Big Float epsilon always the same independant of precision?

General Usage
1

When I allocate a big float, I can set the precison as a keyword argument, like so:

eps(BigFloat(1.0, precision=10))

However I always seem to get the same floating point epsilon for some reason, no matter the precision I set:

julia> eps(BigFloat(1.0, precision=64))
1.727233711018888925077270372560079914223200072887256277004740694033718360632485e-77

julia> eps(BigFloat(1.0, precision=128))
1.727233711018888925077270372560079914223200072887256277004740694033718360632485e-77

julia> eps(BigFloat(1.0, precision=256))
1.727233711018888925077270372560079914223200072887256277004740694033718360632485e-77

I would like to understand what the precision argument actually means, but it seems like its even less clear now. Can someone tell me what is going on? How large is the floating point error when I set it to a certain precision?

EDIT:
Never mind, this seems to work fine:

julia> setprecision(1); eps(BigFloat(1.0))
1.0

julia> setprecision(2); eps(BigFloat(1.0))
0.5

julia> setprecision(3); eps(BigFloat(1.0))
0.25

julia> setprecision(4); eps(BigFloat(1.0))
0.125

I dont know what the precision keyword is supposed to do, but I think that it doesnt work.