I'm going by what I've read in Jennings and Bell. As soon as the pulse reaches the end of a pipe or an expansion, the negative wave is reflected back. From what I understand, the cone doesn't "lengthen" the pipe, but prevents the loss of kinetic energy that a pulse loses hitting the open end of a pipe by being a more efficient inverter. This means the reflected negative wave retains much of its energy.
I've only skimmed through Jennings a while back, haven't read any Bell in over twenty years, Bell was my kart engine builders bible.
Yes, absolutely correct.
From Smith and Morrisons work, they took pressure tappings and determined that the length and divergence angle changes the duration and amplitude of the scavenging pulse relative to a parallel pipe of the same length. Doubling the length of the parallel pipe yielded similar results to the optimum cone on the half length pipe on the test engine in terms of pressure curves. (a single cylinder four stroke) But a megaphone is superior as;
"A pressure wave entering the diffuser or megaphone is gradually reduced in amplitude with decrease in kinetic energy, This energy is however imparted to the waves resulting from the continuous partial reflection of the main pulse, and moves in opposition thereto, to the open large diameter end. Providing the megaphone is designed correctly, almost the whole of the energy content can be imparted to the returning train of reflected waves assisting in the scavenge, with only a small fraction of the original energy rejected to the atmosphere.
In addition, because of the continuous reflective action of the megaphone in comparison to the instantaneous action of a plain pip-end, the total period over which the scavenging waves arrive arrive at the exhaust valve is considerably prolonged."
I don't have the book (big $) but the latest and greatest treatise on performance two strokes is by Prof. Blair from Queens University Belfast, published by the SAE.
He was heavily involved in GP bike racing for the last few years of the Two Stroke era.
BTW, the MOTA software is based on Blair's work, and can supposedly be applied to any two stroke, including saws. Could save a lot of experimentation for someone that was serious about saw racing.
This could be interesting too, haven't read it yet, just downloading.
http://upetd.up.ac.za/thesis/submitted/etd-10312005-115545/unrestricted/01dissertation.pdf