After doing a lot of reading and some practical experiments recently, I just thought I'd pass on a few things I've learned (or think I've learned). This is not meant to be a comprehensive treatment of 2T exhaust designs, just a few observations that try and explain how they work in simple terms, and might help in selecting the right pipe for your application or in modifying an existing pipe to better suit your needs.
Here's an ultra-simplified breakdown of the pipe sections and what they do (in order from cylinder to tailpipe):
Head Tube (before the 1st cone): this basically determines the maximum rpm (minimum pulse width) at which the pipe will act to suck out (scavenge) exhaust gas. Most modern head tubes are either super short or non-existent to make sure the divergent cone's scavenging effect extends up to redline and beyond, although a long head tube might be used to boost mid and low-end power at the expense of top end. The tube diameter can also have a pronounced effect on overall exhaust-system back pressure.
Divergent Cone(s): this section determines the strength and duration of the scavenging effect during the initial part of the exhaust-port-open period, i.e. while pulling exhaust from the cylinder and fresh mixture from the crankcase. Short, steeply-angled cones give a strong suction effect over a narrower time (rpm) span, while longer/shallower cones provide a reduced scavenging effect over a wider rpm range. Apparently it's quite common for this pipe section to do its job too well, and pull fresh mixture out into the pipe (see 'Convergent Cone(s)' below).
Center Section: delays the action of the convergent cone to reduce the possibility of burnt exhaust being pushed back into the cylinder before the exhaust port closes.
Convergent Cone(s): provides a positive pressure pulse 'echo' that (ideally) pushes any fresh mixture lost into the forward part of the pipe back into the cylinder before the exhaust port closes. Cone angle considerations are similar to the divergent cone, although the length is usually shorter to keep the pulse strong.
Tailpipe/'stinger': provides a means of escape for the exhaust gases, and largely determines the overall exhaust-system back pressure unless a restrictive head tube is used.
Speculation: What follows is based on things I've read, recent mods to my 175's exhaust pipe, and checking out some of my other pipes which I've been happy (and unhappy) with:
The pipe does pretty much the same thing at all rpm's, as the pulse width of even the lowest-rpm initial exhaust pulse is extremely short compared to the pipe length. For example, at high rpm a low-end pipe will continue to pull fresh mixture into the pipe after a high-rpm-only pipe would have stopped, and will still send pressure pulses back to the exhaust port after they'll do any good, perhaps even after the port is closed.
In the absence of pipe action, cylinder scavenging efficiency decreases as rpm increases (assuming a 'ballpark-normal' exhaust port height in the vicinity of 50% of stroke), due to reduced exhaust duration combined with increasing gas inertia. If this were NOT true, increasing exhaust duration by raising the top of the exhaust port would not result in additional high-rpm power. Probably for this reason, every modern pipe design I've seen incorporates head pipes that start diverging practically from the exhaust manifold, providing additional scavenging at high rpm. On a related note ...
Long straight head pipes are not for me. Giving up strong acceleration in the top part of the rpm range - which, to me, is the most enjoyable part of the 2T experience - for the sake of more mids and bottom end just isn't my idea of a good time, although I've tried really hard to like it for many months now. It seems like most modern pipe makers are in agreement with me on this one. I grafted a tapered (divergent) head section onto my 175's pipe, and the difference was night-and-day; like getting back a normal 2T powerband (hard pull to redline) instead of the original surge-and-die response - and in this case, I didn't lose any low end, just some of the mid hit, which mainly just resulted in wheelspin anyway. Vintage pipes seem far more likely to have long head pipes than those used on EVO or modern bikes.
Too-short center sections should be avoided unless absolute max top end power is the goal. It seems better to me to 'go long' with the center section to set the convergent cone's timing to enhance only low and midrange power (when the exhaust port is open longer and there's a greater chance of losing fresh charge into the exhaust) than run the risk of jamming exhaust gas back into the cylinder at lower revs for the sake of a high rpm power gain. I recently tried a pipe with an extremely short center section and convergent cone, and it threw away a bunch of power under 7K for a short burst way up high; another recent experience with an aftermarket 'fatty' pipe using this design was almost identical.
To expand on the above - in my opinion, the convergent cone's design and placement/timing has a lot of potential to reduce low and mid power, through premature pulses stuffing exhaust back into the exhaust port. It seems that while low-mid pipes can also have decent top end, mainly top-end pipes will always have reduced low end power. For most riding, It seems better to me to design a pipe to achieve a guaranteed cylinder full of clean fuel mixture at all times - perhaps accompanied by some fresh-mixture loss out the pipe and loss of fuel efficiency - than one with more potential top-end power combined with the risk of a cylinder (or worse yet, crankcase) contaminated with exhaust gas at low and mid revs. Of course, some applications (road racing, etc.) demand absolute max top-end power, with lows/mids being expendable.
I wonder if the major manufacturers and/or pipe makers use infrared cameras in the pipe headers to find out just how much fresh mixture escapes into the pipe at various engine speeds? I Googled it but didn't find anything relating to motorcycles.
I don't know how many audio/electronics guys we have here, but one thing that really helped me in understanding pipes was realizing that a 2T exhaust is an analog device with a digital input. A pulse (exhaust port opening) is fed into a megaphone (horn), delayed slightly by the pipe's center section, then reflected by the convergent cone (echo chamber). The pipe works on one exhaust event at a time, so it was helpful for me to think solely in terms of elapsed time and single event duration, rather than exhaust repetition rate (which mainly becomes important if the pipe is overly restrictive and builds up excessive pressure at higher rpm's).
All the above is just my $.02. I'm no pipe expert, I just read the papers.
Below are some fairly good examples of the different types of pipes. Note the long head pipe on the low-end unit, the shorter head pipe and longer cones/center section of the wide-range pipe, and the shortness/fatness of the high-rpm pipe at the bottom.