Maybe it would work better if you used a line-noise map, rather than a dot-noise map. Like the kind of image you get with Eye Candy fur function. Do one of those, then greyscale it and use it for the displacement channel. I think Stewer did a tutorial on this a few months ago.