OK, I think I've got the clutch system figured out. I've read Olav Aaen's Clutch Tuning Handbook cover to cover twice, read the snowtech arcticle a couple of times, spoken to Olav on the phone this afternoon for about an hour, and performed a bunch of calculations & free body diagrams on the clutch setups. (BTW, I'm a mechanical engineer and when I tweak or custom build something I tend to go a little overboard)
It turns out I was wrong with the helix assumption earlier. As minus40 said it starts out at the high angle and transfers to a smaller angle (I was looking at the picture wrong). This would help to compensate for a higher initial load spring - but, there is a drawback to running a variable helix as well as other benifits.
Starting with an ideal, maximum track hp setup:
In the ideal setup the secondary helix and spring should be setup to provide minimum clamping on the belt, but enough so that it doesn't slip. Since the forces on the belt are highest at low track speeds (low gear ratios) and lower at high track speeds, the only way (that I've seen with current designs & read about) to achieve this is through a small helix angle (say 34 degrees) and a soft secondary spring (Polaris red is good for about 80 ft-lbs of engine torque). Any more clamping than required causes extra belt friction and wasts hosepower. The reason this works is the cam provides a clamping force proportional to the torque being applied. Setting up engagement rpms and wot max rpms are then a function of the primary spring, weights, and cam profile (shape of weight contacting roller).
This "ideal" setup does generate a very large amount of upshift & backshift. ie. when you ease off on the throttle the cv transmission will upshift into a higher gear ratio. This forces the machine to feel less peppy when you get back on the throttle because it takes time for the transmission to backshift to the correct max hp gear ratio.
Running a steeper helix produces less variation in rpms (backshift & upshift)with throttle (torque), keeping the engine running closer to the peak hp rpms at all times (for faster, snappier response). A variable helix improves the throttle response at lower speeds, making the machine snappy out of corners, etc. The drawback to both high angle and worse for variable helix angles is the belt clamping forces are higher than required at high track speeds and therefore some hp is wasted in friction (we're only talking a few percent here - but it's still some). The snappier engine response is the reason most people prefer a variable or higher angle helix. The variable helix also allows the transmission to upshift at high cruising speeds better than just a steep helix, so it is a good compromise for performance trail use.
Since changing the helix angle changes the clamping forces on the belt, it requires appropriate clamping forces at the primary to maintain the correct gear ratios. This is done through the primary spring selection and the weights.
There are lots of options to get what I want. One is I could keep the weights the same (10AL), install a higher initial spring load spring (like a 165/280) and install a 42-34 variable helix. This would give me the higher engagement rpm, better low speed acceleration, the same mid-high speed accleration at 8000 max rpm (after some tweaking), and better throttle response on the trails - but I'd loose a tiny bit of top end and it would cost a bunch of $ (Cdn prices: Helix - $100, primary spring - $25, secondary spring - $25 extra springs, etc. to get it tweaked right) in parts (I still might do it).
Another cheaper option is: Since the system works quite well right now everywhere except off the line and does provide close to optimal track hp, I could simply grind the engagement area on the weights flatter or slightly notched to get the engagement rpm where I want it and leave everything else alone. I'll still have the minor delay when cracking the throttle after a curve on the trails, but everything else should work well, including giving me the maximum top speed (provided the shift rpms are set to max hp).
One other option (that may not get what I want) is to go with one of the EPI kits. They have a kit for my machine that sells for $150 Cdn. They won't tell me anything about what is in it, other than they use a variable helix and two non-stock springs. It might be a well optimised kit, with a high engagement rpm, but it might have way more belt clamping force than required and a lower rpm engagement (than I want).
I had a good look at both the supertorquer web page and snowtech's review of the heel clickers and wasn't all that impressed with the technology. Basically what they are doing is producing a weight with a center of gravity that is farther from the pivot point at small primary clutch movements. This will lower the engagment rpms and increase clamping loads (as they say) at low speeds. While everything stated is reasonably accurate, this can all be done with the shape of more traditional weights. One benifit to their system is it is adjustable, but so are Thunder Products kits (the Super Torquer system is $190 US or $285 Cdn and the TPS kit is only $150 Cdn). It's likely that the Super Torquer profiles are not available stock in the standard Comet or Polaris weights, but they could be reproduced.
Anyway this has been quite a learning experience (and I haven't even started tweaking yet). As Olav says, even when you've got the theory and the math for these systems totally figured out, once you start making changes you'll find you still have to tweak to get exactly what you want.
I'll have to see how things go when I start tweaking.
The great news is it looks like we'll be getting the first decent snow Monday night!