The circular law states that if a random matrix is sufficiently large enough and the entries of the matrix is scaled to have mean 0 and variance 1/n, then the distribution of the matrices' eigenvalues will be uniform over a unit circle centered at (0,0).
I know that if the entries of the matrix are scaled to have mean 0 BUT NOT scaled to have variance 1/n, you will still get a circle, but you won't get a unit circle (ie. you won't get a circle with radius = 1).
So if the entries of the matrix are not scaled to have mean 0, will you still get a circle AND/OR will this circle be centered at (0,0)?