the top baffle anywhere on the H, default beeing 0.5 of the depth - where the H's baffle also is placed default. I didn't noticed that at first.
The argument against non-symetrical baffles and non-symmetrical placement of the element is that, even if it will help against diffraction effects, it will also corrupt polar response and thus make worse constant directivety. Symmetrical placement will give a little greater baffle midrange hump. This might be countered by listening a bit off axis letting speaker axises cross a bit in front of you in the sweetspot.
Assuming that MJK's model actually takes into account effects that are affected by the top baffle placement. I would if building a setup like this place the element symmetrical on a rectangular top baffle, say 9" from side and 8" from bottom of the simulated baffle - symmetrical thus, but place the top baffle a little behind the H-baffle to even better timealign elements. This will according to simu lessen the hump a bit and preserv the better polar response.
/Erling
