Taff
Thanks for the very misguided vote of confidence, but I can't add much from personal experience. My pre 2001 600 behaves the same as Simon's on e-start. I still tend to use use both e-start & kick together.
Simon
One thing I intended to check out but never did came from Dale Lineaweaver & it makes sense. He said the problem with the SEM system is that the CDI trigger voltage is marginal at low rotational speeds, even though the charge voltage is probably OK.
With a dirty big cylinder (650!) sitting atop minimal flywheel the rotational speed of the crank will vary considerably across the two-revolution cycle (ie between strokes that slow it down & speed it up). Its slowest point is just after compression - exactly when you need it to spark! While the average engine revs at cranking might seem OK, you really need the compression stroke to be almost totally bled away by the decompressor to keep crank speed high at TDC and get sufficient trigger voltage. I suspect that is why these things are so finicky with regard to decompressor lift and timing. You say these seem as OK as they can be.
You probably only need a poofteenth more rotational speed just at the end of compression. While adding another battery might not seem to increase average cranking speed cranking noticeably, it probably markedly increases the starter torque as it slows down against compression.
At light loads, the starter speed really just depends on battery voltage. As the load on it increases, its speed depends on both battery voltage and the effective series resistance in the starter circuit, most of which is inside the batterty itself. A bigger battery (or two in parallel) doesn't increase its voltage, what it does do is decrease its internal resistance.
Assuming you are stuck with the battery type you have (YTZ7A?) and can't change it, anything you can do to minimise resistance in the starter circuit will help. Bigger starter leads and running a dedicated fat copper ground wire from the frame earth connection to the starter body might help a bit but my experience was not that much.
It might also be worth trying bypassing the starter solenoid main contact somehow and seeing if it helps. I suspect the solenoid contacts are probably marginal at cranking currents against compression - the solenoid itself looks pretty small. I don't know how you might do this - perhaps put both positive cables under the one solenoid bolt and try clamping and unclamping the battery ground wire to the engine somewhere to start / stop the starter. If it does help a larger solenoid from something else might be the go?
Also, I think it was Scully? who had a story about fitting other starters a few years back. It might help if you havent found it already.
Scully's thread
If the starter sounds like it is working correctly, it is unlikely that playing with the brushes will help much.
Sorry for the essay - the first post in a long dryspell......
Hope all is well with you, BW