At the beginning of winter I asked the same question and no one here could really provide me with a black/white answer on the effects of roller size. So, I bought a few different sizes and have been testing intermittently as time allows. I'll take a stab at this and share a few things I've learned along the way, but be advised these are just my opinions.
The roller weight you mention is not as important as roller size. The first part number you list is for 14.5 mm rollers used in 17/18 Winders. The second part number I'm assuming is the updated roller for 19 Winders which are 14.8 mm. I dont think there is enough difference between the two to notice any performance gains/losses. I'm sure the bigger rollers weigh more, hence the descrepency in weight.
Because bigger rollers tuck the weight in more they raise engagement to the tune of 70 rpm per .5 mm increase in roller size. So, if you have 14.5 rollers that engage at 3000 rpm, simply installing 16 mm rollers will bring engagement to 3200 rpm. (1.5 ÷.5 = 3 x 70 = 210 rpm more engagement +/-) So in a sense, you can utilize a primary spring with a low initial rate and raise your engagement by installing bigger rollers. Hold on it gets more convoluted.
Bigger rollers also apply less shift force, which in turn slows clutch shift. Since the shift is slowed, the byproduct is a raise in rpm across the shift range to the tune of 100 rpm extra per .5 mm increase in roller size. Hence, the raise in rpm and reduced shift force can both be remedied by adding more weight and/or more helix angle.
Adding more helix can harness the gain in rpm, but the lower shift force created with the bigger roller coupled with the bigger helix might yield belt slip and produce more heat. In my experience, if you keep the helix constant, I would say every .5 mm increase in roller size will require an additional .5 gram per weight to compensate. Keep in mind all the rules listed above can/will vary depending on primary/secondary springs.
Big rollers permit the use of heavier weights and bigger helix angles. For example, running 76 gram weights, 14.5 mm rollers and a 41/35 helix makes my sled doggy in the trail. Turbo spool is a little delayed, hence for trailriding I prefer a straight 35. Now.....if I install 78 gram weights, 16mm rollers and 41/35 helix, sled lights up and is awesome in trail, but clutches get black from slip. There is always a trade off in everything you do it seems.
I've got 5 sets of rollers I've been testing: 14.5, 14.8, 15.5, 16, and 16.5. I also have a toolbox full of various primary/secondary springs and helixes. In conjunction to all this, I run Dalton adjustable weights and change my clutch setup every ride. Why? Because I'm always looking for that magical setup.....I'm just weird like that.
I like to spin my winder 9150 and I can easily rattle off 10 different clutching combinations consisting of various rollers, springs, weights, helixes that will all produce the desired 9150 rpm, however this is what keeps me up at night:
Which setup is the best on fuel?
Which one will yield cooler clutch temps?
Which one will extend belt life?
Which one will extend roller life?
Which one of is the fastest?
Or, how about which one will do all of the above?
And the testing continues...........
