I have 2 xia-class miners (machine detail: http://www.lketc.com/engoods/show/252.aspx), each works at 1000GH/s and present to their pool worker by mining_proxy, and the mining_proxy shows normal diff 1880. My Internet connection is 1xEV-DO, which redials every 3 hours. Thus, every time redial, the diff will reset to 8, then increase to about 1800 slowly. Before diff increase to 1800, pool shows my hashrate lowly. If Slush pool find a block in a short time, I'll get a very poor income.
So, problem 1: I want to fix my diff at 1024 or higher, is it feasible? and how?
and problem 2: Is the normal diff 1880 be sent to every xia-class completely, or half? even if I have 3 xia-class, diff will be sent one third to everyone?