I've been using layer grid to emit particles from a precomposed text layer, in which words appear and disappear. The precomp is about 300x300 pixels so i thought of using that as the grid size too.
But when particles are emitted from one word, visible in the precomp, and less than half a second later, another word appears and particles should be emitted, they are not. Probably because the max of 200.000 particles/sec and a grid size of 300x300 (90.000 particles) roughly allows 2 full grids a second. Now there is only a small area visible for each word (i guess about 700 pixels), and with only those, and with the max particles/sec i can create many grids per second.
So why do the invisible particles count as particles, isn't there any way around this? (except for resizing the precomp, that would mean i have to position a lot of keyframes (100+), or making a solid with particular for each appearance of a word)
It's an inherent principle of layer grid emitters - by switching to this mode, you explicitely tell Particular to sample all pixels on a plane, regardlesss of color, Alpha and whatnot. In your case the normal layer mode would be a better option, though it's sampling randomly distributed. I agree that some sort of ordered layer mode but without the overhead of the grid would be nice. There is no real solution to your problem. For the time being you should simply limit your particle emission by properly adjusting your layer and grid sizes.