The newer machines is in the fmole range. I dont think detection is a problem, I think it is sample purity. The machine may be able to pick up fmoles, however if you have billions of more abundant "garbage" peptides (proteins) you are picking up, your protein (peptide) of interest may never be found even if its in the detecable limit. This I guess is a needle in the haystack problem, your more abundant proteins are masking your less abundant ones.
Thats why you need to get a pretty low complex sample, a few hundred protein sample can easily be purified even if there are only a few peptides of each.
welcome to the forum!